Queen Essays

Differences in Profit and Nonprofit Organizations: A Study of Effectiveness and Efficiency in General Short-Stay Hospitals

Author(s): William Rushing

Source: Administrative Science Quarterly , Dec., 1974, Vol. 19, No. 4 (Dec., 1974), pp. 474-484

Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management, Cornell University

Stable URL: https://www.jstor.org/stable/2391804

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

Differences in Profit and Nonprofit Organizations: A Study of Effectiveness and Efficiency in General Short-Stay Hospitals

William Rushing

1

Research for this paper was supported by Grant Number HS-00028 from the National Center for Health Services Research and Development. Support of the Tennessee Mid-South Regional Medical Program also facilitated the research, Comments by Anthony Obserschall on an earlier version of this paper are appreciated. as is the computer assistance of Thomas James.

Results of a study of small short-stay profit and nonprofit hospitals show significant differences in organizational rela- tionships. The proportionate number of management and support personnel is negatively associated with the occupancy rate and the proportionate number of production personnel is more positively associated with occupancy rate in profit hospitals but not in nonprofit hospitals. Average daily charges are associated with community wealth for profit hospitals, but not for nonprofit hospitals. It is suggested that the profit- making orientation of hospitals is a significant contextual property which influences the relationships of hospital struc- ture and community wealth to hospital efficiency and effectiveness.'

Despite the apparent fundamental distinction between profit and nonprofit organizations, their differences in organizational efficiency and effectiveness have received little systematic conceptual analysis by organizational theorists and few if any organizational studies have been conducted comparing these characteristics.

The concepts of organizational efficiency and effectiveness date to Barnard's (1 938) analysis. For Barnard, effectiveness was defined in terms of organizational goal attainment and efficiency in terms of satisfaction and cooperation of organi- zational participants. Thompson (1 967: 4-6), however, noted that in scientific management, administrative science, and bureaucratic theory, it is efficiency that is viewed in terms of goal attainment. Etzioni (1 964: 8) spoke of effectiveness in terms of goal attainment, with efficiency being defined in economic terms-the "amount of resources used to produce a unit of output." Thompson's (1 967: 86) perspective was similar; efficiency is assessed in terms of economic criteria and, although Thompson (1 967: 14) did not use the term effectiveness, he did speak of instrumental orientations and criteria, which refer to the extent to which the organizations are able to achieve their goals.

Efficiency will be defined in this study in the same way that Etzioni and Thompson used it, but the approach to effective- ness will be in terms of Yuchtman and Seashore's (1 967) system resource perspective. The general hypothesis pre- sented in this article is that the organizational correlates of efficiency and effectiveness differ depending on whether the organization is a profit or nonprofit organization. Type of organization is viewed as a contextual property for the relationships between organizational characteristics and criteria of efficiency and effectiveness.

THE STUDY AND DATA

Data are from a sample of profit and nonprofit hospitals. The bulk of the data is based on responses to a questionnaire sent to hospital administrators in the 1 05 general short-stay hospitals in the Tennessee Mid-South region that were members of the American Hospital Association in 1 968. The questionnaire was mailed in the spring of 1 969 and included 1 40 occupational titles commonly found in general hospitals. Respondents were asked to indicate the full-time equivalent personnel employed in their hospitals for each category and to add any occupational category present in their hospital, but not listed on the questionnaire; 14 titles were added by all the hospitals responding. Following additional requests, a total of 91 hospitals responded, for an 87 percent response rate.

474/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

2 For brevity, the AHA Guide to the Health Care Field will henceforth be referred to as the Hospitals Guide Issue in this article.

Profit and Nonprofit Organizations

In consultation with experienced hospital administrators and registered nurses, the author and an associate grouped occu- pations into three categories: administrative, production, and hotel. Administrative includes two subcategories-managerial and clerical. The former includes all department heads, super- visors, administrators, and assistant administrators, while the latter includes all persons engaged in information processing activities, such as secretaries, clerks, accountants, bookkeepers, and data processing operators. The production component includes all persons who perform a direct or indirect patient service or whose training qualifies them to perform such a service. Specific occupations include registered nurses, excluding nurses in supervisory positions-directors, super- visors, and head or charge nurses; practical nurses; nurse aides; orderlies; salaried physicians, for example, pathologists; and all types of technicians and therapists. Hotel personnel are concerned with the physical maintenance and cleanliness of the physical plant, as well as the room and board needs of staff and patients, and include such activities as dietary or kitchen, for example, cooks or tray girls; housekeeping, such as janitors or maids; laundry-laundry workers; and building and grounds, for example, painters and yardmen.

Some error is usually involved when occupations are classified according to the organizational functions performed. In the present instance, the performance of administrative functions by registered nurses is especially noteworthy and has been recognized as a source of strain in this role (Christman and Jellinek, 1 967; Lysault, 1 970). There is little doubt that the role of a nonsupervisory registered nurse contains a higher proportion of production activity than the role of assistant administrator, clerk, cook, janitor, or painter. In some instances, however, there are clearly ambiguous cases. For example, dietician could be classified as hotel as well as production, but was finally classified as hotel. In most instances, agree- ment on the classification of the title as primarily administra- tive, production, or hotel posed little difficulty. Moreover, since all ambiguous cases involved very few personnel, errors stemming from the classification of such titles would have no practical significance.

In addition to occupational data, information concerning several characteristics of each hospital and the county in which each hospital is located was obtained from the 1 970 AHA Guide to the Health Care Field2 and the United States census.

Profit and Nonprofit Hospitals

Of the 91 hospitals responding to the questionnaire, 22 were profit making. All but 1 had 95 beds or less-the exception having 1 99. Only 40 of the 69 nonprofit hospitals were in this size range. In order to avoid comparing predominantly small profit hospitals with a substantial number of large nonprofit hospitals, analysis was limited to hospitals with 95 beds or less. Moreover, data for computing measures of efficiency and effectiveness were available only for 37 non- profit and 1 6-efficiency-and 1 4-effectiveness-profit makers. Although conclusions must be qualified with respect to the size range of hospitals, limitation of the study to hospitals of a common size has an advantage, since the effects of size are minimized. Nevertheless, even within this restricted size range, profit hospitals are smaller than nonprofit hospitals. The average number of personnel is 65.67 and 103.50, respectively.

The designation of hospitals as profit making or nonprofit

making was based on information provided in the 1 970

475/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

3 Davis' (1972: 11 ) analysis of different facilities in profit and nonprofit hospitals indicates that for hospitals with over 100 beds, differences in types of service offered have decreased over time. Along with radioisotope, the presence of a premature nursery is the facility that shows the greatest difference between the two types of hospitals, but there has been a decline in premature nurseries in both types of hospitals.

4 The difference may not be limited to hospitals, but may characterize differences between profit-making and nonprofit- making organizations in general. For example, according to Congressional testimony, the profit retained by TVA for investment in its physical plant, expansion of service, and improvements since 1 967 has run almost twice that for private utilities (Nashville Tennessean, May 8, 1 973: 1 7).

Hospitals Guide Issue. Fourteen of the profit makers have corporate structures, none of which at the time of the study were part of chain corporations, while 4 are individually owned and 3 are partnerships. Twenty-six nonprofit hospitals are government owned and 14 others are in the other non- profit category. No significant differences were found between these distinctions on the variables included in this study.

Several differences between profit and nonprofit hospitals have been observed in other studies. Berry (1 967) reported that the former tend to be smaller and to offer, on the average, a smaller range of services. Davis (1 972) observed that from 1 961 to 1 969 profit hospitals increased their admissions and capacity at a higher rate and had fewer assets and personnel per patient than nonprofit hospitals. Davis notes little differ- ence in the occupancy rate, however. Other comparisons indicate that hospital charges to patients may be lower for profit makers in certain areas of the country (Owen, 1970).

A particularly controversial issue concerns differences in the services offered by the two hospitals. Critics of profit-making hospitals charge that they tend to dump the expensive services on which it is difficult to make a profit-such as emergency rooms and obstetrics. They have also been accused of cream skimming-admitting only those patients who will be in the hospital for a limited stay, for example, those with appendectomies, and who therefore will have a high per day charge, since the most expensive services tend to be provided shortly after admission. It is not possible to address these general issues with-the current data, since, as suggested by the results of others (Berry, 1 967), to a substantial degree the type of service rendered is a function of hospital size. Since the hospitals in this study were of a common size range, the problem of differences in services is minimized. A comparison of the two types in the average number of services provided from among 34 of 35 services reported in the Hospi- tals Guide Issue-hospital auxiliary service is excluded- shows an average of 4.69 services for profit makers versus 5.1 5 for nonprofit makers. Comparisons for each of the 34 services yielded only one statistically significant relationship, based on Chi-square; premature nurseries are more apt to be present in nonprofit than in profit hospitals. In this analysis, therefore, it will be necessary to consider the possible influence of this differences

Davis (1 972) has noted that the distinction between profit and nonprofit hospitals may be a misnomer, since both types make a profit. All hospitals tend to bring in more revenue than they spend, with the excess in nonprofit hospitals usually being invested in the expansion of hospital services and operations. This may constitute a significant difference never- theless: nonprofit hospitals may be more disposed to invest their profit in expanded services and operations than profit makers.4

Probably one of the most significant differences between the two types of hospitals is in terms of ultimate control. This, in turn, has an important implication for the decision-making process. Decision making in all organizations is complicated and its relationships to structural properties of organizations are not well understood. Studies of hospitals suggest that decision making in hospitals often is accompanied by conflict between administrators, physicians, and boards of trust (Burling, Lentz, and Wilson, 1 956; Lentz, 1 957; Smith, 1 958; and Perrow, 1965). It is plausible to assume, however, that there is less conflict between these groups in profit than in

476/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

Profit and Nonprofit Organizations

nonprofit hospitals, if for no other reason than that ownership and control are more likely to be lodged in the hands of one group, physicians. Physicians and administrators have no community board of trust to which to answer and physicians may control the administration in profit-making more than in nonprofit-making hospitals. Forty-five percent of the profit makers have physicians as the head administrator in com- parison to only 10 percent of the nonprofit institutions.

This is not to say that conflict between community need, administrative functions, and physicians' concern for medical service and their own professional autonomy do not exist. Even when such functions are lodged in one group, such as when physicians own and control the hospital, individuals in that group may experience considerable role conflict between contending organizational functions. The assumption here is only that such conflict in the decision-making process is less likely in profit-making hospitals, because competing criteria for making decisions are more apt to be systematically subordinated to only one criterion, the economic interest of the hospital.

Because of this, criteria used in making decisions may be clearer in profit-making hospitals. In nonprofit hospitals, decisions are more apt to be based on a variety of criteria: perception of community need; the wishes and desires of members of the board of trust, which may or may not be consistent with community needs or the economic interest of the hospital; empire building tendencies of the hospital administrator; and the outcome of the give and take between administrators, board of trust and physicians. Although deci- sions may be made predominantly on the basis of the phy- sician's profit, as one economic analysis suggests (Pauly and Redisch, 1 973), this does not mean that decisions are made in terms of institutional economic interests.

Thus, there is reason to believe that profit and nonprofit hospitals differ in at least two important respects. First, economic criteria per se are more important in decision making in profit hospitals. Second, more criteria are apt to influence decision making in nonprofit hospitals. It is anticipated that these differences will make a difference in the way the hospital operates as a system.

EFFICIENCY

Efficiency is used here to refer to "whether a given effect [is] produced with least cost or, alternately, whether a given amount of resources [is] used in a way to achieve the greatest result" (Thompson, 1 967: 86). Assessment of efficiency, therefore, requires measurements of both resources and system outcomes or results.

From a purely economic point of view-which is the perspec- tive used for efficiency-hospitals may be assessed in terms of their occupancy rate. All other things being equal, a hospital is more efficient to the extent that it utilizes its facilities more, which may be indexed by the occupancy rate. A given amount of resources produces a greater result when the hospital occupancy rate is high than when it is low.

There have been a number of studies of occupancy rates in hospitals, most concerned with the relationship between hospital size and occupancy. Evidence indicates that occu- pancy increases at a decreasing rate across the range of size (Davis, 1969; Hefty, 1969). The results of the current study are consistent with this, in that the correlation for all 91

477/ASQThis content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

5 Results for the occupancy rate for only one year (1 969) give essentially the same results as those based on a five-year average.

6 The product-moment correlations between total number of beds and total employees are .91 and .78 for profit and nonprofit hospitals respectively.

7 'The proprietary hospital, by definition, regards itself as an economic enterprise, It must do so in order to survive" (Bonnet, 1967: 66-67).

hospitals is positive, with the occupancy rate increasing at a decreasing rate as total hospital size (total employees) in creases. With a logarithmic transformation of size-base 10, the product-moment correlation increases from .22 to .34. Other studies have concentrated on the relationship between types of administration and occupancy rates and on the rela- tionship between occupancy rates and quality of care (Neuhauser, 1971 ).

The present measure of occupancy differs from that used in other studies, since it is based on a five-year average, 1 965 to 1 969, rather than on one year. This was done because inspection of reported occupancy rates in several Hospitals Guide Issues over a period of years revealed that rates of occupancy sometimes fluctuated widely from year to year. Consequently, the rate reported for any particular year might not be representative of the general occupancy rate for a particular hospital. Averaging over a period of years reduces this problem.5

Measures of hospital resources are for personnel resources. They are the proportion of personnel in each of the three occupational categories to the total number of hospital em- ployees. The higher the proportion for each of the categories, the greater the personnel resources in that category.

Since occupancy rate is a function of the ratio of patients to hospital beds, the denominator for occupancy is highly correlated with the denominator for measures of personnel resources (total personnel).6 Ratio terms which have the same denominator are standardized with respect to the denominator, and in this sense the denominator is statistically controlled in the correlation between the ratios. As several writers have noted, however, correlations between variables having a common denominator may be inflated simply because the variables have common terms; zero-order correlations between ratios with a common denominator are not necessarily identical to partial correlations between the numerators in which the denominator is controlled. (See Kuh and Meyer, 1 955; Fuguitt and Lieberson, 1 973; Schuessler, 1 973; and Freeman and Kronenfeld, 1 973.) A central issue in most discussions is the conceptual status of the ratios and whether the ratios themselves rather than the component parts of the ratios (the numerators and denominator separately) are of primary interest. In our case, it is the number of patients in relation to the number of hospital beds and the number of different types of personnel as a proportion of all personnel, rather than the total number of patients and personnel, that is of theoretical interest. The absolute number of patients will increase with the total number of personnel, as well as with the totals for different categories of personnel, but there is no theoretical interest in that. However, some recommend that even when theoretical interest is in the ratios, relations between the component parts may be profitably explored (Fuguitt and Lieberson, 1 973: 141). For this reason, analysis for the ratio terms will be supplemented with partial correla- tions in which total employees is the control variable.

Our hypothesis is that relationships between measures of personnel ratios and occupancy rates will vary between profit and nonprofit hospitals. Profit hospitals are in business to make a profit first and to provide a service second.7 The reverse is the case for nonprofit hospitals. This is not to say that the former are unconcerned with community service or that the latter are not concerned with a profit. The fact is, however, that profit hospitals are primarily economic organi-

478/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

8 Hospital administrators and physicians from the region have volunteered com- ments that hiring policies of community- owned hospitals are not based as much as they should be on the qualifications of the individuals being hired or on the needs of the hospital. The employee's relationship with members of the board of trust is viewed as a factor that is sometimes involved in personnel practice.

9 It is assumed that occupancy rate is the dependent variable. As the discussion indicates, however, this is problematic in the sense that resources may be added in anticipation that they will influence the occupancy rate. In this case, one may view resources as the dependent variable, which, in profit hospitals, vary depending on their subsequent estimated effect on the occupancy rate. The relationship is probably one of reciprocal effects over time. Given the limitations of the current data to one point in time, the investigation of such reciprocal effects is not possible,

Profit and Nonprofit Organizations

zations and nonprofit hospitals are not. Consequently, one would expect economic criteria-that is, a concern for efficiency-to be more salient in decision making in profit hospitals. Specifically, the employment and utilization of personnel is more apt to be made in light of anticipated effects on economic outcomes, for example, bed occupancy. In nonprofit hospitals, other criteria are more apt to influence these decisions.8

There are two aspects to this hypothesis. First, it stipulates that more factors besides the anticipated effect on the occu- pancy rate impinge on decision making in nonprofit hospitals; consequently, as is the case when a number of imperfectly correlated variables besides the one under consideration exert effects on a dependent variable, a lower correlation coefficient would be expected than when only one variable exerts an effect. Hence, the profit-making orientation of a hospital is viewed as a significant contextual variable which operates as a causal factor in the number of variables that influence decision making. A higher measure of association for profit hospitals would constitute support for that causal hypothesis.

Second, regardless of the presence of other factors, the influence of efficiency considerations is viewed as stronger in profit hospitals, since these hospitals are primarily economic organizations and secondarily community service organiza- tions. In this instance, the profit-making orientation of hospitals is considered a significant contextual variable which influences the actual effect of personnel resources on eco- nomic outcomes. Therefore, the regression coefficient for the occupancy rate on the level of resources should be higher for profit hospitals.9

It is in terms of production personnel that the hypothesis is most relevant, since these are the personnel who actually perform services that are unique to hospitals and which patients come to the hospital to receive. In addition, in hospitals whose goals are primarily economic, as high a positive correlation between the occupancy rate and the proportionate number of support personnel, that is, adminis- trative and hotel personnel would not be expected, since these personnel make no direct contribution to patient care and, hence, no visible contribution to economic outcomes. Accordingly, differences between profit makers and nonprofit makers would be expected in the relationship between occupancy rate and the three personnel ratios.

Table 1

Comparison of Characteristics of Profit and Nonprofit Hospitals-Average Scores

Profit Nonprofit hospitals- hospitalst

Production personnel .58 .57

Administrative personnel .25 .23

Hotel personnel .20 .22

Ratio of production personnel to administrative and hotel personnel .19 .17

Occupancy rate 74.40 80.02

N=21, except for occupancy rate, in which case N=16.

t N=40, except for occupancy rate, in which case N=37.

479/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

Results

Before presenting the results, average values for the different variables will be presented separately for the two groups of hospitals. As Table 1 shows, all values are virtually identical, except for the occupancy rate, which is slightly higher for the nonprofit hospitals.

On the average, nonprofit makers are larger. Examination of correlations between size and the independent and dependent variables, however, shows that correlations for the two types of hospitals are almost identical (see Table 2). On this basis, size would be expected to have little or no influence on the relationships under investigation.

Table 2

Product-Moment Correlations between Hospital Size-Total Number of Employees-and 4 Hospital Characteristics of Profit and Nonprofit Hospitals

Profit Nonprofit hospitals hospitalst

Production personnel -.06 .07

Administrative personnel -.20 -.22

Hotel personnel -.1 4 .04

Ratio of production personnel to administrative and hotel personnel .21 .03

Occupancy rate .32 .24

N= 21 except for correlation for occupancy rate, in which case N= 1 6.

t N=40, except for correlation for occupancy rate, in which case N=37.

Product-moment correlation (r) and Spearmen rank-correlatiyn (rs) coefficients are presented in Table 3. Correlations are presented for the three personnel ratios. In profit hospitals, the r is positive for production personnel, but negative for the administrative and hotel personnel; for nonprofit hospitals, the relationship is negative for production personnel and positive for one of the two nonproduction components. All coefficients for nonprofit hospitals are quite small and none is statistically significant. In all three comparisons, the coefficient is considerably higher for profit makers, although only one is significant at beyond the .05 level-for sample size of 1 6, a product-moment correlation must be at least .48 to reach statistical significance at the .05 level with a two-tailed test. Another approximates this level, being significant at the .1 0 level (r=.46).

Results are not due to hospital differences in total number of personnel, since analysis for the absolute number of employees in each of the three personnel categories controlling for total employees yields the same general results. A difficulty in conducting this analysis, however, is that the correlations between total personnel and the total for each category are very high for both groups of hospitals; product-moment correlations vary between .79 and .97. This, of course, poses a problem of multicollinearity. (It thus indicates that analysis based on the components of ratios is not inherently superior on statistical and mathematical grounds to analysis based on the ratios.) In any case with size controlled, partial correlations between number of personnel in each of the three categories and the occupancy rate for profit hospitals are .31 (production personnel), -.32 (administrative personnel) and -.44 (hotel personnel), in comparison to corresponding correlations of

480/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

Profit and Nonprofit Organizations

Table 3

Product-Moment Correlations and Spearman-rank Correlations between 3 Measures of Occupational Structure and Occupancy Rate Profit and Nonprofit Hospitals

Product-moment Partial Spearman-rank correlations correlations correlation Profit Nonprofit Profit Nonprofit Profit Nonprofit hospitals hospitals hospitals hospitals hospitals hospitals

Production personnel .46 -.1 0 .49 -.1 1 .57t -.04

Administrative personnel -.28 .1 3 -.1 8 .1 8 -.36 .09

Hotel personnel -.681 -.01 -.66 -.05 -.721 -.06 Ratio of production to administrative and hotel

personnel .38 -.16 .34 -.14 .671 -.18 .

Profit hospitals (N= 16); nonprofit hospitals (N=37).

t p<.05.

p<.01.

10 In some studies, effort is made to control for the common term by computing partial correlations between the ratios and then controlling for the common term (Freeman and Kronenfeld, 1 973). Fleiss and Tanur (1 971 ) provide mathematical proof, however, that the resulting partial correlations do not differ from the zero- order correlations. Since our two de- nominators (total beds and total em- ployees) are so highly correlated, we would expect very little difference between the zero-order correlations in Table 3 and partial correlations between occupancy and the personnel ratios with total number of personnel controlled. Results are as expected. Comparisons between r's and partial r's are as follows, with partial correlations in parenthesis. For profit hospitals, .46 (.49), -.28 (-.32), and -.68 (-.66); for nonprofit hospitals, -.20 (-.11) , .13 (.18) and -.01 (-.05).

11 The formula for determining probabilities for rs is Z = (rs-0)/(1 IN-1). (Blalock, 1 972: 41 7.)

12 Interpretation of b coefficients is in terms of proportional or percentage increases in the independent variables-the occupa- tional ratios. For example, the b of .761 for production personnel means that on the average an increase of 1 percentage point (.01 ) in the production personnel ratio is associated with an increase of 7.76 in the occupancy rate.

-.07, .09, and -.08 for nonprofit hospitals. Results in Table 3 would not appear to be statistical artifacts of correlations between ratios in which the denominators are highly correlated.10

In addition, inspection of scatterdiagrams indicates that the results are not due to extreme values for a small proportion of hospitals in either group. This can be seen by comparing the magnitude of the rank correlations, which would not be influenced by extreme cases, with the magnitude of the r's. For nonprofit hospitals, each rs is about the same as the corresponding r and it is higher for profit makers in every case. Two are statistically significant at beyond the .05 level." Results, therefore, are not due to extreme values for a small proportion of cases in either group of hospitals.

Table 4 presents the regression coefficients (raw b), with occupancy rate as the dependent variable. All b's are con- siderably larger for profit hospitals and, as noted for the cor- relation coefficients, the signs are in the predicted direction for profit makers and are different from the signs for nonprofit hospitals in two of three instances. Also, in all instances, the standard error of estimate is lower than the regression coeffi- cient for profit makers, but higher in all instances for nonprofit makers.12

Table 4

Regression Coefficients and Standard Error of Estimates of Occupancy Rates on 3 Measures of Hospital Occupational Structure

Profit Nonprofit hospitals hospitals (N= 16) (N= 37)

Production personnel .761 (?.394) -.328 (?.541)

Administrative personnel -.940 (?.870) .642 (?.844)

Hotel personnel -1.75 (?.510) -.245 (?.007) Ratio of production to administra-

tive and hotel personnel .064 (?4.042) -.067 (4?.070)

Results thus indicate that in profit hospitals, but not in non- profit hospitals, increases in utilization are more apt to be linked to production processes than to management and other

481 /ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

support processes, whereas this is not true in nonprofit- making hospitals. Since there are no other apparent differences between the two types of hospitals, results appear to be due to differences in profit orientation. Moreover, analysis of the differences among nonprofit makers with and without a premature nursery-data not shown-yields no significant difference in the variables included in Table 1 or in the coefficients reported for all nonprofit makers in Tables 3 and 4.

Data do not indicate that profit hospitals are more efficient than nonprofit hospitals. If anything, since profit makers have a slightly higher occupancy rate (Table 1), nonprofit hospitals may be slightly more efficient overall. This difference, however, may be due to differences between the two types of hospitals in average hospital size, since size and occupancy are slightly positively related. The focus here, however, is not on differ- ences between the two types of hospitals in efficiency, but on differences in the relationships between the criterion of efficiency and other hospital characteristics. It is the contextual effect of type of hospital on relationships between specified variables that is theoretically significant, not differences be- tween types of hospitals in the specific variables themselves.

EFFECTIVENESS

There is an increasing tendency in conceptual frameworks to consider organizations in terms of organization-environment relationships. Consistent with this tendency is Yuchtman and Seashore's (1 967) conceptualization of organization effective- ness, which views organizations in terms of their ability to exploit resources in the environment. The more an organization can realize in the way of positive inputs from the environ- ment, the greater its effectiveness.

One measure of the degree to which a hospital exploits the resources in the community is the economic resources it obtains from the community, as indexed by average daily cost per patient. Although there are relevant resources besides those measured in dollars, it is clear that the more the hospital takes from the environment in terms of dollars, the more re- sources it obtains from the environment. The significance of income for nonprofit hospitals cannot be discounted; without it the hospital would have no resources. Since the exploitation of resources probably varies according to the availability of resources and hence the opportunity to exploit the environ- ment, a relationship between community wealth and average daily cost per patient would be expected.

A stronger relationship would be expected for profit hospitals than for nonprofit makers, however. For organizations oriented primarily to serving community need, effectiveness would not be gauged in terms of organizational success in exploiting the economic resources of the community. If anything, the reverse might be argued; the less a nonprofit hospital exploits the community, the more effective it is in meeting community needs, assuming that services are constant. In any case, no particular relationship between opportunity for economic exploitation and cost per patient would be expected.

The hypothesis presented refers to the closeness of the relationship as well as the actual statistical effect of one variable on the other. A closer relationship-higher r-between community wealth and patient cost would be expected for profit hospitals, because fewer considerations besides eco- nomic factors enter into the setting of patient costs. A stronger effect of community wealth-a higher b-would be expected

482/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

13 For profit hospitals, total costs are $50.66 and personnel costs $28.94, in comparison to figures of $39.29 and $22.42 for nonprofit hospitals. The percentage of total costs due to personnel costs is virtually identical in the two types-56 percent and 57 percent. For all United States general short-stay profit hospitals with less than 1 00 beds, the average daily total costs and average personnel costs are $63.61 and $32.73, which is somewhat higher than figures for this sample, al- though the percentage of total costs going for personnel is almost the same-58 percent (Hospitals Guide Issue, 1970: 476). For nonprofit hospitals, corresponding figures are $49.99 and $27.25, which is again higher than in this sample, with the percentage of total costs due to personnel -51 percent-slightly lower than in this sample. The proportional difference in costs between the two types of hospitals in this sample and in the United States are virtually identical: total costs are 28.94 percent higher for profit hospitals in this sample, which compares with the figure of 27.27 percent for all United States general short-stay hospitals with less than 1 00 beds. Thus, with the exception that average daily costs are below al' other United States general hospitals of com- parable size-which may be due to regional differences in hospital costs (Hinchey, 1970), figures for this sample are quite similar to aggregative figures for all United States hospitals with fewer than 1 00 beds.

Profit and Nonprofit Organizations

because economic criteria are apt to be given greater weight in hospitals primarily designed to make a profit than in hospitals not so designed.

Results

Two measures of average daily patient costs are used: total cost and personnel costs. Both are based on data from the Hospitals Guide Issue and are computed by first determining the average daily hospital expense, which is obtained by dividing the reported annual expense by 365. This figure, in turn, is divided by the average daily census, reported as census in the Guide Issue.13

Community wealth is indexed by the median family income (1 960) of the county in which the hospital is located, as reported in the United States census (United States Bureau of the Census, 1967). Because there is more than 1 hospital in some counties, more hospitals than counties are included in the analysis. In all, the 1 4 profit hospitals which report cost data are from 11 counties and the 37 nonprofit hospitals are from 33 counties. The average for the median family income is slightly higher for the profit hospitals-$3,542 versus $3,1 06. It is the relationship of average daily costs to com- munity income that is of significance, however.

Results for both total costs and personnel costs are consistent with the hypothesis. For total costs, r=.53 (p <.05) for profit hospitals and .1 5 for nonprofit hospitals, while b's are .0041 6 (?.001 93) and .00108 (?.0011 9). For personnel costs, r's are .56 (p <.05) and .14 and b's are .00274 (?.00117) and .00056 (?.00066). Results support the hypothesis that profit hospitals exploit the economic resources of their environments more effectively than do nonprofit hospitals.

Again, the magnitude of the coefficients for profit hospitals is not due to one or two extreme cases: Spearman rank correla- tions for the relationship between community income and average daily total costs and average daily personnel costs are .58 (p < .04) and .79 (p < .01 ), respectively, for profit hospitals and .21 and .1 8 for nonprofit makers.

CONCLUSION

The differences in the correlates and probable determinants of efficiency and effectiveness between profit and nonprofit hospitals stem from the fact that one type is primarily an economically oriented organization and the other is not. Results are consistent with the postulate that in profit-making organizations, economic outcomes and criteria exert greater influence on decision making and hence on organizational process. It is possible, of course, that results are due to factors other than those that have been postulated. In any case, it would appear that the profit-making orientation of hospitals, and possibly other types of organizations as well, is a sig- nificant organizational property that influences the relationship between intraorganizational variables. Research on other types of profit and nonprofit organizations, as well as profit and non- profit hospitals of different size ranges, will be required before it is certain that these findings are specific to the hospitals included in the current study or if they characterize organiza- tions in general. Such research may yield significant theoretical results by indicating that many propositions in organizational theory do not apply equally to profit and nonprofit organiza- tions and, hence, need to be specified with respect to this factor.

William Rushing is professor of sociology in the Department of Sociology and Anthropology at Vanderbilt University.

483/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

REFERENCES

American Hospital Association 1 970 AHA Guide to the Health

Care Field. August.

Barnard, Chester I. 1 938 Functions of the Executive.

Cambridge, Mass.: Harvard University Press.

Berry, R. E., Jr. 1 967 'Returns to scale in the

production of hospital services." Health Services Research, 2: 1 23-1 39.

Blalock, Hubert M., Jr. 1 972 Social Statistics. New York:

McGraw- Hill.

Bonnet, Philip D. 1 969 "The proprietary hospital."

In The Changing Composi- tion of the Hospital System- A Report of the 1 969 National Forum on Hospitals and Health Affairs: Duke University, Durham, N.C., Graduate Program in Hospital Administration, 63-72.

Burling, Temple, Edith M. Lentz, and Robert N. Wilson 1956 The Give and Take in

Hospitals. New York: Putnam.

Christman, Luther, and Richard C. Jelinek 1 967 "Old patterns waste half

the nursing hours." Modern Hospital, 108: 78-81.

Davis, Karen P. 1 969 A Theory of Economic

Behavior in Non-Profit Private Hospitals. Doctoral dissertation, Rice University.

1 972 "Economic theories of behavior in nonprofit, private hospitals." Economic and Business Bulletin, 24: 1-1 3.

Etzioni, Amitai 1 964 Modern Organizations.

Englewood Cliffs, N.J.: Prentice- Hall.

Fleiss, Joseph L., and Judith M. Tanur

1 971 ''A note on the partial correlation coefficient." The American Statistician, 25. 43-45.

Freeman, John Henry, and Jerrold E. Kronenfeld 1 973 ''Problems of definitional

dependency: the case of administrative intensity." Social Forces, 52: 108-21.

Fuguitt, Glenn V., and Stanley Lieberson

1 974 "Correlations of ratios or difference scores having common terms." In Herbert L. Costner, editor, Socio- logical Methodology, 1 973-74. San Francisco: Jossey-Bass, Inc.

Hefty, T. R. 1 969 "Returns to scale in

hospitals: a critical review of recent research." Health Services Research, 4: 1 67-1 80.

Hinchey, S. B. 1 970 An Analysis of Interregional

Variation in General Hos- pital Expenses. Doctoral dissertation, University of Missouri.

Kuh, Edwin, and John R. Meyer 1 9 55 "Correlation and regression

estimates when the data are ratios." Econometrica, 23: 400-1 6.

Lentz, Edith M. 1 957 "Hospital administration:

one of a species." Adminis- trative Science Quarterly, 1: 444-463.

Lysault, Jerome P. 1 970 An Abstract for Action.

New York: McGraw-Hill Book Company.

Nashville Tennessean 1973 May8:17.

Neuhauser, Duncan 1 971 The Relationship Between

Administrative Activities and Hospital Performance.

Chicago: Center for Health Administration Studies.

Owen, Arthur 1 970 ''Can the profit motive save

our hospitals?" Medical Economics, 47: 76-111.

Pauly, Mark, and Michael Redisch

1 973 "The not-for-profit hospital as a physicians' coopera- tive." American Economic Review, 63: 87-99.

Perrow, Charles 1 965 'Hospitals: technology,

structure and goals." In James G. March (ed.), Handbook of Organizations, 910-971, Chicago: Rand McNally.

Schuessler, Karl 1 973 "Ratio variables and path

models." In Arthur S. Gold- berger and Otis Dudley Duncan, editors, Structural Equation Models in the Social Sciences. New York: Seminar Press.

Smith, Harvey L. 1 958 "Two lines of authority:

the hospital's dilemma." In E. Gartley Jaco (ed.), Patients, Physicians and Illness: Source Book in Behavioral Science and Medicine, 468-478, Glencoe, Illinois: Free Press.

Thompson, James D. 1 967 Organizations in Action.

New York: Wiley.

United States Bureau of Census 1 967 County and City Data

Book, 1 967 (A Statistical Abstract Supplement). Washington: U.S. Govern- ment Printing Office.

Yuchtman, Ephraim, and Stanley E. Seashore 1 967 "A system resource ap-

proach to organizational effectiveness." American Sociological Review, 32: 891-903.

484/ASQ

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 06:58:01 +00:00������������

All use subject to https://about.jstor.org/terms

ORIGINAL RESEARCH published: 05 May 2020

doi: 10.3389/fpubh.2020.00124

Frontiers in Public Health | www.frontiersin.org 1 May 2020 | Volume 8 | Article 124

Edited by:

Jason Scott Turner,

Rush University, United States

Reviewed by:

Daniel Skinner,

Ohio University, United States

Kate E. Beatty,

East Tennessee State University,

United States

*Correspondence:

Tatiane Santos

[email protected]

Specialty section:

This article was submitted to

Public Health Policy,

a section of the journal

Frontiers in Public Health

Received: 10 February 2020

Accepted: 27 March 2020

Published: 05 May 2020

Citation:

Santos T (2020) Non-profit Hospital

Targeted Health Priorities and

Collaboration With Local Health

Departments in the First Round

Post-ACA: A National Descriptive

Study. Front. Public Health 8:124.

doi: 10.3389/fpubh.2020.00124

Non-profit Hospital Targeted Health Priorities and Collaboration With Local Health Departments in the First Round Post-ACA: A National Descriptive Study Tatiane Santos*

Health Systems, Management and Policy Department, Colorado School of Public Health, Aurora, CO, United States

We examined the community health needs assessments (CHNA) and implementation

strategies of a national sample of 785 non-profit hospitals (NFPs) from the first round

after the ACA. We found that the priorities targeted in the implementation strategies

were well-aligned with the top community health priorities identified in CHNAs as

reported in previous studies. The top five targeted priorities included obesity, access

to care, diabetes, cancer, and mental health. We also found that 34% of sample NFPs

collaborated with their local health department (LHD) to produce a single CHNA for their

jurisdiction. Non-profit hospitals that collaborated with a LHD on the CHNA had higher

odds of selecting behavioral health community issues (i.e., substance abuse, alcohol,

and mental health), while hospitals located in counties with high uninsurance rates had

lower odds of targeting these community issues. Our contribution was 3-fold; first, we

examined a large sample of implementation strategies to extend on previous work that

examined CHNAs only. This gives a more complete picture of which community issues

identified in the CHNA are actually targeted for implementation. Second, this study

was the first to present information on the status of NPF collaboration with LHDs to

produce a single CHNA (from the NFP perspective). Third, we examined the association

between targeted priorities with NFP and county-level characteristics. The community

benefit requirement and Section 9007 of the ACA present an opportunity to nudge

NFPs to improve the conditions for health in the communities they serve. The ACA has

also challenged institutions in the health care sector to approach health through the

social determinants of health framework. This framework moves beyond the provision

of acute health services and emphasizes other inputs that improve population health. In

this context, NFPs are particularly well-positioned to shift their contribution to improve

population health beyond their four walls. Section 9007 is one mechanism to achieve

such shift and has shown some promising changes among NFPs since its passage as

reflected in the findings of this study. This study can inform future research related to NPF

community benefit and local health planning.

Keywords: non-profit hospital, community benefit, implementation strategy, community health needs assessment,

local health department, collaboration

Santos Non-profit Hospital Community Health Prioritization

INTRODUCTION

Non-profit hospitals (NFP) are exempt under Section 501(c)(3) of the Internal Revenue Code. This tax exemption comes with a community benefit requirement which obliges NFPs to invest in the health and healthcare of the communities they serve. This community benefit requirement was first introduced in 1969 by the Internal Revenue Service (IRS) but the agency never specified what community benefit meant and what it should entail. Prior to that, the IRS required NFPs to provide charity care to the uninsured and underinsured. Hospitals had a relatively great degree of flexibility in determining the amount of charity care they would provide. This was a much narrower obligation compared to the concept of community benefit which was not limited to the direct provision of healthcare services but also included education, research, and activities that promote community health (1).

It was not until decades later, in the 1990s and 2000s, that many government organizations and advocates started voicing their concerns about the practices of NFPs in respect to this requirement. Their main concern was whether NFPs were making sufficient community benefits investments to justify their tax exemption. Public concern was well-justified considering the sizeable value of tax exemption for NFPs which was estimated to be $24.6 billion in 2011 (2). This study was deemed exempt from review by an Inter-Institutional Review Board.

In 2009, the IRS added Schedule H to Form 990 which all NFPs must file in order to keep their tax-exempt status. Non- profit hospitals are required to report their community benefit expenditures in eight categories under “Financial Assistance and Certain Other Community Benefits at Cost” (Part I of Schedule H), and nine categories under “Community Building Activities” (Part II of Schedule H). Schedule H was a clear improvement in increasing the accountability of NFPs through reporting; however, it still fell short on providing a clear definition of what was entailed in each of the new community benefit spending categories. It also did not provide clear guidance on how NFPs should allocate their community benefit dollars across categories. A 2011 study reported that NFPs spent ∼$62 billion on community benefit of which 92% went to charity care, subsidized health services, and education and research (2). While these areas of spending are beneficial to the community, they represent only a partial fulfillment of the community benefit requirement per the IRS (2–10). Non-profit hospitals are also expected to improve the overall health of the communities they serve by providing health care and prevention activities outside its four walls.

Section 9007 of the Patient Protection andAffordable Care Act (ACA) further defined the role of NFPs in improving population health through its requirements for a triennial community health needs assessment (CHNA) and implementation strategy; and further clarification of their financial assistance policies (11, 12). This was another regulatory attempt to steer NPFs toward higher levels of engagement in community health.

While this new requirement increased accountability and transparency, it left NFPs to decide how to approach the actual implementation of the CHNA. The IRS instructions for Form

990 and Schedule H explain that “CHNA must take into account input from. . . those with special knowledge of or expertise in public health. . . ” (13). The IRS only loosely suggests that NFPs should engage experts in public health but leaves room for wide variation across NFPs in how they obtain such input. Furthermore, NFPs have a lot of flexibility when selecting priorities to target through interventions (i.e., as reflected in their implementation strategy).

Non-profit hospitals are required to make their CHNAs publicly available. While there’s no requirement to make implementation strategies available, the majority of NFPs also make these documents publicly available on their websites. This has provided researchers with a wealth of data on how NFPs conduct their CHNAs, how community issues are prioritized, and importantly, which community health priorities are actually targeted in implementation strategies.

Many studies have conducted content analyses of CHNAs and implementation strategies to better understand how NPFs are engaging with their communities to improve community health (14–20). The majority of these studies focused on single states or specific community issues (e.g., violence) (16–19). There were two larger studies that examined a national sample of CHNAs, but not the accompanying implementation strategies (14, 15). The first one examined 300 CHNAs mostly from the first round after the ACA (14). It found that the top five drivers of community health needs identified by NFPs were: access to care, preventive and screening services, chronic condition management, socioeconomic factors (e.g., poverty, housing), and insurance coverage (14). The authors also found that the top five conditions identified in their CHNAs included: obesity, behavioral health, substance abuse, diabetes, and cancer (14). The second study examined 300 CHNAs by NFPs in the second round after the ACA. The coding framework was slightly different for the second study, but overall the findings aligned with the earlier study. For example, they found that the top five health conditions identified in the CHNAs were: obesity, behavioral health, diabetes, substance abuse, and chronic disease (cancer was ranked 6th) (15). Both of these larger studies examined only the community needs identified in the CHNA but not the priorities selected by NFPs for actual implementation.

Non-profit hospitals take into account many factors that go beyond the most prevalent community issue in order to select CHNA-identified priorities for targeting through interventions. Specifically, NFPs use a combination of the following criteria to prioritize and select community issues to address in their implementation strategies: prevalence and incidence, local stakeholder input, available resources and community assets, community readiness and engagement, needs of medically underserved/low income population, the hospital’s expertise in the health priority, the hospital’s mission, availability of evidence- based interventions, and an evaluation of whether other local organizations are addressing the health priority. The result of this process is that while the CHNA may identify several community issues, the NFP usually selects only a handful of local priorities to target during the ACA-imposed 3-year cycle. Sometimes the selected priorities are not necessarily the most pressing need in the community. One study of NFPs located in Pennsylvania

Frontiers in Public Health | www.frontiersin.org 2 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

found that while 87% of hospitals in the sample identified dental health as a community need, none actually targeted dental health in their implementation strategies (17). Other examples from this study include 100% of study hospitals identifying access to primary care as a community need, while only 50% targeted interventions toward the identified need (17).

Our study seeks to fill a gap in the literature by examining both the CHNAs and implementation strategies completed in the first round post-ACA by 785 NFPs. We performed content analysis of implementation strategy documents and identified the top 13 community needs that were actually targeted for interventions. We described the organizational, financial, community benefit expenditures, and community characteristics of these NFPs. We also collected information on the number of community needs targeted per NPF, and whether NFPs and local health departments (LHD) worked together to produce a single CHNA for their communities in 2012–2013. Finally, we examined the relationship between the community needs targeted and hospital characteristics, community benefit spending, collaboration with LHD, and community characteristics.

MATERIALS AND METHODS

Data Sources We obtained copies of publicly-available CHNAs and implementation strategies on the websites of NFPs between April 2019 and August 2019. All of these reports were from the first round after the ACA. More specifically, all CHNAs were conducted in 2012 and all implementation plans were completed in 2013. The study sample of NPFs represent diversity in geographic area (33 states represented in sample), urban/rural status, hospital size, system membership, and teaching status.

Data on hospital characteristics came from the Centers for Medicare & Medicaid Services (CMS) Healthcare Cost Report Information System and the American Hospital Association (AHA) Annual Survey. Data on community benefit spending by NPFs came from the IRS Statistics of Income database (Schedule H). On Schedule H, hospitals report net expenditures (cost minus offsetting revenues) for selected categories of community benefit. County-level demographic, socioeconomic, and labor market measures came from the American Community Survey. We also collected information from the Henry J. Kaiser Family Foundation and Center for Medicare and Medicaid Innovation to define Medicaid expansion status and State Innovation Model participation, respectively.

We used NFP and county-level data from 2013 for our main analyses because the implementation plans used in this study were completed in 2013 for all NFPs in our sample. We also ran analyses using 2012 data (results not presented here but available upon request) and the findings were virtually the same. On average, NFP organizational and financial characteristics do not change substantially from one year to the next. Some circumstances under which characteristics change more significantly include hospital mergers, closures, switching to for profit status, among other local market shocks that may influence hospital finances. The same applies to county- level characteristics. These tend to be stable from year to

year, unless significant shocks occur. One example, would be the 2007 great recession in the US which had a significant impact on unemployment, uninsurance, and other county- level characteristics.

Methods There were three main components to our methods including: content analyses and coding of NFP CHNAs and implementation strategies; descriptive statistics for the NFPs in the study sample; and bivariate analyses to examine the association between the priorities targeted by NFPs and a set of hospital and community characteristics.

We conducted content analysis of CHNAs and implementation strategy reports prepared in the first round after the ACA by 785 NFPs (i.e., 2012 and 2013). The inclusion criteria for this study were counties: (1) that had a 1:1 ratio of LHD to county; and (2) that had one to five NFPs. We wanted to ensure that counties were comparable from a public health resource and capacity perspective because the CHNA process is directly tied to both characteristics. Furthermore, we also wanted to identify whether NFPs collaborated with their LHDs to produce a single CHNA. This type of collaboration may be more straightforward in cases where there is only one LHD in the county. We did not limit to counties with only one NFP because it would have significantly reduced our sample size. Figure 1A shows the geographic distribution of study sample NPFs across the United States. As shown in Figure 1A, there is reasonable geographic diversity in the study sample.

We developed a coding framework based on previous studies of NFPs CHNAs and implementation strategies (14, 15, 17). Specifically, we grouped selected priorities under two main groups: drivers and conditions. Drivers include the structural and social factors that are associated with health status, while conditions are the diseases and health concerns experienced in the community (14, 15). Examples of drivers include access to care, care coordination, and public planning. Examples of conditions include obesity, diabetes, and cancer. We further collapsed the drivers using the County Health Rankings framework for clinical care which includes access to care. Access to care as conceptualized by the County Health Rankings framework includes areas such as transportation, insurance coverage, and primary care (17, 21).

We primarily used the 2013 implementation strategy reports because these documents include information on the selected health priorities and their respective initiatives to be implemented by NFPs over the 3 years following the CHNA. We used the 2012 CHNAs when the implementation strategy for 2013 could not be located. Some hospitals combine the CHNA and implementation strategy in one report, in which case, the targeted priorities and implementation plan can be identified. For a few cases, we identified NFP’s 2013 targeted priorities and their respective implementation strategies using 2015/16 CHNA reports because the 2012 CHNAs were no longer available. Non-profit hospitals are required to report their progress on previously targeted priorities in subsequent CHNAs. The majority of hospitals for which we could not identify their selected 2013 health priorities included hospitals that closed

Frontiers in Public Health | www.frontiersin.org 3 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

FIGURE 1 | (A) Geographic distribution of sample non-profit hospitals. (B) Geographic distribution of sample non-profit hospitals: by status of collaboration with local

health department. Authors’ analysis of data from the IRS, Centers for Medicare and Medicaid Services, and non-profit hospital (NFP) community health needs

assessments (CHNA) and implementation strategies.

Frontiers in Public Health | www.frontiersin.org 4 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

during the study period, opened after 2013, or switched to for- profit or public status.

The CHNAs and implementation strategies were coded by the author and a research assistant using Nvivo software (QSR International Pty Ltd., Version 11, 2015). Both researchers coded 44 randomly selected documents to compare the consistency of coding. Coding was compared through an iterative process until reaching agreement greater than 90%. All remaining reports were coded by the study author.

We provided the descriptive statistics of sample NFPs, county- and state-level factors. We also provided the descriptive statistics stratified by NPFs that collaborated with LHDs and those that did not. We compared the two groups using bivariate analyses (chi-square test for categorical variables and two-sample t- tests for continuous variables). We used two-tailed tests for these comparisons and report findings at the conventional 0.05 significance level.

Finally, we conducted logistic bivariate regression analysis to examine the relationship between the targeted priorities and a set of NFP and county-level characteristics. We report the two-tailed p-values at the 0.1, 0.05, and 0.01 significance levels.

Study Measures We present the findings for the top 13 targeted priorities as reflected in implementation strategies including: access to care, obesity, heart, diabetes, cancer (prostate, lung, breast, colon, and cervical coded separately), substance abuse (use of prescription and/or illicit drugs), mental health, alcohol, and tobacco. Other categories were selected by a small percentage of sample NFPs (e.g., housing, oral health, and liver disease) which aligns with findings from previous studies (14, 15). We also coded whether or not the NFP and LHD produced a single CHNA for their jurisdiction. This information came primarily from the CHNAs, as well as additional Web searches to ascertain that both institutions had collaboratively developed only a single report. Figure 1B shows the distribution of jurisdictions in which NFPs and LHDs produced a single CHNA.

Hospital organizational characteristics were extracted from CMS’ Healthcare Cost Report Information System and AHA’s Annual survey. These included: hospital bed size, number of psychiatric beds, system membership, teaching status, church affiliation, rural status, critical access status, and whether the hospital was a children’s hospital. Hospital financial indicators were extracted from CMS’ Healthcare Cost Report Information System database. The financial indicators included: net patient revenues (total dollars earned from providing patient care after contractual allowances and charity care); operating margin (ratio of the hospital operating income to operating revenues); and total margin (ratio of the hospital total income to total revenue). We also included two indicators of community benefit spending by NFPs which were extracted from the IRS Statistics of Income database. The community benefit spending indicators included total community benefit spending and population health spending (total spending on community health improvement, cash and in-kind contributions, and community building activities). We standardized the community benefit spending measures by dividing each indicator by the NFPs total

TABLE 1 | Ranking of top 13 priorities targeted in 2013 implementation strategy.

Rank Priority Non-profit hospitals

n (%)

1 Obesity 590 (75.2)

2 Access 557 (71.0)

3 Diabetes 400 (51.0)

Cancer 419 (53.4)

4 Breast cancer 160

5 Colon cancer 82

6 Lung cancer 73

7 Prostate cancer 63

8 Cervical cancer 41

9 Mental health 397 (50.6)

10 Cardiovascular disease 307 (39.1)

11 Tobacco 303 (38.6)

12 Substance abuse 268 (34.1)

13 Alcohol 136 (17.3)

Authors’ analysis of data from non-profit hospital community health needs assessments

and implementation strategies.

operating expenses. Hospital market characteristics included market concentration (Herfindahl-Hirschman Index, HHI). Data from CMS was used to calculate HHI.

County-level factors were extracted from the American Community Survey and included: uninsurance and unemployment rates, median income, and race distribution (white, black, and other). Finally, state-level indicators included Medicaid expansion status in 2014 (i.e., extracted from the Henry J. Kaiser Family Foundation) and participation in Round One State Innovation Models during 2013 (i.e., extracted from CMS’ Center for Medicare & Medicaid Innovation). All study measures were operationalized using 2013 data with the exception of Medicaid expansion which reflected the state’s decision to expand Medicaid in 2014.

RESULTS

Table 1 presents the ranking of the top 13 health priorities targeted by NFPs in 2013. Over three quarters of NFPs targeted obesity making it the priority that was targeted most often by NPFs in the 2013–2015 implementation cycle. Access to care was a close second with ∼71% of NPFs targeting interventions to address it. Diabetes was also targeted by the majority of NFPs, and ranked third for 2013 targeted health priorities. Interestingly, these findings align with previous studies that examined national random samples of NPF CHNAs (14, 15). The ranking order of the remaining 10 targeted health priorities was different in comparison to those found in previous studies; however, they were found to be among the top 10 priorities in the studies (14, 15).

Table 2 presents the descriptive statistics for all sample NFPs. We also stratified NPFs by whether they collaborated with a LHD to produce a single CHNA. Some notable differences exist between NPFs that collaborated with a LHD and those that did

Frontiers in Public Health | www.frontiersin.org 5 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

TABLE 2 | Descriptive statistics: non-profit hospital and county- and state-level characteristics.

All sample hospitals Hospitals collaborated

with LHD

Hospitals did not

collaborate with LHD

Comparisona

n = 785 n = 265 n = 520 p-value

COMMUNITY BENEFIT SPENDING (% OF OPERATING EXPENSES)

Total Community Benefit, mean (SD) 8.96 (4.8) 8.56 (3.8) 9.16 (5.3) 0.12

Population Health, mean (SD) 0.66 (0.9) 0.62 (0.8) 0.67 (0.9) 0.51

HOSPITAL CHNA AND IMPLEMENTATION STRATEGY CHARACTERISTICS

Total Priorities Addressedb, mean (SD) 4.3 (2.2) 4.39 (2.3) 4.23 (2.1) 0.42

Hospital-LHD Collaborationc, % 33.8

HOSPITAL ORGANIZATIONAL CHARACTERISTICS

No. of beds, mean (SD) 174.4 (127.8) 178.32 (128.2) 172.51 (127.2) 0.56

No. of psychiatric beds, mean (SD) 9.27 (17.6) 10.60 (18.9) 8.59 (16.9) 0.13

System membership, % 73.4 72.2 74.0 0.60

Teaching hospital, % 5.0 6.8 4.0 0.09

Church affiliation, % 24.2 22.3 25.2 0.36

Children’s hospitals, % 1.3 0.8 1.5 0.35

Critical Access Hospital, % 9.8 8.7 10.4 0.45

DSH, % 67.1 67.2 67.1 0.99

HOSPITAL FINANCIAL CHARACTERISTICS

Net Patient Revenues, mean in 1,000s

(SD)

241,658 (287,891) 276,224 (360,572) 224,035 (241,122) 0.02

% Operating Margin, mean (SD) −1.2 (17.8) 0.7 (13.2) −2.1 (19.6) 0.04

% Total Margin, mean (SD) 5.3 (1.4) 6.9 (8.9) 4.4 (16.3) 0.02

LOCAL MARKET CHARACTERISTICSd (%)

Non-metropolitan area, % 18.1 17.4 18.5 0.70

Herfindahl-Hirschman index, mean (SD) 18.6 (14.0) 19.19 (13.74) 18.3 (14.1) 0.41

% Uninsurance rate, mean (SD) 15.3 (4.6) 14.5 (4.3) 15.8 (4.7) <0.01

% Unemployment rate, mean (SD) 7.5 (2.1) 7.3 (1.7) 7.6 (2.2) 0.02

Median income, mean in 1,000s (SD) 53,603 (14,548) 54,420 (14,935) 53,187 (14,344) 0.26

% Race, mean (SD)

Black 8.2 (9.6) 8.4 (8.9) 8.1 (9.9) 0.75

White 81.7 (12.9) 81.0 (13.1) 82.1 (12.8) 0.26

Other 7.1 (7.2) 7.7 (8.1) 6.9 (6.6) 0.13

STATE CHARACTERISTICS (%)

Medicaid expansion in 2014e, % 58.3 53.2 61.0 0.04

State Innovation Model Participationf , % 26.0 25.7 26.2 0.89

Authors’ analysis of data from the IRS, CMS, the Center for Medicare and Medicaid Innovation, the Henry J. Kaiser Family Foundation, the Census Bureau, and non-profit hospital

(NFP) community health needs assessments (CHNA) and implementation strategies. LHD stands for local health department. ap-values from bivariate analyses comparing two groups

of NFPs (collaborated with LHD on CHNA vs. did not collaborate with LHD on CHNA); bTotal priorities targeted in 2013 implementation strategies from top 13 priorities (detail in text); cNFP and LHD produced a single CHNA in 2012-13; dLocal market characteristics are at the county level with the exception of HHI which is based on the hospital referral region; ePercentage of NPFs located in states that expanded Medicaid in 2014; fPercentage of NPFs located in states that participated in Round 1 State Innovation Models.

not collaborate. A higher percentage of NPFs that collaborated with LHDs were teaching hospitals (6.8 vs. 4.0%; p = 0.09). On average, NPFs that collaborated with a LHD performed better financially than their counterparts as can be seen by the hospital financial characteristics. For instance, total margin was 2.5 percentage points higher among NFPs that collaborated with a LHD. Non-profit hospitals that collaborated with LHDs tended to be located in counties with slightly lower uninsurance (14.5 vs. 15.8%; p< 0.01) and unemployment (7.3 vs. 7.6%; p= 0.02) rates. A lower percentage of NFPs that collaborated with their LHDs

were located in states that later expanded Medicaid in 2014. All other characteristics were similar across the two groups.

Tables 3A,B present the results from the bivariate analyses. The factors that had more significant associations with each of the targeted priorities were whether a NPF collaborated with a LHD to produce a single CHNA, and county-level uninsurance rate. Non-profit hospitals that collaborated with a LHD had higher odds of targeting obesity (OR: 1.982; p < 0.01), mental health (OR: 1.442; p < 0.05), substance abuse (OR: 1.437; p < 0.05), and alcohol (OR: 1.841; p < 0.01), but lower odds of

Frontiers in Public Health | www.frontiersin.org 6 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

TABLE 3A | Bivariate analyses: association of targeted priority with hospital and county characteristics.

A

Obesity Access Diabetes Cancer Mental health

n = 785

HOSPITAL CHNA CHARACTERISTICS

Hospital-LHD Collaboration 1.982*** 0.782 0.795 0.797 1.442**

(0.377) (0.128) (0.120) (0.140) (0.219)

HOSPITAL ORGANIZATIONAL CHARACTERISTICS

Number of beds 1.000 1.000 1.000 1.000 1.000

(0.0004) (0.0002) (0.0003) (0.0002) (0.0002)

Number of psychiatric beds 1.010* 1.000 1.006 1.003 1.003

(0.005) (0.004) (0.004) (0.004) (0.004)

System membership 1.040 1.552** 1.085 0.861 0.976

(0.194) (0.268) (0.176) (0.156) (0.158)

Teaching hospital 3.028** 0.814 1.403 1.435 1.284

(1.618) (0.284) (0.468) (0.502) (0.425)

Church affiliation 0.704* 0.881 0.782 0.836 0.944

(0.131) (0.160) (0.131) (0.162) (0.157)

Children’s hospitals 1.326 3.728 0.237* 0.309 0.977

(1.054) (3.941) (0.188) (0.327) (0.622)

Critical Access Hospital 0.540** 0.734 0.549** 0.715 1.422

(0.137) (0.186) (0.136) (0.210) (0.346)

DSH 1.310 1.000 1.608*** 0.881 0.821

(0.226) (0.168) (0.246) (0.151) (0.125)

HOSPITAL FINANCIAL CHARACTERISTICS

Total Margin 1.256 1.284 1.172 0.165** 1.893

(0.693) (0.683) (0.600) (0.123) (1.093)

LOCAL MARKET CHARACTERISTICS

Herfindahl-Hirschman index 1.544 0.807 0.496 0.304* 1.698

(0.939) (0.449) (0.255) (0.191) (0.872)

Non-metropolitan area 0.967 0.520*** 0.861 0.711 0.764

(0.206) (0.100) (0.160) (0.159) (0.142)

Uninsurance rate 0.942*** 1.013 1.045*** 1.003 0.960***

(0.017) (0.018) (0.016) (0.018) (0.015)

Unemployment rate 0.992 0.971 1.025 0.941 0.932**

(0.039) (0.036) (0.036) (0.039) (0.033)

Median income 1.000 1.000 1.000 1.000 1.000***

(5.56−6) (5.61−6) (4.91−6) (5.44−6) (5.03−6)

targeting cardiovascular disease (OR: 0.667; p < 0.01). Non- profit hospitals located in a county with higher uninsurance rates had higher odds of targeting diabetes (OR: 1.045; p < 0.01) and cardiovascular disease (OR: 1.039; p < 0.05), but lower odds of targeting obesity (OR: 0.942; p < 0.01), mental health (OR: 0.960; p < 0.01), substance abuse (OR: 0.969; p < 0.10), and alcohol (OR: 0.914; p < 0.01). These patterns are almost exactly the inverse of one another (e.g., higher odds of targeting obesity for NPFs that collaborated vs. lower odds of targeting obesity for NFPs located in counties with a higher uninsurance rate). The number of psychiatric beds was not significantly associated with targeting mental health or substance abuse (illicit, prescription, alcohol, and tobacco). Non-profit hospitals located in a county with higher unemployment rates had lower odds of targeting

mental health (OR: 0.932; p < 0.05), substance abuse (OR: 0.936; p < 0.10), and alcohol (OR: 0.893; p < 0.05).

DISCUSSION

Our study examined a sample of 785 NPFs CHNAs and implementation strategies from the first round post-ACA. To date, this is the largest sample of such documents to be examined. In fact, this is the first study to examine a large national sample of implementation strategies after the ACA and to describe the community priorities actually targeted through hospital interventions. Several studies have contributed to our understanding of the process used by NFPs for community needs assessment and prioritization,

Frontiers in Public Health | www.frontiersin.org 7 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

TABLE 3B |

B

Cardiovascular disease Tobacco Substance abuse Alcohol

n = 785

HOSPITAL CHNA CHARACTERISTICS

Hospital-LHD Collaboration 0.667*** 1.017 1.437** 1.841***

(0.105) (0.158) (0.226) (0.353)

HOSPITAL ORGANIZATIONAL CHARACTERISTICS

Number of beds 1.000 1.000 1.000 1.000

(0.0003) (0.0002) (0.0002) (0.0005)

Number of psychiatric beds 1.003 1.009** 1.006 1.003

(0.004) (0.004) (0.004) (0.005)

System membership 1.273 1.075 1.069 0.881

(0.214) (0.179) (0.183) (0.185)

Teaching hospital 1.346 1.535 1.515 1.240

(0.444) (0.505) (0.503) (0.506)

Church affiliation 0.962 0.829 0.731* 1.104

(0.165) (0.144) (0.132) (0.239)

Children’s hospitals – – 0.211 0.527

– – (0.223) (0.557)

Critical Access Hospital 0.879 1.368 1.260 1.402

(0.219) (0.330) (0.311) (0.410)

DSH 1.340* 0.966 0.905 0.843

(0.212) (0.151) (0.144) (0.166)

HOSPITAL FINANCIAL CHARACTERISTICS

Total Margin 0.192** 0.568 0.673 0.407

(0.136) (0.310) (0.354) (0.243)

LOCAL MARKET CHARACTERISTICS

Herfindahl-Hirschman index 0.457 1.441 0.411 0.780

(0.246) (0.750) (0.232) (0.537)

Non-metropolitan area 1.259 1.488** 1.142 0.964

(0.236) (0.278) (0.221) (0.238)

Uninsurance rate 1.039** 0.989 0.969* 0.914***

(0.017) (0.016) (0.016) (0.021)

Unemployment rate 0.998 1.044 0.936* 0.893**

(0.035) (0.037) (0.036) (0.045)

Median income 1.000 1.000*** 1.000 1.000***

(5.11−6) (5.52−6) (5.10−6) (5.99−6)

Authors’ analysis of data from the Centers for Medicare and Medicaid Services, the Census Bureau, and non-profit hospital (NFP) community health needs assessments (CHNA) and

implementation strategies. LHD stands for local health department. Results are reported as odds ratio. ***p < 0.01, **p < 0.05, *p < 0.1.

as well as the community issues most often identified in CHNAs (14–20). Our contribution was 3-fold; first, we examined a large sample of implementation strategies to extend on previous work that examined CHNAs only. This gives a more complete picture of how NFPs move from identifying all community issues to actual targeted priorities. Second, we also presented information on the status of NPF collaboration with LHDs to produce a single CHNA in the first round after the ACA, which hasn’t been recorded in previous studies. Third, we examined the association between targeted priorities with NFP organizational characteristics and county-level factors.

We uncovered interesting findings, especially when contrasted with the other two larger national studies on NPF CHNAs. We found that most of the health priorities identified in the CHNAs were also targeted with concrete interventions in the implementation strategies. The ranking of these priorities was strikingly similar, especially related to community health issues. Obesity, access to care, diabetes, mental health, and substance abuse ranked in the top 5 for all studies, including our study. Non-profit hospitals have a high level of discretion when selecting priorities from the CHNA to target in their implementation strategies. The findings reported here may be indication that NFP community benefit work reflects

Frontiers in Public Health | www.frontiersin.org 8 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

community priorities as opposed to a stronger focus on strategic organizational priorities which may not necessarily align with community needs. One of the goals of the ACA requirement for a CHNA was to engage NFPs with the communities they serve and to help them gain a more in-depth understanding of community needs. This improved understanding would then facilitate more targeted NFP financial and human capital investment on specific community issues, which hopefully can lead to improved population health. It is promising that there is an alignment between the top priorities identified in the CHNA and those targeted in the implementation strategies.

We also found that∼34% of NPFs in our sample collaborated with their LHD to produce a single CHNA. Collaboration between NFPs and LHDs in conducting CHNA can avoid wasteful duplication of efforts and resources, especially in the context of LHDs seeking to be accredited by the Public Health Accreditation Board (PHAB). Prior to applying for accreditation, LHDs have a set of prerequisites that must be met, including: community health assessment, community health improvement plan, and a department strategic plan (22). The first two are equivalent to the requirement for NFPs to conduct a CHNA and develop an implementation strategy. According to National Association of County and City Health Officials (NACCHO), in 2016, 78% of LHDs had completed a community health assessment and 67% had completed a CHIP (23). This presents an unprecedented opportunity to engage NFPs and LHDs in meaningful collaboration in local health planning.

Non-profit hospital collaboration with LHDs holds the potential for more efficient and effective allocation of resources, and perhaps greater motivation for non-profit hospitals to financially invest in population health. We found some evidence of collaboration in local health planning by NFPs and LHDs; however, there is still much unrealized potential as many jurisdictions across the United States have yet to engage in this type of collaboration.

Some states have aligned their policies with Section 9007 to encourage collaboration between NFPs and LHDs in local health planning. For instance, the New York Prevention Agenda requires NFPs and LHDs to collaborate in local health planning, and has recently aligned the CHNA cycles for both institutions to be on a 3-year schedule (24, 25). Other state policies that are moving in a similar direction include Maryland’s Local Health Improvement Coalitions, Maine’s Shared Community Health Needs Assessment, and North Carolina’s Community Health Improvement Collaborative (26–28). Ohio recently mandated all its non-profit hospitals to collaborate with their LHDs on CHNA and community health improvement plans by 2020 (29).

The state policies mentioned above reflect a common belief that collaboration between LHDs and NFPs may be especially important in improving community health and population health investment by non-profit hospitals. Based on NACCHO’s Profile Studies, LHD collaboration with hospitals decreased by about 22 percentage points from 2008 to 2016 (23). These findings indicate that a requirement may need to be in place for LHDs and hospitals to work together.

We also examined the association of NFP and county-level factors with the targeted priorities. The two factors that showed a stronger pattern of association were NFP-LHD collaboration and county uninsurance rate. Non-profit hospitals that collaborated with a LHD had higher odds of targeting needs related to behavioral health (i.e., mental health, substance abuse, and alcohol) and obesity. County uninsurance rate showed an inverse pattern than collaboration. One explanation could be that addressing substance abuse and alcohol rely more heavily on resident insurance status. In other words, community resources (e.g., treatment, therapy, rehabilitation) are less likely to be available when there are higher rates of uninsurance (i.e., because of a lack of reimbursement for services). Consequently, NFPs may decide that it would take a substantial investment from their part to make a difference in those areas and may choose to invest on a different community issue. This rationale is further supported by the findings related to unemployment rate which followed a very similar pattern as uninsurance rate. Unemployment is closely related to uninsurance because most insured individuals obtain it through their employers. Furthermore, unemployed individuals do not have the means to afford behavioral health treatment. The lack of reimbursement (via insurance or directly purchased by residents) for behavioral health services may lead NFPs to determine that this particular community issue (i.e., behavioral health) is outside their means to reasonably address. This is one way to explain the results observed in our study, but we need to be cautious as these are cross-sectional bivariate regression analyses which are not reliable for causal interpretation.

Some areas for future research emerged from our study and we highlight a few here. The first one is to further investigate the organizational process of selecting priorities to be targeted from the list of several priorities identified in the CHNA. It would be interesting to better understand whether NFPs are targeting priorities that truly reflect community needs, if they give preference to community issues that align with their strategic planning and financial goals, or a combination of both. The second area is related to NFP collaboration with LHDs in local health planning. There are several interesting research questions related to this area. For instance, do NPFs invest more or less on population health when they collaborate with LHDs? Is NFP-LHD collaboration in local health planning associated with improved community health outcomes? As more states align their policies with Section 9007 to encourage NFP-LHD collaboration, we will have the ability to design rigorous studies to examine these and other questions, and to provide the evidence needed to sustain collaborative local health planning efforts. Finally, it will be key to understand the types of interventions being implemented by NFPs to address community issues. One approach would be to place interventions on the spectrum of down, mid, and upstream factors using a social determinants of health framework. This will give us an understanding of whether NFPs continue to focus most of their efforts on the downstream factors (e.g., provision of acute care services) or if some are also addressing the mid and upstream factors (e.g., investment in housing capital projects).

Frontiers in Public Health | www.frontiersin.org 9 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

LIMITATIONS

Non-profit hospital reporting on CHNAs and implementation strategies is not standardized and NFPs may sometimes use a different approach for grouping health priorities. For instance, some NPFs group all substances under the umbrella priority of “substance abuse” which often can include illicit and prescription drugs, alcohol, and tobacco. Sometimes, substance abuse may be nested under “mental health”. Another health priority that seems to vary widely in terms of what community needs are covered is the ubiquitous “access to care” which may cover insurance coverage, primary care, prescription drug costs and other needs. As a result, previous coding frameworks have differed especially for the priorities that fall under “drivers” (e.g., access to care), which is why we collapsed some drivers under “access to care” using the County Health Rankings framework (described earlier).

As previously described, in some cases we had to use 2015/16 CHNA reports to identify NFP’s 2013 targeted priorities because the 2012 CHNAs were no longer available. While NFPs are required to report their progress on previously targeted priorities in subsequent CHNAs, we can’t ascertain whether it includes information on all targeted priorities as listed in the previous implementation strategy.

Finally, the bivariate analyses are exploratory and do not aim to establish causality. In fact, there is a high likelihood of reverse causality. For example, NFPs may seek to collaborate with LHDs to implement interventions to address obesity, but they would have selected obesity as a target community need regardless of having collaborated with a LHD. As such, bivariate analysis results must be interpreted with caution.

CONCLUSION

The community benefit requirement and Section 9007 of the ACA present an opportunity to nudge NFPs to make larger investments in population health and to improve the conditions for health in the communities they serve. Population health has received a renewed focus since the passage of the ACA and its many provisions for health delivery and payment reforms that seek to move our health care system from a volume-based to a value-based one. The ACA has also challenged institutions in the health care sector to approach health through the social determinants of health framework. This framework moves beyond the provision of acute health services and emphasizes other inputs that improve population health (e.g., education, secure and safe housing, employment, etc.). In this context, NFPs are particularly well-positioned to shift their contribution to improve population health beyond their four walls. Section 9007 is one mechanism to achieve such shift and has shown some promising changes among NFPs since its passage.

DATA AVAILABILITY STATEMENT

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHOR CONTRIBUTIONS

TS was the sole author of this study and manuscript.

FUNDING

This study was partially funded by an Agency for Healthcare Research and Quality (AHRQ R01HS24959-01).

REFERENCES

1. Rosenbaum S, Byrnes M, Hurt N. Community Benefit and the ACA: A Brief

History and Update. Washington, DC (2015).

2. Rosenbaum S, Kindig DA, Bao J, Byrnes MK, O’Laughlin C. The

value of the nonprofit hospital tax exemption was $24.6 Billion

In 2011. Health Affairs. (2015) 34:1225–33. doi: 10.1377/hlthaff.20

14.1424

3. Young GJ, Chou CH, Alexander J, Lee SY, Raver E. Provision of community

benefits by tax-exempt U.S. Hospitals. N Engl J Med. (2013) 368:1519–27.

doi: 10.1056/NEJMsa1210239

4. Young GJ, Flaherty S, Zepeda ED, Singh SR, Rosen CG. Community benefit

spending by tax-exempt hospitals changed little after ACA. Health Affairs.

(2018) 37:121–4. doi: 10.1377/hlthaff.2017.1028

5. Singh SR, Young GJ, Daniel Lee SY, Song PH, Alexander JA. Analysis of

hospital community benefit expenditures’ alignment with community health

needs: evidence from a national investigation of Tax-exempt hospitals. Am J

Publ Health. (2015) 105:914–21. doi: 10.2105/AJPH.2014.302436

6. Seay JD. Community benefit prevails. Are radical changes in hospital tax-

exemption laws necessary? Health Prog. (1992) 73:42–7.

7. Leider JP, Tung GJ, Lindrooth RC, Johnson EK, Hardy R, Castrucci

BC. Establishing a baseline: community benefit spending by not-for-profit

hospitals prior to implementation of the affordable care act. J Publ Health

Manage Pract. (2017) 23:e1–9. doi: 10.1097/PHH.0000000000000493

8. Nicholson S, Pauly MV, Burns LR, Baumritter A, Asch DA. Measuring

community benefits provided by for-profit and nonprofit hospitals. Health

Affairs. (2000) 19:168–77. doi: 10.1377/hlthaff.19.6.168

9. McGovern JJ. In the line of fire: the battle over tax-exempt status continues –

revise or preserve? Health Prog. (1992) 37–41:64.

10. Singh SR, Young GJ. Tax-exempt hospitals’ investments in community

health and local public health spending: patterns and relationships.

Health Serv Res. (2017) 52:2378–96. doi: 10.1111/1475-6773.

12739

11. Internal Revenue Service. Additional Requirements for Charitable Hospitals;

Community Health Needs Assessments for Charitable Hospitals; Requirement

of a Section 4959 Excise Tax Return and Time for Filing the Return. (2014).

Available online at: https://www.federalregister.gov/documents/2014/12/31/

2014-30525/additional-requirements-for-charitable-hospitals-community-

health-needs-assessments-for-charitable (accessed March 23, 2020)

12. Internal Revenue Service. New Requirements for 501(c)(3) Hospitals Under

the Affordable Care Act. (2018). Available online at: https://www.irs.gov/

charities-non-profits/charitable-organizations/new-requirements-for-

501c3-hospitals-under-the-affordable-care-act (accessed March 23, 2020)

13. Internal Revenue Service. Instructions for Schedule H (Form 990) –

2017. (2017). Available online at: https://www.irs.gov/pub/irs-pdf/i990sh.pdf

(accessed March 23, 2020)

14. Health Research and Educational Trust.Hospital-Based Strategies for Creating

a Culture of Health. Chicago, IL (2014).

Frontiers in Public Health | www.frontiersin.org 10 May 2020 | Volume 8 | Article 124

Santos Non-profit Hospital Community Health Prioritization

15. Rosenbaum S, Byrnes M, Rothernberg S, Gunsalus R. Improving Community

Health Through Hospital Community Benefit Spending: Charting a Path to

Reform. Washington, DC (2016).

16. Fischer KR, Schwimmer H, Purtle J, Roman D, Cosgrove S, Current JJ,

et al. A content analysis of hospitals’ community health needs assessments

in the most violent U.S. cities. J Commun Health. (2018) 43:259–62.

doi: 10.1007/s10900-017-0413-9

17. Powell RE, Doty AMB, Rising KL, Karp DN, Baehr A, Carr BG. A content

analysis of nonprofit hospital community health needs assessments and

community benefit implementation strategies in Philadelphia. J Publ Health

Manage Pract. (2018) 24:326–34. doi: 10.1097/PHH.0000000000000621

18. Beatty KE, Wilson KD, Ciecior A, Stringer L. Collaboration among missouri

nonprofit hospitals and local health departments: content analysis of

community health needs assessments. Am J Publ Health. (2015) 105:S337–44.

doi: 10.2105/AJPH.2014.302488

19. Pennel CL, McLeroy KR, Burdine JN, Matarrita-Cascante D, Wang J. A

mixed-methods approach to understanding community participation in

community health needs assessments. J Publ Health Manage Pract. (2017)

23:112–21. doi: 10.1097/PHH.0000000000000362

20. Pennel CL, McLeroy KR, Burdine JN, Matarrita-Cascante D. Nonprofit

hospitals’ approach to community health needs assessment. Am J Publ Health.

(2015) 105:103–13. doi: 10.2105/AJPH.2014.302286

21. Hood CM, Gennuso KP, Swain GR, Catlin BB. County health rankings:

relationships between determinant factors and health outcomes. Am J Prev

Med. (2016) 50:129–35. doi: 10.1016/j.amepre.2015.08.024

22. Public Health Accreditation Board. Guide to National Public Health

Department Accreditation. Alexandria, VA (2015).

23. National Association of County and City Health Officials. 2016 National

Profile of Local Health Departments. (2016). Available online at: http://

nacchoprofilestudy.org/wp-content/uploads/2017/10/ProfileReport_

Aug2017_final.pdf

24. NYDOH. New York State Prevention Agenda. (2018). Available online

at: https://www.health.ny.gov/prevention/prevention_agenda/2013-2017

(accessed March 23, 2020)

25. Live!, Public Health. Prevention Agenda 2013: A Blueprint for Community

Action to Improve Health in New York State. (2013). Available online at:

https://vimeo.com/63173195 (accessed March 23, 2020)

26. Maine Department of Health and Human Services.Maine Shared Community

Health Needs Assessment. (2019). Available online at: https://www.maine.gov/

dhhs/mecdc/phdata/MaineCHNA/ (accessed March 23, 2020)

27. Maryland Department of Health. Local Health Improvement Coalitions.

(2019). https://pophealth.health.maryland.gov/Pages/LHIC.aspx (accessed

March 23, 2020)

28. University of North Carolina Gillings School of Global Public Health.

NC Community Health Improvement Collaborative. (2019). Available online

at: https://sph.unc.edu/nciph/nciph-home/nciph-ph-hosp-collab/ (accessed

March 23, 2020)

29. Ohio Revised Code – Title [37] XXXVII Health – Safety – Morals – Chapter 3701:

Department Of Health – 3701.981 Completion and Submission of Assessments

and Plans. (2019). Available online at: http://codes.ohio.gov/orc/3701.981

(accessed March 23, 2020)

Conflict of Interest: The author declares that the research was conducted in the

absence of any commercial or financial relationships that could be construed as a

potential conflict of interest.

Copyright © 2020 Santos. This is an open-access article distributed under the terms

of the Creative Commons Attribution License (CC BY). The use, distribution or

reproduction in other forums is permitted, provided the original author(s) and the

copyright owner(s) are credited and that the original publication in this journal

is cited, in accordance with accepted academic practice. No use, distribution or

reproduction is permitted which does not comply with these terms.

Frontiers in Public Health | www.frontiersin.org 11 May 2020 | Volume 8 | Article 124

Why Most Private Hospitals Are Nonprofit

Author(s): Carson W. Bays

Source: Journal of Policy Analysis and Management , Spring, 1983, Vol. 2, No. 3 (Spring, 1983), pp. 366-385

Published by: Wiley on behalf of Association for Public Policy Analysis and Management

Stable URL: https://www.jstor.org/stable/3324447

REFERENCES Linked references are available on JSTOR for this article: https://www.jstor.org/stable/3324447?seq=1&cid=pdf- reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

and Wiley are collaborating with JSTOR to digitize, preserve and extend access to Journal of Policy Analysis and Management

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Carson W. Bays Nonprofit

In recent decades, restrictions that have been imposed on hospitals organized for profit have served to restructure the industry, generating a constant trend toward nonprofit organizations. Small proprietary hospitals in particular have disappeared while corporate chains have come to dominate what is left of the for-profit hospital sector. The trend toward nonprofit hospitals is

not explained by the failure of the health service markets and is not Abstract the result of a desire to serve the public interest more effectively.

Although a number of arguments have been advanced to explain the shift, the hypothesis that seems most consistent with the existing evidence is that the nonprofit form of organization serves most effectively to strengthen the restrictive character of the market for physicians' services and thereby to serve the individual economic interests of the physicians.

Of over 7000 hospitals registered with the American Hospital Association in 1977, only about one in ten was organized on a for-profit basis. For-profit hospitals once dominated the private sector of the hospital industry, but their absolute and relative numbers have fallen since the early 1900s. As Table 1 shows, for-profit hospitals accounted for approximately 56% of all regis- tered hospitals in the United States in 1910' but now they consti- tute about 11% of all hospitals and only 5% of all hospital beds. Within the shrinking for-profit sector, another major shift has been taking place, especially notable within the past few decades. The number of traditional "proprietary" hospitals-small organiza- tions owned by one or a handful of doctors-has been reduced drastically while corporate chain profit hospitals, which are typi- cally larger more diversified institutions, have increased substan- tially both in number and in total beds.

I wish to thank, but not implicate, William A. Glaser and three anonymous referees for comments on earlier versions of this article.

Journal of Policy Analysis and Management, Vol. 2, No. 3, 366-385 (1983) ? 1983 by the Association for Public Policy Analysis and Management Published by John Wiley & Sons, Inc. CCC0276-8739/83/020366-20$03.00

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

Table 1. Trends in hospital ownership.

Nongovernment Total for-profit Corporate chain Total nonprofit hospitals hospitals for-profit hospitals

number of

hospitals of Beds Beds Beds Year all types Number (thousands) Number (thousands) Number (thousands)

1910 4359 2441 1928 6852 2435 1946 6125 2584 301 1076 39 1955 6956 3097 389 1020 37 1966 7160 3440 533 852 48 1967 7203 3461 550 821 47 1968 7137 3430 566 769 48 1969 7215 3263 571 748 50 567 40 1970 7123 3386 592 769 53 1971 7165 3207 596 733 55 602 46 1972 7061 3326 617 738 57 1973 7123 3320 629 757 63 1974 7250 3241 641 763 71 666 63 1975 7156 3364 659 775 73 1976 7082 3368 671 752 76 1977 7176 3230 669 737 81 661 72

Sources: (1) Steinwald, Bruce, and Neuhauser, Duncan, "The Role of the Proprietary Hospital," Law and Contemporary Problems, 35(2) (Autumn 1970): 819; (2) American Hospital Association, Hospital Statistics (Chicago: AHA, 1978), pp. 3-4; (3) 1969, 1971, 1974, and 1977 national survey tapes of the American Hospital Association.

Several factors have had direct negative impacts on for-profit hospitals over the period of their decline: Total capital costs have been higher because the for-profit hospitals have not benefited from the private philanthropy and government subsidies that have been provided to the nonprofits; labor costs have been higher because of the preferential treatment of the nonprofits in the labor laws; the for-profits have paid taxes to which nonprofits have not been subject; they have not enjoyed the preferential legal status which until recently made nonprofits immune to most lawsuits; and they have been burdened more heavily by the direct govern- ment regulation of hospital growth.2 All of these factors suggest a distinct preference for the nonprofit form of hospital organization. But what are the sources of this preference? Do they arise from some innate characteristics of hospital services or from other factors?

THE SPECIAL

CHARACTERISTICS OF HOSPITAL SERVICES

It is widely recognized that hospitals have special attributes which distinguish them from other economic organizations. These dif- ferences can be summarized into three broad categories: exter- nalities, conditions of information and uncertainty, and issues of distributional equity.

367

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

Externalities As economists have long recognized, there are certain situations in which the market fails to generate the appropriate level of services because it does not measure properly the benefits or costs of those services to society. Such externalities may be important for health care in general and hospitals in particular. Most public health measures, for example, involve services which either would not be supplied at all by the private market or would be grossly un- dersupplied by profit sellers. Some consumers would be willing to pay the full cost of innoculating themselves against a contagious disease but the total value to society of an innoculation program exceeds the sum of these private demands because persons not buying immunity receive indirect benefits from those who do. Accordingly, most mass innoculation programs provide immuni- zation at either no direct charge to consumers or at a price much below cost.

An externality which may have some implications for the structure of hospitals is that which has been described as the economics of sharing.3 For altruistic reasons, some persons may be willing to contribute to providing health services for others so that the distribution of such services will not be determined solely by their income. If society were to rely entirely upon voluntary contributions to change the shares of medical consumption, how- ever, a "free rider" problem would arise. A problem of this sort occurs in its pure form if society tries to provide for its military defense solely through private citizens buying their protection from private firms. Those who refused to buy the service, the free riders, would still share some of its benefits. In a similar manner, the total amount of giving occurring under a health system supported by purely voluntary donations will be less than the socially optimal amount.

The traditional policy for dealing with this type of externality has been to compel redistribution through taxation and subsidies; that is why military defense is normally provided by governments. With respect to hospitals, there are alternative ways in which the redistribution could be achieved. Hospital care could be produced solely in government hospitals financed by general tax revenues; the redistribution rationale, for instance, has been used in justify- ing the British National Health Service. Or hospital care could be produced in private institutions-either nonprofit or for-profit- which could be subsidized through government grants and helped by tax provisions and other legal advantages.

Information and In addition to the problem of externalities, markets can prove Uncertainty inefficient because of the cost and availability of information to

consumers. The technical complexity of modern medical science raises the possibility that consumers may be unable to make rational choices regarding health problems because they are in- competent to evaluate the costs and benefits of various treatment alternatives. But it is not the technical complexity per se that differentiates medical care from other markets. Many consumer markets involve products of extraordinary technical complexity

368

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 369

about which consumers may be ignorant. In such cases, as many economists would insist, independent sellers-who presumably are competent to judge the relative merits of different brands of a complex good-will have an economic interest in providing con- sumers with accurate information regarding such products.

However, acquiring information for medical care differs in various important respects from acquiring information in typical consumer markets. To begin with, to some extent the product being sought is information itself, such as diagnosis of symptoms and a prescription of treatment; in such cases, the information can only be evaluated after it has been purchased. In addition, because one individual, the physician, supplies both the medical informa- tion and the medical services, a conflict of interest may arise; the physician could lean in the direction of providing diagnoses that generated expensive treatment. To be sure, the conflict would conceivably be limited in a number of ways. Patients after all can seek second opinions. Moreover, most physicians would presuma- bly be sufficiently influenced by the spirit of the Hippocratic oath so that the issue of conflict of interest would not arise. Finally, one might hope that placing hospitals on a nonprofit basis would reduce the risk of such potential conflicts.

Unfortunately, however, there is little evidence to support the view that the nonprofit form reduces the conflict problem. To be sure, it is frequently suggested that for-profit hospitals owned by their admitting physicians have an incentive to overtreat patients as compared with nonprofit hospitals with independent physi- cians4; but this proposition has not been tested adequately.5 Moreover, the contention that for-profit hospitals are more lucra- tive for physicians who practice in them than nonprofits is demon- strably false on a theoretical level.6 Furthermore, as was noted earlier, most for-profit hospitals today are not of the historical "proprietary" form which are run by physician-owners but are parts of corporate chains in which the participating physicians may have little or no ownership interest. The special conditions relating to information in the medical market therefore appear to provide a weak basis with which to explain the prevalence of the nonprofit form among private hospitals.

Apart from information, there is also the problem of uncertainty as it relates to medical care. Uncertainty is present in medical care both in choosing the treatment and in the patient's response. Sociologists and economists have suggested that the nonprofit status of hospitals may be a way of providing protection to patients against this uncertainty by insuring that the special trust relation that inures to the physician-patient relationship is not contaminated by profit-making.7

This rationale for the nonprofit status of hospital industry, however, does not explain the several historical changes in the relative sizes of the nonprofit and for-profit hospital sectors. For centuries, hospitals in Europe were almost the exclusive domain of churches and religious organizations. "Voluntary" nonprofit hos- pitals arose in the Protestant countries, and in Catholic countries

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

that had secularized the religious hospitals.8 Early hospitals in the United States were nonprofit as illustrated by the Pennsylvania Hospital established in Philadelphia in 1751, and the Presbyterian Hospital in New York in 1868. Local government also provided hospitals for the poor, exemplified by the Boston City Hospital established in 1858.9

For-profit hospitals did not rise to significance until the latter part of the 19th century. And as noted earlier, they lost their relative position to nonprofits in the succeeding century. The chain corporate hospitals that have appeared in the for-profit sector since the 1960s roughly correspond in timing to the introduction of widespread government health insurance. With insurance, the physician was in a position to prescribe care without concern for the patient's ability to pay. That development, it could be argued, gave the for-profit sector a new impulse for growth because the presence of more complete insurance coverage reduced the need for nonprofit status to protect the trust relation between doctor and patient.

This hypothesis, however, fails to explain other periods of change in relative sizes of the nonprofit and for-profit sectors. For example, hospital insurance offered by commercial and Blue Cross insurance plans has been expanding since approximately the end of World War II. Yet the slight resurgence in the for-profit sector has not been apparent until the 1960s. Moreover, it is not clear why the resurgence has occurred only among chain corporate hospitals: For-profit hospitals organized as proprietorships and partnerships have continued their historical decline. Therefore, we must look further for explanations of the factors that have been shifting the structure of the hospital service industry.

Distributional Equity Perhaps no other aspect of medical care draws more attention than the issue of equity. Few other goods have characteristics that make their denial life-threatening. Food and shelter are certainly "neces- sities" in this sense, but only limited amounts of each are truly necessary for survival. A catastrophic disease or accident, how- ever, may require resources that could easily exhaust the resources of even the well-to-do. Yet our social norms demand that critical medical treatment should not be denied to those who cannot pay the cost.

The question to be asked in this context is whether equity considerations such as these have been responsible for the domi- nance of nonprofit organizations among private hospitals. The answer is unequivocally no.

The fact is that nonprofit private hospitals provide relatively little "free" care. For the period 1962-1966, the most recent years for which national data are publicly available, free care provided by nonprofit hospitals was equal to about 3% of their total patient revenue. This was actually proportionately less than the amount of charity provided by for-profit hospitals. The latter provided char- ity equal to slightly over 4% of their total revenue.10

A more complicated issue is whether nonprofit hospitals dis-

370

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 371

criminate in the prices that they charge for different services, by profiting from some and losing from others-and whether that discrimination has implications for the support of the poor. The practice of cross subsidization among services within hospitals is widespreadl and has been the subject of several previous pa- pers.12 The most sophisticated of these-that by M.I.T. economist Jeffrey Harris-is one of the few to conclude that cross subsidiza- tion does yield an improvement in social welfare. Harris argues that substantial welfare gains are possible through a hospital's internal price discrimination and concludes that "… this form of discriminatory pricing can be completely consistent with non- profit objectives."13

Harris' analysis, however, does not explore the question of whether cross subsidization by nonprofit hospitals is equitable in another sense. Where public hospitals are used to deliver health services, voters in effect are taxing themselves to provide charity. When a private hospital provides for the poor through price discrimination, however, the individuals being "taxed" have no opportunity to vote on the direction or degree of redistribution. Not only is the outcome inequitable, it also has been demonstrated in theory to be a less than optimal solution for the free-rider problem.14

Harris' analysis also ignores the economic consequences of the role of admitting physicians. These are doctors who admit their private patients to the hospital and then control the allocation of the hospital's resources on behalf of their patients. Health care costs depend not only upon the pricing of hospital services but upon the pricing of physician services as well. Because Harris deals only with the hospital component of cost, his conclusions regarding possible welfare gains from discriminatory pricing are unpersuasive. Moreover, his conclusion that the cross subsidiza- tion is the result of conscious attempts by hospital administrators to redistribute income and spread risks among patients has neither theoretical nor empirical support. Without an explicit model of nonprofit hospital behavior, it is unclear why the administrator would behave in this manner.15 Furthermore, given the influence which admitting physicians have in hospital operations, it is unlikely that the administrator could behave in such a manner if it were inconsistent with the private interests of the admitting physicians. It is critical, therefore, to determine where the inter- ests of the admitting physician lie.

To an important extent, hospital inputs are complementary to those of admitting physicians in the production of hospital care. Physicians make the decisions regarding admission and treatment but are not employed by the hospital nor charged rental for the use of its facilities and personnel. Instead, the patient, or his insuror, is billed separately by the hospital for the use of its inputs and by the physician for services rendered. This arrangement has frequently been rationalized as a consequence of the agency relation between doctor and patient16: To insure that treatment is motivated by medical need rather than by profitability, the physician is insu-

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

lated from the question of hospital costs, which are billed sepa- rately.

Data collected by Harris for a sample group of hospitals reflect the strong tendency of hospitals to engage in cross subsidization of different services. Table 2 compares the hospitals' prices for selected services with their long-run marginal costs in providing each of the services. Harris explains the cross subsidization as an attempt in part to hold down costs for those who would have difficulty in paying. The hypothesis, however, is unsupported by the data. There is no reason to assume, for instance, that the most heavily subsidized service-surgery-is more commonly provided for the poor than the more profitable services listed on the table.

The obvious alternative explanation is that administrators are aware that the price which the physician can charge in any given case is dependent in part upon the size of the hospital's charges. That alternative explanation is suggested with particular force by the fact that surgery is the activity that is subsidized most heavily. If hospitals charge less than cost for such products as surgery and anesthesia, surgeons are in a position to charge more for their services. This loss can be subsidized by charging more than cost on services that do not have such a direct link to the physician's services, such as rooms and therapy.

An additional factor that complicates the distributional issue is the incentive structure that is created for hospitals because of the way they are reimbursed for their services. When Medicare, Medicaid, and Blue Cross refuse to compensate hospitals for bills that the patients themselves are unable to pay, these charges are loaded onto the bills of other patients; cross subsidization occurs therefore through these channels. Moreover, the higher mark-ups of price over cost typically are encountered in those hospital departments with the greater proportions of Medicare patients.17 Patients in those departments therefore subsidize patients in other departments such as the delivery room. The distributional impact of these subsidies is ambiguous and may even be perverse.

In summary, cross subsidization in hospital pricing is wide-

Table 2. Prices and price-cost margins for selected hospital services.

Markup (or markdown) of Type of service Average price price over cost (percent)

(1) Surgery $193.00 – 44 (2) Intensive care unit 211.00 – 38 (3) Special diagnostic 233.00 – 16 (4) Rooms 63.00 + 49 (5) Diagnostic x-ray 29.00 + 62 (6) Routine therapy 11.00 + 69 (7) Routine diagnostic 2.98 +158 (8) Chest x-ray 12.00 +175

Source: Harris, "Pricing Rules for Hospitals," p. 233.

372

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

spread and has distributional consequences. The conclusion that it reflects attempts by nonprofit hospitals to achieve socially desira- ble income transfers is not supported by available data. The alternative hypothesis-that cross subsidization by hospitals sup- ports the private economic interests of their admitting physicians-seems more plausible and is supported by the avail- able evidence.

NONPROFIT HOSPITALS

AND PRIVATE INTERESTS

The Historical Pattern

None of these considerations explains very well why the nonprofit form of organization has come to predominate among hospitals. But some indication of why that development has occurred can be gleaned from history and from interest group theory.

The hospitals of the 19th century in the United States were largely charity agencies in which private physicians donated their time. Private physicians valued appointments at larger hospitals be- cause a research or teaching arrangement with such a hospital enhanced the doctors' reputation among the private patients who paid for the physicians' services. Individuals who could afford to pay for medical care went to doctors for treatment in their personal offices because hospitals could offer little in the way of additional diagnosis and treatment.18 The hospital's role during this early period, therefore, was basically as a repository for the indigent sick. Physicians were trained, as were other professionals, within the context of the guild system which had been a widely accepted institutional arrangement for maintaining quality and protecting the consumer against the incompetence of the supplier of services.19 It also, of course, had the effect of sharply limiting entry into the profession. In order to become a physician, young men had to indenture themselves to a master doctor for several

years, paying a fee for the opportunity.20 Entry was restricted further by occupational licensing. Between 1799 and 1826 all but three states adopted some type of licensing arrangement for doctors which restricted medical practice to those with training deemed appropriate by the state or county medical boards.2

The development of the germ theory of disease, the introduction of aseptic and antiseptic techniques, the discovery of x rays, and the advent of antibiotics converted the hospital from a medical almshouse into a diagnostic and treatment center. Hospitals began to offer services which attracted paying patients, and private physicians initially viewed this as unfair competition: "As late as 1900, the Journal of the American Medical Association regularly ran editorials on the problem of 'hospital abuse' complaining about patients who could have afforded to pay for medical services, but preferred 'cheating [their] own family physician at home out of his fee' by going to hospitals (Jour. A.M.A. 35:20, p. 1240."22 How- ever, the development of new diagnostic and surgical techniques made hospitals more attractive to physicians. New capital- intensive technology such as x-ray machines were more readily accessible to physicians, and sterile surgical fields and recovery

373

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

374 Why Most Private Hospitals Are Nonprofit

facilities gave hospital-based care a scale advantage over medical care rendered in the doctor's private office.

These quality improvements, combined with rapid population growth, resulted in increases in demand for medical care that exceeded the capacity of the guild system of physician training. Gradually medical training shifted to medical schools, which expanded in number from 5 in 1810 to 160 in 1900.23 Most of this growth was among proprietary medical schools which offered students an alternative to the lengthy apprenticeships of the guild system. The guild control over quality and entry into the profes- sion dissolved as the number of physicians increased. Medical schools took over the role of licensing from the state and county medical boards and by 1864 virtually all of the state licensing laws had been repealed. The reasons for repeal are a continuing con- troversy among medical historians. On the one hand, it has been regarded as resulting from the anti-intellectual and antieduca- tional biases of the Jacksonians24; but it has also been attributed to the public's becoming disgruntled with the inability of state licensing to insure competence.25

Shortly after its founding in 1847, the American Medical Associ- ation began to lobby state legislatures to reinstate licensing for medical doctors. Between 1880 and 1900 the AMA had succeeded in having all states establish standards promulgated by the AMA.26 It also began to establish standards for medical education that eventually resulted in the closing of approximately one-half of the medical schools in existence. The primary vehicle of this change was the 1910 Report on Medical Education in the United States and Canada by Abraham Flexner. Although the study was commis- sioned and published by the Carnegie Foundation for the Ad- vancement of Teaching, it is clear that it was done at the behest of the AMA. The Carnegie Foundation lent an air of prestigious independence to a series of policy changes recommended by the AMA four years before.27 The ensuing restrictions on entry into the profession and the consequent increase in physician income have been well documented.28

Available data indicated that the for-profit hospital sector reached its zenith about 1930, somewhat later than did the proprietary medical schools. The historical record regarding the decline of the for-profit hospitals is sparse, and the interpretations of it are somewhat contradictory. The traditional view has been that for-profit hospitals serve an essentially transitory role in the medical care system. They arise for two reasons. First, they are organized and controlled by physicians who are denied admitting privileges at nonprofit hospitals. Second, they enter areas of rapid population growth where nonprofits are sluggish to respond be- cause of the absence of private philanthropic support. According to this hypothesis, once population stabilizes and personal income reaches a level sufficient to support voluntary hospitals, for-profit hospitals either close or convert to nonprofit status.29

Empirical support for this hypothesis is mixed and deals with only fairly recent periods: Through the mid-1960s, one could

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 375

discern a conversion from for-profit to nonprofit hospitals, but this is not true for later periods.30 Moreover, as hospital chains began to dominate the for-profit sector in the late 1960s, the relative importance of for-profit hospitals has tended to be higher in the areas with slower population growth, a tendency that runs con- trary to the hypothesis.31

Perhaps there are two reasons for the ambiguous results of studies of the growth of for-profit hospitals. First, part of the cause for the decline of for-profit hospitals had nothing to do with ownership status per se but with changes in hospital size brought about by technological change. The medical discoveries and scien- tific innovations that converted the hospital from an almshouse to a diagnostic and treatment center had the combined effects of increasing the minimum efficient size of the hospital and of increasing the attainable quality of care. A large number of smaller hospitals, both for-profit and nonprofit, disappeared as a result. Because for-profit hospitals on average were smaller than non- profits, the effect on the for-profit sector appeared dispropor- tionate. Second, the emergence of for-profit chain hospitals in the late 1960s was not simply a reversal of previous historical trends but reflected instead the financial advantages of the chain form.

A Cartel Model As the hospital has come to assume a critical role in medical care, the withdrawal of hospital admitting privileges has at various times been very effective in punishing physicians guilty of "unpro- fessional" conduct, such as affiliation with prepaid group prac- tice.32

In several models of the economic behavior of hospitals, the relationship with admitting physicians occupies a central fea- ture.33 Although these models vary slightly in their predictions, they each stress the important role which admitting physicians have in both the short- and long-run decisions of the hospital. It should be noted that the economic interest of an individual physician may differ from the interests of physicians as a group.34 For example, an individual physician may be able to increase his income by ordering unnecessarily large amounts of nursing care for his patients. Because the cost of such care is typically included in the hospital's per diem rate paid by all patients, the physician's patient (or insurance carrier) will pay only a portion of the cost of this overutilization. If all physicians overutilize in this manner, however, the resulting inflation in hospital charges will have adverse effects on physicians as a group: The increases in the hospital component of total charges will have a constraining effect on the price which the physician can charge for his services. The tendency of individual members of a cartel to act counter to the interests of the group has long been recognized.35 Efforts to control such tendencies in medical care are to be found in various widespread hospital policies, such as limiting the list of physicians with admitting privileges and instituting internal peer review.

The problem of an individual physician acting contrary to the collective interests of his peers manifests itself, of course, in

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

376 Why Most Private Hospitals Are Nonprofit

contexts other than that of the hospital. For example, price cutting could be an effective way for an individual physician to increase the size of his practice; but to engage in price cutting effectively, the physician must be able to communicate his prices cheaply to potential patients. When advertising is prohibited, information about prices cannot be communicated readily, a fact that reduces the benefit to a physician in charging less than "usual, customary, and reasonable" fees. Indeed, physicians in private practices are encouraged by organized medicine to limit their output, a ten- dency that reduces the pressure on prices. For example, doctors in private practice typically use fewer aides than the number that would maximize their individual profit.36 Standards promulgated and enforced by state medical licensing boards about the kinds of tasks that can be performed by physician's aides such as nurses also limit potential productivity growth among physicians.37 Al- though the AMA has supported federal subsidies to help students meet the cost of their nursing education, it has opposed changes that would allow nurses to become substitutes rather than com- plements to physicians38; these measures reduce the cost of nursing services to the physician without increasing his competition.

The cartel perspective helps to explain the policy stands of organized medicine on hospitals. For example, the AMA strongly supported the federal aid for hospital construction and loan subsidies to nonprofit hospitals. The so-called Hill-Burton pro- gram provided funds and loans guarantees for the expansion of existing hospitals, which lowered the cost of existing admitting physicians' most important complementary input, the hospital; but the program restricted funds for new hospitals to areas with shortages of hospital beds, thus opening up opportunities for new admitting physicians only in the areas where new competition could be absorbed most easily.39 Since the early 1960s the AMA has favored the concept of hospital planning but has insisted that planning agencies-which can sharply limit the growth of new and existing hospitals-must include physicians.4

The cartel model is also a useful hypothesis for explaining historical changes in the relative importance of for-profit hospi- tals. As was observed earlier, the first hospitals in the United States were nonprofit because they were primarily charities. Entry into medicine and professional conduct standards were controlled by the guilds, but there was little need for control over hospitals by organized medicine because they were not yet crucial to the typical doctor's practice. In the late 19th and early 20th centuries, it will be recalled, this situation changed drastically and medical schools and hospitals organized for profit entered in response to the increases in medical care demand engendered by technological change. The guild system was disrupted but was finally reestab- lished under the auspices of the AMA in the early 1900s. For-profit hospitals were tolerated in areas of rapid population growth because the cartel's loss of market share to the interlopers was less noticeable (and less costly) when total market demand was ex- panding. Once population growth stabilized, however, the for-

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 377

profit hospitals either were forced to convert to nonprofit status- that is, to join the cartel-or were forced to close by discriminatory tax treatment, refusal of accreditation, or discrimination by third-party payers, which were dominated by nonprofit interests. According to this interpretation, physicians as a group prefer nonprofit hospitals not because they are allergic to the notion of profit, but because restrictions against for-profit hospitals have been one way of controlling entry. Individual physicians-who are already in the cartel-prefer nonprofit hospitals because favorable tax treatment and private and government subsidies to nonprofits lower the total cost of the complementary hospital inputs and therefore increase the price that physicians can charge.

The cartel model thus implies that the tax treatment and favored legal status of nonprofit hospitals exist in part because they support the economic interests of private physicians. A comprehensive analysis of the political origins of various state and federal laws that discriminate against for-profit hospitals would help test that conclusion. So far, no such analysis has been undertaken. Still, the casual evidence in support of that view is compelling. For instance, hospital planning agencies appear to have a bias against for-profit hospitals, according to anecdotal evidence,41 empirical research,42 and opinion surveys of health planning agency managers.43 In some states-a notable exception is California-Blue Cross originally refused reimbursement to for-profit hospitals or reimbursed them at a lower rate than that for nonprofit hospitals.44 In negotiations between the Social Secu- rity Administration and the American Hospital Association regard- ing the original Medicare legislation, the AHA convinced the SSA to structure the reimbursement formula in such a way as to discriminate against for-profit hospitals.45 A 1948 federal antitrust suit meticulously documented the role of organized medicine in systematically undermining the for-profit hospital associations of Oregon and Washington.46 These were early versions of the health maintenance organization, which combines the role of hospital, physician, and insurance carrier and constitutes a competitive threat to the doctor-hospital cartel.

Evaluating the Cartel There are questions regarding the cartel model, however. For Model example, why have chain for-profit hospitals apparently been

exceptions to the rule that hospitals run for profit are driven out of the market eventually by a cabal of private physicians and non- profit hospitals? Also, if public attitudes reflect a preference for the nonprofit form, the cartel model implies that citizens and policymakers have been deceived into favoring an organizational structure that benefits one group-physicians-without providing any net benefits to society at large.47 How can so many have been misled so long by so few?

These questions are much less perplexing when placed in the context of a general theory of interest group action.48 The general formulation of the theory is that different interest groups, defined as individuals with a common economic interest, compete for

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitls Are Nonprofit

regulations which redistribute wealth from the rest of society to interest group members. The mechanism that links interest group preferences with regulation is the politician (regulator) who arbit- rates among interest groups to serve his own interests. His inter- ests may be to secure a political majority, or to acquire income, or both.49

The success of an interest group in "taxing" the rest of society depends not only on the tradeoffs the regulator must make among the competing interest groups but also upon the organizational constraints common to all interest groups. For example, condi- tions that enhance the power of the group to secure favorable regulation are the size and homogeneity of its membership, the level of income of its members, and the ability to organize at relatively low cost.50 A conclusion of the model is that in regulatory equilibrium an institutional structure will evolve that-whether it involves price fixing, entry restriction, professional and quality standards, direct subsidies, or some combination of these-will be efficient in the narrow sense that the marginal benefit and margi- nal cost of further action are equated for each participant. That is, any change in policy on the part of the regulator would result in a net loss of votes and any further political contribution or organiza- tional action by interest groups would cost more than the addi- tional regulatory benefits received. It follows that the politician will not usually promote the interests of a given group fully because other groups will have competing demands and because each group will face diminishing returns to further organizational and political action.

This broader perspective can be used to explain the historical fluctuations in the importance of for-profit hospitals in the follow- ing way. Until about the middle of the 19th century, medical training in the United States was in a regulatory equilibrium which relied largely on the guild system of training and quality control. The technological improvements that began in the mid- 1800s disrupted this equilibrium by changing the nature of medi- cal practice and increasing the demand for medical care. Eventually a new regulatory equilibrium was approached in which organized medicine restricted entry by control over medical school curricula, licensing standards, and discrimination against for-profit hospitals. Elected officials responsible for the financing of public hospitals also had an interest in eliminating for-profit hospitals because their alleged "skimming" of profitable patients increased the deficits of the public institutions.51 The ability to close for-profit hospitals was enhanced by the quality argument: Small hospitals of all types offer fewer services than larger hospi- tals.52 Closing smaller hospitals, of whatever type, therefore could be justified on the basis of improvements in quality of care. Nevertheless, efforts at closing hospitals have had dispropor- tionate effects on for-profit hospitals of all sizes and appear to have inhibited their expansion to more competitive sizes.53

Such hospitals have survived and prospered in a few states (most of the survivors are located in California, Texas, and Florida)

378

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 379

where long periods of population growth during the late 19th and early 20th centuries begot for-profit sectors that were sufficiently large to counter the influence of competing interest groups. The fact that chains operating for profit have expanded recently while the single for-profit hospitals have continued to decline apparently reflects the greater financial abilities of the chain organization form. The evidence regarding the relative efficiency of the chain hospitals and nonprofits is mixed.54 If the sources of the recent growth among chain hospitals are economies of scale in purchas- ing and administration, then these efficiencies in principle are attainable by nonprofit chains as well. Nonprofit hospitals in fact have begun to combine into chains in recent years but the actual efficiency gains from such consolidations appear to have been quite small.55 It seems likely that the growth of chain hospitals reflects their ability to raise capital through the sale of equity as well as of debt. The total capital costs may be lower for nonprofits because of past donations and subsidies, but nonprofit hospitals have been increasingly compelled to rely on the private capital market since the 1960s because of decreases both in charity and in government construction grants.56 If the cost of the marginal unit of capital is cheaper for the chains because of their superior access to the capital markets, then this could account for their relative growth. It has also been suggested that chain hospitals have exploited the current doctor "surplus" by writing contracts with their pathologists and radiologists that are less costly than those of nonprofit hospitals.57

The view of regulation as a rational outcome of interest group rivalry also helps to explain how small interest groups can success- fully lobby for a regulatory framework that imposes economic losses on the balance of society. A common explanation as to why interest groups prefer indirect subsidies such as price controls or entry restrictions to direct cash grants is that cash grants are much more visible than the subsidies to those who must pay for them.58 However, even if citizens had perfect information regarding the cost and incidence of indirect subsidies, the subsidies may still be tolerated. For example, if we assume that the cumulative effect of the various restrictions inflated the 1979 median physician income of $70,000 by 40%-a conservative estimate59-then the typical consumer's out-of-pocket costs for physician care was increased by less than $20 in that year.60 The per capita cost of organizing a political effort to remove the subsidy would be greater than the benefit to any individual involved. Nor does the existence of a regulatory structure that benefits an interest group necessarily mean that the group in the past had the prescience to contrive that particular structure. Specific regulatory policies can be motivated by reasons other than to increase income of a professional group, but the economic impact of those policies on different groups plays an important role in determining the regulatory structure that evolves. The Flexner report discussed earlier, for example, was not intended (at least by the Carnegie Foundation) as a blueprint for raising physician income. It was motivated by the abysmal quality

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

of training provided by some of the medical schools of the day. Nevertheless, its main policy proposal-to limit the number of medical schools and graduates-served the long-run economic interest of physicians and therefore helped to create a powerful and cohesive lobby that continues to rationalize entry restrictions in the name of quality care.

A complete test of the interest group explanation for the fluctua- tions in ownership structure of U.S. hospitals would require a good deal more empirical work. Perhaps the primary value of the explanation is in casting light on the course of present and future health policy. Legal scholars now recognize the cartel aspects of organized medicine,61 and serious legal questions have been raised regarding the reduced tax liability of nonprofit hospitals.62 More- over, spokesmen for organized medicine continue to espouse policies that can be evaluated within the context of the interest group model. For example, a 1980 policy statement of the Mas- sachusetts Medical Society decried the growth in the number of private physicians who have direct financial interests not only in hospitals but also in nursing homes, diagnostic laboratories, dialysis units, and other for-profit corporations providing health care services or supplies.63 It urged organized medicine "… to act decisively in separating physicians from the commercial exploita- tion of health care."64

The interest group model would predict that the eventual effect of any formal attempt on the part of organized medicine to "separate" doctors from commercial interests would be limited largely to for-profit hospitals, health maintenance organizations, and perhaps nursing homes, because these pose the greatest competitive threat to the doctor-hospital cartel. Physicians' interests in medical supply organizations that are not directly competitive with the cartel would be expected to escape any effective sanctions. These predictions, by the way, are different from those that would be made if the reason for the dominance of nonprofit hospitals was the doctor-patient trust rationale; the prediction therefore provides a potential way of selecting between the two theories. If maintaining the trust relation between doctor and patient is the true justification for limiting the commercial interests of doctors, then we would expect organized medicine to attempt to enforce prohibitions equally against all potentially conflicting interests.

Another potential test of the two theories is provided by the evolution of the Professional Standards Review Organization. This is a federal program begun in 1972 that established 192 regional review agencies charged with improving the quality of care in hospitals and lowering hospital costs. The mechanism created for this was a formal peer review system in which a panel of physi- cians monitors in-patient treatments, lengths of stays, and charges for all hospitals in its area. Organized medicine was frankly hostile to the PSRO concept during the program's initial years but opposition gradually changed to grudging acceptance and in some cases even support.65 Preliminary tests of the effects of PSRO

380

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit

concluded that the program had beneficial effects on hospital utilization,66 but these studies have been criticized as conceptually flawed.67 In principle the program can economically benefit admit- ting physicians as a group to the extent that it constrains the inflation of hospital costs that is caused by overutilization. In other words, PSRO provides the doctor-hospital cartel with another way of limiting noncooperative behavior of physicians in using their joint hospital inputs.

A useful piece of future research would be to compare the rates of increase in physician income between areas having strong PSROs with areas in which they were late in organizing or ineffectual. The interest group theory predicts that physician incomes will be increased by effective PSROs and allows the inference that the decline in organized medicine's initial hostility to the PSRO has an economic basis. The trust model, on the other hand, is silent on any predicted relationship between PSRO effectiveness and physician income; that model would explain the shift in attitude toward PSROs as a reaction to the improvements in the quality of care brought about by the program.

CONCLUSION The current policy debate regarding government regulation versus free market allocation may be improperly drawn as it applies to the health care industry. The choice is not between government control of allocation and unfettered competition. Instead, it in- volves a choice along a continuum from market allocation through varying degrees of nonmarket control, both public and private. Removing government regulation does not mean a "return to the market." It is debatable whether market control has ever been the primary method of allocation in the health care industry, with the possible exception of the 19th century. The movement toward more market control in health care now has intellectual respecta- bility and political momentum.68 But a comprehensive overhaul of health regulation is an enormous task requiring both executive and legislative changes at the federal and state levels. Various interest groups can be expected to attempt to exploit these policy changes so as to foster a structure congenial to their economic interests. The recent attempt in the Congress to limit the power of the Federal Trade Commission over the professional activities of physicians is a particularly disturbing example. Antitrust action-for all its imperfections-may be required in the health care field as an indispensable counterweight to interest group action.

CARSON W. BAYS is a member of the Department of Sociology, Anthropology and Economics at East Carolina University.

NOTES 1. This percentage is a very crude estimate both because of the sparse data on health institutions of the time and because of ambiguities in the definition of a hospital. Nevertheless, the data are probably the

381

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

382 Why Most Private Hospitals Are Nonprofit

best obtainable. See Steinwald, Bruce, and Neuhauser, Duncan, "The Role of the Proprietary Hospital," Law and Contemporary Problems, (2) (August 1970): 817-838.

2. For a discussion of the impacts of early health planning legislation on hospitals of different ownership types see May, Joel J., "Health Planning-Its Past and Potential," Health Administration Perspectives No. A5 (Chicago: Center for Health Administration Studies, University of Chicago, 1967). The differential effects of the more recent Certificate of Need laws are discussed in Hyman, Herbert H., Health Regulation: Certificate of Need and 1122 (Baltimore: Aspen Systems Corporation, 1977).

3. Lindsay, Cotton M., "Medical Care and the Economics of Sharing," Economica, 4(4) (November 1967): 351-362.

4. Klarman, Herbert E., The Economics of Health (New York: Columbia University Press, 1965); Somers, Anne R., Hospital Regulation: The Dilemma of Public Policy (Princeton, NJ: Princeton University Press, 1969); Johnson, Richard L., "Data Show For-Profit Hospitals Don't Provide Comparable Service," Moder Hospital, 65(4) (April 1971): 116-118; Stewart, David A., "The History and Status of Proprietary Hospitals," Blue Cross Reports, 7 (March 1973): 10-16.

5. There is evidence of "cream skimming" by for-profit hospitals but this does not necessarily imply either overtreatment or lower quality of care. See Bays, Carson W., "Case-Mix Differences Between Nonprofit and For-Profit Hospitals," Inquiry, 14(1) (March 1977): 17-21; and Clark, Robert Charles, "Does the Nonprofit Form Fit the Hospital Industry?" Harvard Law Review, 93 (1980): 1419-1489.

6. Kwon, Jene K., "On the Relative Efficiency of Health Care Systems," Kyklos, 51 (1974): (fasc 4) 821-837; Pauly, Mark V., and Redisch, Michael, "The Not-For-Profit Hospital as a Physician's Cooperative," American Economic Review, 63(1) (March 1973): 87-99; Rossett, Richard N., "Proprietary Hospitals in the United States," in The Economics of Health and Medical Care, Mark Perlman, Ed. (New York: Wiley, 1974), pp. 57-65; Shalit, Sol S., "A Doctor-Hospital Cartel Theory," Journal of Business, 50(1) (January 1977): 1-20.

7. See, respectively, Parsons, Talcott, The Social System (London: The Free Press of Glencoe, 1951), p. 464; and Arrow, Kenneth J., "Uncer- tainty and the Welfare Economics of Medical Care," American Eco- nomic Review 53(5) (December 1963): 965.

8. Glaser, William A., Social Settings and Medical Organization: A Cross- National Study of the Hospital (New York and Chicago: Atherton- Aldine, 1970).

9. White, William, "A Brief History of the Hospital Industry Since 1900," in Research in Health Economics, Vol. II, Richard Scheffler, Ed. (Greenwich, CT: JAI Press, 1982), pp. 143-170.

10. Davis, Karen, and Foster, Richard W., Community Hospitals: Inflation in the Pre-Medicare Period (Washington, DC: DREW, SSA, OES, 1972).

11. Davis, Karen, "Hospital Costs and the Medicare Program," Social Security Bulletin 23(3) (August 1973): 19-28.

12. Hellinger, Fred J., "Hospital Charges and Medicare Reimbursement," Inquiry, 12(4) (December 1975): 313-319; Joseph, Hyman, "On Inter- departmental Pricing of Not-For-Profit Hospitals," Quarterly Review of Economics and Business, 12(4) (December 1975): 33-44; Harris, Jef- frey E., "Pricing Rules for Hospitals," Bell Journal of Economics, 10(1) (Spring 1979): 224-243; Danzon, Patricia Munch, 'Profits' in Hospital Laboratories: The Effects of Reimbursement Policies on Hospital Costs and Charges (Santa Monica, CA: The Rand Corporation, 1980); Caul-

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 383

field, Stephen, Cross Subsidies in Hospital Reimbursement (Washington, DC: Government Research Corporation, June 1981); Monheit, Alan C., and Hornbrook, Mark C., "Redistributive Effects of Hospital Reimbursement: Do Private/Charge Paying Patients Sub- sidize the Care of Public/Cost Paying Patients?," paper presented at the annual meetings of the Econometric Society, Washington, DC, De- cember 29, 1981.

13. Harris, op. cit., p. 240. 14. Feldman, Paul, "Efficiency, Distribution, and the Role of Government

in a Market Economy," Journal of Political Economy, 79(3) (May-June 1971): 508-526.

15. For such a theory, see James, Estelle, "How Non-profits Grow: A Model," Journal of Policy Analysis and Management, 2(3) (Spring 1983): 350.

16. Arrow, op cit., Feldstein, Martin S., "Quality Change and the Demand for Hospital Care," Econometrica, 45(4) (October 1977): 1781-1802; Harris, Jeffrey E., "The Internal Organization of Hospitals: Some Economic Implications," Bell Journal of Economics 8(2) (Autumn 1977): 467-482.

17. Hellinger, op. cit. 18. White, op. cit., p. 145. 19. Benham, Lee, "Guilds and the Form of Competition in the Health Care

Sector," in Competition in the Health Care Sector, Warren Greenburg, Ed. (Germantown, MD: Aspen Systems Corporation, 1978), pp. 363-374.

20. Shafer, Henry B., The American Medical Profession 1783-1850 (New York: AMS Press, 1968), pp. 33-34; Packard, Francis R., History of Medicine in the United States, Volume I (New York: Hafner Publishing Company, 1931), pp. 273-274.

21. Shryock, Richard H., Medical Licensing in America 1650-1965 (Balti- more: The Johns Hopkins University Press, 1967), p. 23.

22. White, op. cit., p. 146. 23. U.S. Department of Health, Education and Welfare, Health Manpower

Sourcebook, Section 9, Physicians, Dentists and Professional Nurses (Washington, DC: U.S. GPO, 1959), p. 9.

24. Shryock, op. cit., pp. 30-42. 25. Berlant, Jeffrey L., Profession and Monopoly (Berkeley, CA: University

of California Press, 1975), pp. 234-235. 26. Ibid., pp. 220-221. 27. Kessel, Reuben, "Price Discrimination in Medicine," Journal of Law

and Economics, 1(1) (October 1958): 25-29; "The A.M.A. and the Supply of Physicians," Law and Contemporary Problems, 35(1) (Spring 1970): 267-269.

28. Friedman, Milton, and Kuznets, Simon, Income from Independent Professional Practice (New York: National Bureau of Economic Re- search, 1945): Rayack, Elton, "The Physicians' Service Industry," in The Structure of American Industry, 6th ed., Walter Adams, Ed. (New York: Macmillan, 1982), pp. 388-426.

29. Steinwald, Bruce, and Neuhauser, Duncan, "The Role of the Proprie- tary Hospital," Law and Contemporary Problems, 35(2) (August 1970): 817-838.

30.Ibid., p. 827; Lave, Judith R., and Lave, Lester B., The Hospital Construction Act (Washington, DC: American Enterprise Institute, 1974); Kushman, John E., and Nuckton, Carol F., "Further Evidence on the Relative Performance of Proprietary and Nonprofit Hospitals," Medical Care, 15(2) (May 1977): 55-67.

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

384 Why Most Private Hospitals Are Nonprofit

31. Bays, Carson W., "Patterns of Hospital Growth: The Case of Profit Hospitals," Medical Care (forthcoming).

32. Kessel, Price Discrimination in Medicine, pp. 33-34. 33. Manning, Willard G., Jr., "Comparative Efficiency in Short-Term

General Hospitals," unpublished Ph.D. thesis, Stanford University, Palo Alto, CA, 1973; Pauly, Mark V., and Redisch, Michael, "The Not-For-Profit Hospital as a Physician's Cooperative," American Eco- nomic Review, 63(1) (March 1973): 87-99; Shalit, Sol S., "A Doctor- Hospital Cartel Theory," Journal of Business, 50(1) (January 1977): 1-20.

34. For a rigorous yet accessible discussion of the conflict between the interests of the group and of the individuals comprising it, see Olson, Mancur, The Logic of Collective Action (Cambridge, MA: Harvard University Press, 1971), especially chapters I and II.

35. Patinkin, Don, "Multiple-Plant Firms, Cartels, and Imperfect Competi- tion," Quarterly Journal ofEconomics, 57(1) (February 1947): 173-205.

36. Reinhardt, Uwe E., Physician Productivity and the Demand for Health Manpower (Cambridge, MA: Ballinger, 1975); Smith, Kenneth R., Miller, Marriene, and Golladay, Frederick L., "An Analysis of the Optimal Use of Inputs in the Production of Medical Services," Journal of Human Resources, 7(2) (Spring 1972): 208-224.

37. Feldstein, Paul J., Health Care Economics (New York: Wiley, 1979), p. 325.

38. Feldstein, Paul J., Health Associations and the Demand for Legislation (Cambridge, MA: Ballinger, 1977), p. 45.

39. Lave and Lave, op. cit., pp. 8-9. 40. Feldstein, Health Associations, p. 45. 41. American Hospital Association, "Study of For-Profit Chains," unpub-

lished report (Chicago: AHA, 1970). 42. See May and Hyman, note 2, supra. 43. Havighurst, Clark C., Deregulating the Health Care Industry (Cam-

bridge, MA: Ballinger, 1982), pp. 363-365. 44. Feldstein, Health Associations, p. 160. 45.Ibid., pp. 147-148, 160-161. 46. Goldberg, Lawrence G., and Greenburg, Warren, "The Emergence of

Physician-Sponsored Health Insurance: A Historical Perspective," in Greenburg, op cit., pp. 231-254.

47. Clark, "Does the Nonprofit Form Fit the Hospital Industry?" pp. 1433-1441, 1446-1447.

48. The theory was first formalized by Stigler, George J., "The Theory of Economic Regulation," Bell Journal of Economics and Management Science, 2(1) (Spring 1971): 3-21. Contributions have been made by Posner, Richard A., "Taxation by Regulation," Bell Journal of Econom- ics and Management Science, 2(1) (Spring 1971): 22-50; "Theories of Economic Regulation," Bell Journal of Economics and Management Science, 5(2) (Autumn 1974): 335-358; and Peltzman, Sam, "Toward a More General Theory of Regulation," Journal of Law & Economics, 29(2) (August 1976): 211-240. Empirical tests of some aspects of the theory are contained in Jordan, William A., "Producer Protection, Prior Market Structure and the Effects of Government Regulation," Journal of Law & Economics, 15(1) (April 1972): 151-176; Eisenstadt, David, and Kennedy, Thomas E., "Control and Behavior of Nonprofit Firms: The Case of Blue Shield," Southern Economic Journal, 48(1) (July 1981); 26-36; Paul, Chris W., II, "Competition in the Medical Profession: An Application of the Economic Theory of Regulation," Southern Economic Journal, 48(3) (January 1982): 559-569; and Ar- nold, Richard J., and Eisenstadt, David, "The Effects of Medical

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

Why Most Private Hospitals Are Nonprofit 385

Society Control of Blue Shield on Fees in the Physician Service Market: Some Preliminary Evidence," Quarterly Review of Economics and Business, 22(1) (Spring 1982): 32-44.

49. See Paul, op. cit. 50. Stigler, op. cit., p. 18. 51. Kushman and Nuckton, op. cit., pp. 190-191. 52. Ferber, Bernard, "An Analysis of Chain-operated For-profit Hospitals,"

Health Services Research, 6(1) (Spring 1971): 49-60. 53. May, J. Joel, "The Planning and Licensing Agencies," Regulating Health

Facilities Construction, Clark C. Havighurst, Ed. (Washington, DC: American Enterprise Institute, 1974), pp. 47-68.

54. Bays, Carson W., "Cost Comparisons of Forprofit and Nonprofit Hospitals," Social Science & Medicine, 13(c) (December 1979): 219-225; Lewin, Larry, Investor-Owned Hospitals: An Examination of Performance (Washington, DC: Lewin and Associates, 1978); Vignola, Margo L., "An Economic Analysis of For-Profit Hospitals," paper presented at the Western Economic Association Meetings, June 18, 1979, Las Vegas, Nevada.

55. Schwartz, William, and Joskow, Paul L., "Duplicated Hospital Facilities," New England Journal of Medicine, 303(12) (December 1980): 1449-1457; Whittaker, Gerald F., "Productivity and Efficiency in Merged Hospital Systems," unpublished manuscript, May 1980.

56. Ginsberg, Paul B., "Capital in Nonprofit Hospitals," unpublished Ph.D. dissertation, Harvard University, 1970.

57. Neuhauser, Duncan, "The Future of Proprietaries in American Health Services," in Havighurst, Regulating Health Facilities, pp. 233-247.

58. Nelson, Philip, "Political Information," Journal of Law and Economics, 19(2) (August 1976): 315-336.

59. Rayack, "Physicians' Service Industry" in Adams, op cit., table 3, p. 411.

60. Computed from Feldstein, Health Care Economics, table 9-2, p. 159. 61. Kissam, Philip C., Weber, William L., Bigus, Lawrence W., and

Holzgraefe, John R., "Antitrust and Hospital Privileges: Testing the Conventional Wisdom," California Law Review, 70(2) (May 1982): 595-685.

62. Hansmann, Henry, "The Rationale for Exempting Nonprofit Organi- zations from Corporate Income Taxation," Yale Law Journal, 91(6) (November 1981): 54- 100.

63. Relman, Arnold S., "The New Medical-Industrial Complex," The New England Journal of Medicine, 303(10) (October 1980): 963-970.

64. Ibid., p. 968. 65. Compare, for example, McCampbell, S.R., "PSRO is a Four Letter

Word," OSMA Journal, 66(1) (January 1973): 3, with Raymond D. Goodman, Ed., PSRO: An Educational Symposium (Los Angeles: UCLA School of Medicine, 1975).

66. American Hospital Association, Reimbursement Survey (Chicago: AHA, October 1978); U.S. DHEW, HCFA, Professional Standards Review Organization 1978 Evaluation (Washington, DC: U.S. GPO, November 1978).

67. Congressional Budget Office, The Effect of PSRO's on Health Costs: Current Findings and Future Evaluations (Washington, DC: U.S. GPO, June 1979).

68. Havighurst, Clark C., Deregulating the Health Care Industry (Cam- bridge, MA: Ballinger, 1982); National Center for Health Services Research, A Synthesis of Research on Competition in the Financing and Delivery of Health Services (Washington, DC: U.S. Public Health Ser- vice, October 1982).

This content downloaded from �����������71.232.63.204 on Mon, 02 Oct 2023 07:02:33 +00:00������������

All use subject to https://about.jstor.org/terms

RESEARCH ARTICLE Open Access

Evidence that collaborative action between local health departments and nonprofit hospitals helps foster healthy behaviors in communities: a multilevel study Geri Rosen Cramer1, Gary J. Young2, Simone Singh3, Jean McGuire4 and Daniel Kim5*

Abstract

Background: The Patient Protection and Affordable Care Act of 2010 (ACA) encouraged nonprofit hospitals to collaborate with local public health experts in the conduct of community health needs assessments (CHNAs) for the larger goal of improving community health. Yet, little is known about whether collaborations between local health departments and hospitals may be beneficial to community health. In this study, we investigated whether individuals residing in communities with stronger collaboration between nonprofit hospitals and local public health departments (LHDs) reported healthier behaviors. We further explored whether social capital acts as a moderating factor of these relationships.

Methods: We used multilevel cross-sectional models, controlling for both individual and community-level factors to explore LHD-hospital collaboration (measured in the National Association of County and City Health Officials (NACC HO) Forces of Change Survey), in relation to individual-level health behaviors in 56,826 adults living in 32 metropolitan and micropolitan statistical areas, captured through the 2015 Behavioral Risk Factor Surveillance System (BRFSS) SMART dataset. Nine health behaviors were examined including vigorous exercise, eating fruits and vegetables, smoking and binge drinking. Social capital, measured using an index developed by the Northeast Regional Center for Rural Development, was also explored as an effect modifier of these relationships.

Results: Stronger collaboration between nonprofit hospitals and LHDs was associated with not smoking (odds ratio, OR 1.32, 95% CI 1.11 to 1.58), eating vegetables daily (OR 1.29; 95% CI 1.06 to 1.57), and vigorous exercise (OR 1.17; 95% CI 1.05 to 1.30). The presence of higher social capital also strengthened the relationships between LHD- hospital collaborations and wearing a seatbelt (p for interaction = 0.01) and general exercise (p for interaction = 0.03).

(Continued on next page)

© The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

* Correspondence: [email protected] 5Bouvé College of Health Sciences and the School of Public Policy and Urban Affairs, Northeastern University, Boston, MA, USA Full list of author information is available at the end of the article

Cramer et al. BMC Health Services Research (2021) 21:1 https://doi.org/10.1186/s12913-020-05996-8

(Continued from previous page)

Conclusions: Stronger collaboration between nonprofit hospitals and LHDs was positively associated with healthier individual-level behaviors. Social capital may also play a moderating role in improving individual and population health.

Keywords: Collaboration, Population health, Local health department, Nonprofit hospital, Social capital, Health behaviors

Background In 2014, the United States ranked last among high- income countries on measures of mortality and life ex- pectancy while health expenditures per person were more than double the highest-ranking country [1, 2]. Within the US, significant health disparities exist within and across populations and regions [3]. The pioneering work of the Robert Wood Johnson Foundation’s “Cul- ture of Health Action Framework” and the Institute for Healthcare Improvement (IHI)‘s “Triple Aim,” have pro- posed that health is a function of community assets and socioeconomic circumstances [4, 5]. Cross-sector part- nerships between local public health departments and healthcare providers are a potentially important resource for improving community health and wellbeing. However, a long-standing chasm has existed between

the public health and healthcare delivery sectors [6, 7]. In general, local health departments (LHDs) have the mission to promote community health through pro- grams aimed at disease prevention and emergency pre- paredness. Meanwhile, healthcare providers focus on patient-level treatment for acute and chronic health con- ditions. Among healthcare providers, hospitals are a par- ticularly important community asset as they possess knowledge, data, and human capital resources that are relevant for improving community health [8]. Although over half of US hospitals are nonprofit, tax-exempt en- tities and therefore carry societal and governmental ex- pectations to provide community benefits, these hospitals have typically sought to meet this expectation by providing charity care to the poor or investing in health education and research activities [9]. These efforts address important needs but do little to directly improve community health [9]. Collaboration between LHDs and hospitals may be impeded by differences in mission, lan- guage and training of those in leadership positions [10]. At the same time, empirical research has explored the

potential benefits that can be gained from greater levels of collaboration between LHDs and hospitals. One of the few relevant studies found that when hospitals and LHDs do collaborate on community health needs assess- ments, the results are more comprehensive and detailed than when either entity (LHD and hospital) conducted an assessment independently [11]. In addition, an eco- logical study conducted by Mays et al. showed that

greater collaboration among various stakeholders was as- sociated with lower county-level mortality rates [12]. Yet, a significant knowledge gap exists in that none of these studies have implemented a multilevel study de- sign to directly test whether collaboration between LHDs and hospitals is associated with health behaviors when measured at the individual-level.

Conceptual framework Although there are many relevant frameworks in the public health, sociology and economics literature that address community health and welfare, we adapt a framework created by the Bay Area Regional Health In- equities Initiative (BARHII, modified framework in Fig. 1) [13]. The BARHII Framework for Reducing Health Inequi-

ties has become a useful tool for LHDs seeking to better address social determinants [13]. This framework is par- ticularly suitable for our work in that it recognizes the importance of strategic partnerships between institutions and articulates that individual health behaviors are on the path to individual disease and mortality. One area of interest not explicitly identified in the

BAHRII Framework is the concept of social capital. In a community context, social capital has been defined as the “source of the ability to identify problems and needs, achieve a workable consensus on goals and priorities, and work in partnership with others to achieve goals.” [14, 15] Putnam, in his seminal work “Bowling Alone”, described community-level social capital as being a “public good” whereby efforts from collective action may benefit the population at large [16]. In this study, we sought to address two key research

questions using a multilevel study design: (1) are individ- uals residing in communities with long-standing LHD- hospital collaborative action on community-wide health promotion more likely to report healthier behaviors than individuals residing in communities with less collabora- tive action?; and (2) does the level of social capital of a community modify the relationships between long- standing LHD-hospital collaborative action and health behaviors of individuals? We hypothesized that individuals residing in commu-

nities with long-standing collaborative action would re- port healthier behaviors. We believe that LHDs and

Cramer et al. BMC Health Services Research (2021) 21:1 Page 2 of 12

nonprofit hospitals that engage in long-standing collab- orative action benefit from the synergies that have been observed through research on collaboration outside of the health context [17]. We also hypothesized that higher levels of community social capital could serve as a facilitator in these efforts. That is, social capital may act as a catalyst for such collaborative action to be effective.

Methods Study population Our study population consisted of 56,826 adults living in 32 metropolitan and micropolitan statistical areas (MMSA) in the United States. Geographies were selected for which we had individual-level responses and a single LHD that reported the presence of at least one nonprofit hospital within its jurisdiction.

Outcomes Information on individual-level health behaviors came from the 2015 Behavioral Risk Factor Surveillance Sur- vey (BRFSS) SMART dataset. The 2015 BRFSS SMART dataset is a sub-sample of the 2015 state BRFSS surveys based on geographies defined as metropolitan statistical areas, micropolitan statistical areas, and metropolitan

divisions (collectively called MMSAs) made publicly available to researchers. The 2015 BRFSS Smart dataset included 132 MMSAs where at least 500 BRFSS surveys were collected [18]. Individual health behaviors selected for this study were

those identified by the Centers for Disease Control and Prevention (CDC) as “Winnable Battles”. Winnable Bat- tles are health outcomes for which the CDC believes that public health can make significant progress in a rela- tively short time frame (i.e., within one to 4 years), have a large-scale public impact, and have evidenced-based interventions readily available for ease of implementa- tion. From the “Winnable Battles” list, we included the following modifiable behaviors/conditions into our ana- lysis: smoking, wearing a seatbelt, binge drinking, eating vegetables daily, eating fruit daily, general exercise in a month, vigorous exercise (300 min) in a week and being overweight or obese (based on self-reported height and weight). Additionally, while not designated by the CDC as a “Winnable Battle”, flu vaccinations are also an im- portant measure of community health where LHDs and nonprofit hospitals may collaborate and was thus in- cluded in our analysis. While we did not have access to specific collaborative action strategies, many of the in- cluded measures are commonly identified in community

Fig. 1 A Modified BARHII Public Health Framework for Reducing Health Inequities [13] (adapted)

Cramer et al. BMC Health Services Research (2021) 21:1 Page 3 of 12

health needs assessments (CHNAs) as being health needs in the community (healthy eating, physical activ- ity, smoking, etc.) [19, 20]. In addition to studying the impact of LHD-hospital

collaboration on specific individual health behaviors, we also conducted analyses using two index measures of the health behaviors. We created one index for risky behav- iors and another for healthy lifestyle behaviors (healthy eating and exercise). The risky behaviors index included wearing a seatbelt,

not smoking, not binge drinking and getting a flu shot. For each individual respondent, we assigned a score for this index based on the number of specific behaviors that the individual reported undertaking (or in the case of smoking and binge drinking reported not undertaking). Thus, if an individual did not report undertaking any of these behaviors, we assigned a score of zero. If an individ- ual reported all of the behaviors, we assigned a score of four. To create the healthy lifestyle index, we combined the specific behaviors of eating vegetable(s) daily, eating fruit daily and vigorously exercising. We found a high cor- relation between general exercise and vigorous exercise and felt the latter was more representative of a healthy lifestyle, so we included that variable only. For this index, we also assigned a score to each individual respondent based on the number of behaviors that reportedly were undertaken. Thus, the healthy lifestyle index ranged from zero (if an individual did not report any of the variables) to three (if an individual reported eating fruit, vegetables and vigorously exercising). Index variables were analyzed as continuous outcomes.

Predictor variables The data for assessing the level of collaboration between LHDs and nonprofit hospitals came from the 2015 Forces of Change Survey administered by the National Association of County and City Health Officials (NACC HO). This survey was developed to measure the impacts of economic forces on the budget, staff, and programs of LHDs. The survey was administered to a subset of the nearly 3000 LHDs across the country using stratified random sampling (on state and size of population in the LHD jurisdiction) [21]. Nine hundred and forty-eight (948) LHDs were randomly selected, of which 690 LHDs participated (73% response rate). Approximately 77% of the included LHDs reported having at least one non- profit hospital in their jurisdiction (n = 519). We used each LHD’s response to a single question as

a proxy for whether it had a long-standing collaboration with nonprofit hospitals in the community for which the LHD was responsible [22, 23]. The survey question asks “Is your LHD included in any nonprofit hospital’s imple- mentation plan for the community health needs assess- ment (CHNA)?” Response options included: no

collaboration, participating in the development of a hos- pital implementation plan, listed as a partner in a hos- pital implementation plan, conducting an activity together in a hospital implementation plan, and using the same implementation plan as the hospital. Because we were interested in identifying established collabora- tions between LHDs and hospitals within local commu- nities, we created a binary variable indicating “long- standing collaboration” for those LHDs that reported conducting an activity together or using the same imple- mentation plan as the nonprofit hospital in their com- munity. Although the survey question did not specify a defined time period for reported LHD-hospital collabor- ation, such CHNA implementation efforts typically en- tail multiple years of activity. Accordingly, we interpreted LHD responses indicating a joint effort for implementing community health needs assessments to be reflective of relatively long-standing relationships (or lack thereof) between an LHD and one or more non- profit hospitals in a community. While this variable lacked granularity in terms of the nature, strength, and scale of LHD-hospital collaboration (e.g., the content of implementation plans was not known), previous research suggests that any level of meaningful, ongoing collabor- ation between these two sectors within the same com- munity is uncommon [24]. Thus, we constructed this variable to measure if such collaboration is associated with positive individual-level health outcomes.

Social capital as a potential effect modifier To assess whether community social capital moderated the relationship between LHD-hospital collaboration and individual-level health behavior, we used the 2014 Northeast Regional Center for Rural Development (NRCRD) social capital index as developed by Rupasin- gha et al. (2006) [25]. This social capital index was cre- ated using principal component analysis based on four factors: the percentage of voters who voted in presiden- tial elections, the response rate to the Census Bureau’s decennial census, the number of non-profit organiza- tions, and the number of social organizations and associ- ations [25]. For social capital, the geographic unit was the county in which the LHD was located. While this was not the unit for the primary analysis, a metropolitan or micropolitan statistical area (MMSA) is defined as a region that consists of a city and surrounding communi- ties that are linked by social and economic factors, as established by the U.S. Office of Management and Budget (OMB) [18]. Thus, social capital at the county- level was deemed an adequate measure for this analysis.

Covariates In all models, we controlled for factors at the individual level (age, marital status, race, education, household

Cramer et al. BMC Health Services Research (2021) 21:1 Page 4 of 12

income, and health insurance status from the BRFSS, all as categorical variables except for age) and the community-level (median household income and per- cent black), derived from the 2010 U.S. Census and Area Resource File [26–28]. Individual survey responses were linked to community-level characteristics from the NRCRD social capital index and NACCHO Profile Sur- vey (2013). We included a variable indicating whether an LHD had more than the average number of full-time equivalents (FTEs) from all LHDs responding to the 2013 NACCHO Profile survey. FTEs relate to a LHD’s resource availability that may affect collaborative action as well as health outcomes [29]. We also controlled for the presence/absence of a local board of health, which has been shown to influence collaboration between non- profit hospitals and LHDs [23]. Finally, we controlled for whether the state had expanded Medicaid as of 2015.

Statistical analysis Due to missing data for two covariates—whether the in- dividual identified as being insured and the individual’s income (35 and 18%, respectively)—we implemented a multiple imputation approach using proc. mi, SAS ver- sion 9.4 [30]. To assess whether there was residual con- founding in missing observations for the health insurance variable, we conducted a sensitivity analysis that excluded the health insurance variable from our models that analyzed the two health behavior indices. Missing data on the outcomes of interest (ranging from 4.3% for not smoking to 13.2% for vigorous exercise) was cause for excluding the individual from the analysis. To examine the relationship between LHD-hospital collab-

oration and an individual’s reported health behavior, we esti- mated hierarchical generalized linear models using proc. glimmix [31]. We generalized the individual-level logistic re- gression model for binary and continuous outcomes and in- corporated the aforementioned community- and individual- level covariates. All models incorporated random intercepts to capture the variation among communities and to adjust the estimates for the lack of independence and clustering of individual responses within each community. Hypothesis testing was two-sided with a type 1 error rate of 0.05.

Results As noted, after data linkages our sample comprised 56, 826 respondents representing 32 communities across 25 states. Geographies were excluded if there was no data from both the Forces of Change survey or BRFSS Smart and if the MMSA (from the BRFSS Smart) represented more than one LHD. Table 1 compares descriptive sta- tistics between respondents from these 32 communities and those from excluded communities. The two groups were largely comparable except for race/ethnicity,

whereby included communities’ members were more likely to self-identify as white (83.3% vs 79.6%). Within our sample, approximately 14% of the popula-

tion reported smoking, 12% reported binge drinking, 11% reported not wearing a seatbelt and 51% reported not getting a flu shot. Despite most of the individuals reporting that they ate fruits and vegetables daily (62 and 80% respectively), 65% of respondents identified as being overweight or obese. Seventy-five percent (75%) of individuals reported that they performed some type of exercise monthly but only 38% reported vigorously exer- cising weekly. Variation also existed for the key independent variable,

the level of LHD-hospital collaboration. Approximately 22% of the communities in our sample reported long- standing collaboration between the LHD and one or more hospitals in its community (7 of 32 LHDs). Thus, 78% of the communities in our sample reported rela- tively little or no LHD-hospital collaboration. Those LHDs answering “I don’t know” were assumed to have no collaboration with nonprofit hospitals, an assumption previously used in the peer-reviewed literature [23]. Thus, in line with previous studies, long-standing collab- oration between LHDs and hospitals appeared to be un- common and thus, the binary variable for LHD-hospital collaboration included in the present study appears to capture an important qualitative distinction among the 32 communities represented in our study sample. From multilevel analyses where both individual and

community characteristics were controlled, we found significant and positive associations between LHD- hospital collaboration and the composite indices of risky behaviors (OR = 1.18; 95% CI = 1.10–1.28; Table 2) and healthy nutrition/lifestyle behaviors (OR = 1.12; 95% CI = 1.05–1.19; Table 3). LHD-hospital collaboration was also significantly and positively associated with three in- dividual outcomes: not smoking (OR = 1.32, 95% CI = 1.11–1.58; Table 2), eating vegetables daily (OR = 1.29; 95% CI = 1.06–1.57; Table 3) and vigorous exercise (OR = 1.17; 95% CI = 1.05–1.30; Table 3). In sensitivity analyses, removing the health insurance variable from the models did not materially change our findings. In models that included the interaction term between

LHD-hospital collaboration and social capital, we found a significant, positive interaction for two of our outcome mea- sures: wearing a seatbelt (p for interaction = 0.01; Table 2) and general exercise (p for interaction = 0.03; Table 3). After stratification, collaborative action was not significantly associ- ated with either behavior. There were no interactions for other health behaviors or index variables (Tables 2 and 3).

Discussion There is growing recognition that improving population health requires multi-sector collaboration. The Patient

Cramer et al. BMC Health Services Research (2021) 21:1 Page 5 of 12

Protection and Affordable Care Act of 2010 (ACA) en- couraged nonprofit hospitals to collaborate with local public health experts in the conduct of CHNAs for the larger goal of improving community health [32]. Long-

standing collaborations between the local public health and hospital sectors, however, appear to be quite limited [10, 11]. A previous study using hospital IRS filings showed that only about half of nonprofit hospitals

Table 1 Comparison of characteristics of individuals living in included vs. excluded communities in full BRFSS Smart cohort (total of 130 MMSAs)

Included communities (n = 32 MMSAs); % of all respondents

Excluded communities (n = 98 MMSAs); % of all respondents

Sex

Male 48.7 48.3

Female 51.3 51.7

Marital

Married 51.4 49.3

Divorced/Widowed/Separated 19.9 19.2

Never Married 23.6 25.6

Member unmarried couple 4.4 5.1

Refused 0.6 0.8

Education

Less than Grade 12 12.6 13.3

Grade 12 or GED (High school graduate) 26.8 25.4

College 1 year to 3 years (Some college or technical school) 32.0 30.0

College 4 years or more (College graduate) 28.2 30.8

Refused 0.4 0.5

Employment Status

Employed 49.9 51.0

Not employed 13.6 13.0

Retired 29.2 28.8

Unable to work 6.3 6.3

Refused 1.0 0.9

Income

Less than $20,000 12.3 13.3

$20,000 to $75,000 41.7 38.7

Greater than $75,000 28.6 30.9

Refused/Don’t know 17.3 17.2

Preferred Race

White 83.3 79.6

Black or African American 8.1 11.1

Asian 1.9 2.9

Other 4.3 4.2

Refused/Don’t know 2.5 2.2

Age

Under 30 11.0 10.5

30–50 25.1 24.7

50–65 29.5 30.4

Over 65 33.1 33.0

Refused/Don’t Know/Missing 1.3 1.4

Cramer et al. BMC Health Services Research (2021) 21:1 Page 6 of 12

Table 2 Unstratified Odds Ratios of Individual-Level Risky Behaviors

Wearing a Seatbelt Not Smoking Not Binge Drinking Getting a Flu Shot Risky Behavior Index

Odds Ratio

95% CI Odds Ratio

95% CI Odds Ratio

95% CI Odds Ratio

95% CI Odds Ratio

95% CI

Primary Predictors of Interest

Collaborative action 1.55 0.96, 2.49

1.32* 1.11, 1.58

1.19 0.92, 1.55

1.03 0.89, 1.20

1.18* 1.10, 1.28

Social capital index 1.09 0.67, 1.76

0.94 0.80, 1.10

0.93 0.75, 1.16

1.41* 1.07, 1.87

1.15* 1.05, 1.25

Individual-level Covariates

Male 0.49* 0.41, 0.58

0.63* 0.48, 0.82

0.47* 0.39, 0.57

0.79* 0.69, 0.91

0.76* 0.69, 0.83

Married 1.48* 1.12, 1.95

1.64* 1.31, 2.05

1.35* 1.00, 1.83

1.13* 1.01, 1.27

1.23* 1.09, 1.38

College 1.37* 1.09, 1.72

1.37* 1.17, 1.61

0.94 0.70, 1.24

1.25* 1.07, 1.46

1.17* 1.07, 1.27

Black 0.86 0.57, 1.30

1.02 0.72, 1.42

1.45* 1.06, 1.98

0.81* 0.69, 0.95

0.89 0.71, 1.12

Hispanic 1.10 0.76, 1.58

1.56* 1.13, 2.15

1.16 0.72, 1.87

0.87 0.68, 1.11

0.90* 0.85, 0.95

Asian 1.05 0.68, 1.64

3.10* 1.05, 9.16

3.08* 1.50, 6.32

1.10 0.77, 1.55

1.04 0.86, 1.25

Other Race 1.05 0.43, 2.57

0.99 .042, 2.32

0.97 0.49, 1.90

0.97 0.48, 1.99

1.01 0.76, 1.32

Age 1.16* 1.11, 1.22

1.09* 1.01, 1.17

1.35* 1.24, 1.48

1.36* 1.29, 1.43

1.14* 1.11, 1.17

Insured 1.22 0.75, 1.97

1.64* 1.31, 2.04

1.06 0.85, 1.31

1.98* 1.46, 2.67

1.28* 1.15, 1.43

Income < $15 k 0.94 0.47, 1.92

0.45* 0.33, 0.62

1.19 0.86, 1.65

0.88 0.64, 1.21

0.93 0.74, 1.17

Income $15 k–75 k 0.88 0.66, 1.19

0.68* 0.54, 0.86

1.13 0.95, 1.36

0.93 0.81, 1.07

0.97 0.88, 1.06

Community-level Covariates

Median household income

1.04 0.93, 1.16

1.03 0.97, 1.11

0.95 0.90, 1.01

1.07 0.96, 1.18

1.00 0.97, 1.04

Percent Black 1.03 0.78, 1.36

0.93 0.83, 1.04

0.98 0.85, 1.12

0.91 0.82, 1.02

0.98 0.94, 1.02

Above Average FTEs 0.99 0.64, 1.55

1.30* 1.09, 1.56

0.82 0.61, 1.12

0.76 0.56, 1.04

0.93 0.81, 1.07

Local Board of Health 0.71 0.49, 1.02

0.92 0.74, 1.13

1.11 0.83, 1.65

0.90 0.67, 1.23

0.92 0.83, 1.01

State-level Covariates

Expanded Medicaid 0.72 0.49, 1.08

0.86 0.74, 1.02

0.79* 0.63, 0.98

0.78* 0.67, 0.91

0.86* 0.80, 0.93

p-value for interaction 0.01 0.46 0.17 0.88 0.91

*P < 0.05. The social capital index was dichotomized into communities with an index above the median (1 = yes) and below the median in our sample (0 = no). Individual-level covariates included being male, being married, graduating college, being Black, Asian, Hispanic, or Other race, having income in the specified range (<$15,000/year, 15,000 to 75,000/year), and having health insurance. The reference category for education was having less than a college degree and the reference for income was greater than $75,000 per year. The remaining dichotomous predictors were coded as 1 = yes; 0 = no. The number of FTEs at the LHD was =1 if the LHD had more than the average FTEs in the Profile 2013 survey (mean = 65). Community-level covariates included continuous measures of the population: percent black (Area Resource File, 2016; rescaled to a one-standard deviation change) and median household income (2014; rescaled to a $10,000 unit change). A second model with an interaction term between collaborative action and social capital was also analyzed. The p-value for interaction is provided.

Cramer et al. BMC Health Services Research (2021) 21:1 Page 7 of 12

Table 3 Unstratified Odds Ratios of Individual-Level Healthy Lifestyle Behaviors

Eating Vegetables

Eating Fruit General Exercise Vigorous Exercise Healthy Weight Healthy Lifestyle Index

Odds Ratio

95% CI

Odds Ratio

95% CI

Odds Ratio

95% CI

Odds Ratio

95% CI

Odds Ratio

95% CI

Odds Ratio

95% CI

Primary Predictors of Interest

Collaborative action 1.29* 1.06, 1.57

1.16 0.99, 1.35

1.17 1.00, 1.38

1.17* 1.05, 1.30

1.06 0.93, 1.21

1.12* 1.05, 1.19

Social capital index 1.10 0.94, 1.29

1.11 0.92, 1.33

0.91 0.73, 1.13

1.05 0.92, 1.19

0.95 0.78, 1.17

1.07* 1.01, 1.14

Individual-level Covariates

Male 0.64* 0.57, 0.72

0.71* 0.58, 0.87

1.15 0.99, 1.33

1.02 0.92, 1.14

0.44* 0.35, 0.57

0.96 0.88, 1.06

Married 1.35* 1.09, 1.66

1.27* 1.09, 1.48

1.18 0.97, 1.45

1.31* 1.07, 1.62

0.80* 0.71, 0.90

1.12* 1.03, 1.21

College 1.16 0.94, 1.42

1.16* 1.08, 1.25

1.14 0.90, 1.44

1.16* 1.01, 1.34

1.15 0.92, 1.43

1.11* 1.05, 1.18

Black 0.59* 0.42, 0.82

1.17 0.92, 1.49

0.69 0.47, 1.00

0.97 0.81, 1.18

0.52* 0.35, 0.78

0.87 0.74, 1.01

Hispanic 1.13 0.80, 1.60

1.25* 1.07, 1.44

0.76* 0.63, 0.92

0.91 0.74, 1.13

0.59* 0.52, 0.68

0.86* 0.79, 0.94

Asian 0.99 0.72, 1.36

1.28 0.86, 1.90

0.71 0.49, 1.04

0.92 0.68, 1.25

1.68* 1.37, 2.07

0.82* 0.75, 0.89

Other Race 1.07 0.57, 2.01

1.02 0.66, 1.59

1.03 0.51, 2.06

1.08 0.48, 2.45

0.95 0.43, 2.08

1.01 0.81, 1.26

Age 1.03 0.98, 1.08

1.07* 1.04, 1.11

0.85* 0.82, 0.89

0.81* 0.79, 0.83

0.81* 0.77, 0.85

0.99 0.96, 1.01

Insured 1.05 0.78, 1.41

1.08 0.93, 1.26

1.46* 1.03, 2.06

1.28* 1.04, 1.57

1.02 0.81, 1.30

1.11* 1.03, 1.21

Income <$15 k 0.54* 0.34, 0.87

0.70* 0.56, 0.88

0.49* 0.37, 0.64

0.73* 0.55, 0.96

0.92 0.70, 1.22

0.93 0.80, 1.09

Income $15 k–75 k 0.79* 0.64, 0.99

0.91 0.77, 1.08

0.72* 0.62, 0.84

0.86* 0.75, 0.98

0.98 0.83, 1.17

0.97 0.90, 1.03

Community-level Covariates

Median household income

1.06* 1.02, 1.10

1.06 0.98, 1.14

1.02 0.96, 1.08

1.05* 1.01. 1.09

1.02 0.95, 1.10

1.00 0.98, 1.02

Percent Black 0.89* 0.82, 0.97

0.88* 0.78, 1.00

0.94 0.82, 1.08

1.01 0.93, 1.11

0.99 0.89, 1.10

0.99 0.96, 1.03

Above Average FTEs 0.97 0.79, 1.18

1.18 0.91, 1.52

0.99 0.75, 1.31

0.91 0.77, 1.07

1.08 0.85, 1.38

1.01 0.91, 1.11

Local Board of Health 0.82* 0.70, 0.95

0.912 0.75, 1.14

0.99 0.82, 1.19

0.97 0.85, 1.10

1.04 0.84, 1.27

0.96 0.89, 1.03

State-level Covariates

Expanded Medicaid 0.74* 0.65, 0.85

1.10 0.94, 1.29

0.92 0.77, 1.10

0.82* 0.73, 0.93

0.96 0.85, 1.08

0.95 0.90, 1.01

p-value for interaction 0.23 0.46 0.03 0.40 0.14 0.13

*P < 0.05. The social capital index was dichotomized into communities with an index above the median (1 = yes) and below the median in our sample (0 = no). Individual-level covariates included being male, being married, graduating college, being Black, Asian, Hispanic, or Other race, having income in the specified range (<$15,000/year, 15,000 to 75,000/year), and having health insurance. The reference category for education was having less than a college degree and the reference for income was greater than $75,000 per year. The remaining dichotomous predictors were coded as 1 = yes; 0 = no. The number of FTEs at the LHD was =1 if the LHD had more than the average FTEs in the Profile 2013 survey (mean = 65). Community-level covariates included continuous measures of the population: percent black (Area Resource File, 2016; rescaled to a one-standard deviation change) and median household income (2014; rescaled to a $10,000 unit change). A second model with an interaction term between collaborative action and social capital was also analyzed. The p-value for interaction is provided.

Cramer et al. BMC Health Services Research (2021) 21:1 Page 8 of 12

reported including public health experts in any of their implementation activities [33]. Our findings are less opti- mistic in that 78% of the LHDs in this analysis reported low to no collaborative action on CHNA implementation with nonprofit hospitals in their community [33, 34]. We set out to study whether collaborative action be-

tween LHDs and nonprofit hospitals may impact com- munity health. We found that after controlling for a number of other factors, LHD-hospital collaborative ac- tion was significantly associated with several healthy be- haviors. Both of the composite health behavior indices showed significant positive associations, indicating that individuals living in communities with some level of LHD-hospital collaboration were more likely to report fewer risky behaviors and a greater number of healthy eating and exercising behaviors than those living in com- munities without documented long-standing collabora- tive action. Associations with all measured behaviors were in the hypothesized positive direction and three of these associations were both positive and significant. These findings are in line with the BARHII Framework’s inclusion of strategic partnership at the institutional level. While the exact mechanism by which long- standing collaboration (i.e. alterations in the physical, so- cial or service environment) may manifest its effects are unclear, these data show interesting potential impacts of such strategic collaboration. For social capital, there was some evidence of moder-

ation for two behaviors: wearing a seatbelt and general exercise. This is consistent with our hypothesis that community social capital resources may allow for a com- munity to benefit from implementation activities put forth through LHD-nonprofit hospital collaborative ac- tion. From these data, we believe our inclusion of social capital to the BARHII Framework is warranted; however, further study of the mechanisms by which social capital can benefit collaborative action between hospitals and LHDs would be useful. Although we report a positive association from LHD-

hospital collaborative action and individual-level healthy behaviors, only 22% of the communities represented by study respondents included an LHD that reported long- standing collaboration with nonprofit hospitals. While understanding why such collaboration is uncommon was not the objective of the current study, extant research points to several barriers. First, nonprofit hospitals may choose to focus on strategies that directly impact their current patient market instead of targeting the social and economic barriers to health of residents, especially those that do not receive health care at their institution. This choice minimizes the need to partner with external organizations, including the LHD. Second, funding for action (i.e., implementing programs) may be a concern to hospitals given the long-term nature of many

strategies required to improve population health. From the LHD’s perspective, budgets have been challenging, leaving staff to do more with less [6, 8]. Third, while current regulations state that nonprofit hospitals must seek input from public health experts to analyze com- munity needs, there is no such expectation for hospitals when prioritizing and implementing actions to meet those needs [9]. Without state or federal laws that re- quire such efforts, nonprofit hospitals may see no ration- ale to include LHDs in executing implementation strategies.

Strengths and limitations To our knowledge, this study offers the first analysis of long-standing collaborative action between LHDs and hospitals and individual-level health behaviors. Our study is exploratory yet methodologically robust in that we used multilevel models to explore impacts on individual-level behaviors, an advance over previous eco- logical studies. We believe we are also the first to assess whether social capital may play a role in modifying these relationships. We further focus on multiple modifiable health behaviors identified by the CDC as “winnable bat- tles” that can be pivotal to substantially improving popu- lation health across the nation. Nonetheless, there are several limitations to our study.

First, our data was observational in nature, being mostly derived from surveys. While we accounted for commu- nity and individual-level covariates, it is possible that factors omitted from our multivariate models could con- tribute to residual confounding. Relatedly, as a cross- sectional study, we could not rule out reverse causation; due to the emerging nature of this research, 2015 was the first, and only year of available data on collaborative action between LHDs and nonprofit hospitals on CHNA implementation from the perspective of the local health department. However, in line with our conceptual framework, we expect our measure of collaborative ac- tion to be indicative of a long-standing LHD-hospital collaboration that has developed over time in the com- munities studied [34]. Hence, the associations observed in this study represent collaborative relationships be- tween hospitals and LHDs with subsequent community health that are not a direct result of regulation. Second, because we relied on a single item assessment

of long-standing collaboration, it was not possible to as- sess which specific strategies the LHD and hospital jointly implemented. Health behaviors we measured as outcomes may not directly relate to the actions being undertaken through collaborative action. A qualitative analysis of CHNA implementation reports would be use- ful to further unpack the elements and strategies of collaboration.

Cramer et al. BMC Health Services Research (2021) 21:1 Page 9 of 12

Third, our measure of long-standing collaborative ac- tion captures the perspective of the LHD only and may not accurately reflect how hospitals in the community perceived the level of collaborative action on community health improvements activities with the LHD. Data from a prior study assessing collaboration with public health experts from the perspective of the nonprofit hospital showed similarly low levels of collaboration [10, 33]. Finally, while individuals from 25 of the 50 states were

included in our sample, our study findings may not ne- cessarily be generalizable to all communities in the US. Future studies that include individuals in the other 25

states would help to expand the external validity of our findings.

Conclusions This study assessed whether collaborative action be- tween local health departments and nonprofit hospitals was linked to individual-level self-reported health behav- iors. Our findings show positive and significant associa- tions for healthier behaviors (fewer risky behaviors and more healthy lifestyle behaviors). Additional research on collaborative action is needed including program evalua- tions that focus on health outcomes from specific

Fig. 2 A Map of Included Communities and their Social Capital Index Values [35]. Permission was received to use and adapt the copyrighted image

Cramer et al. BMC Health Services Research (2021) 21:1 Page 10 of 12

population health initiatives involving the collaboration of local health departments and hospitals. However, given that many LHDs still do not report any collabora- tive action with nonprofit hospitals in their area, re- search is also needed to better understand the barriers to such collaboration. Should our findings be replicated in future work including program evaluations and obser- vational longitudinal studies, they may serve to encour- age policy makers to consider incentives for LHDs and hospitals to take collaborative action to impact the health of their communities.

Abbreviations ACA: Affordable Care Act; BARHII: Bay Area Regional Health Inequities Initiative; BRFSS: Behavioral Risk Factor Surveillance Survey; CDC: Centers for Disease Control; CHNA: Community Health Needs Assessment; CI: Confidence Interval; FTE: Full-Time Equivalents; IHI: Institute for Healthcare Improvement; LHD: Local Health Department; MMSA: Metropolitan and Micropolitan Statistical Area; NACCHO: National Association of County and City Health Officials; NRCRD: Northeast Regional Center for Rural Development at Penn State; OMB: Office of Management and Budget; OR: Odds Ratio

Acknowledgements We thank John Griffith, PhD for biostatistical support.

Authors’ contributions GRC conceived and designed the project, acquired and analyzed the data, interpreted data and drafted the manuscript. GJY assisted in the conception and design, interpretation of the data, assisted in drafting manuscript and provided substantive revisions. SS acquired and interpreted data and provided substantive revisions to the manuscript. JM provided substantive revisions to the manuscript. DK assisted in the conception and design of the project and interpretation of the data, provided substantive revisions to the manuscript, and supervised the study. All authors have read and approved the manuscript.

Funding No funding was provided for this research.

Availability of data and materials Data used in this study came from: https://www.cdc.gov/brfss/smart/smart_2015.html (publicly available). https://aese.psu.edu/nercrd/community/social-capital-resources/social-capital- variables-for-2014 (publicly available). http://nacchoprofilestudy.org/forces-of-change/2015-forces-of-change/ (not publicly available). http://nacchoprofilestudy.org/data-requests/ (not publicly available). https://www.hilltopinstitute.org/our-work/hospital-community-benefit/ hospital-community-benefit-state-law-profiles/ (publicly available). https://data.census.gov/cedsci/table?d=ACS%205-Year%20Estimates%2 0Data%20Profiles&table=DP02&tid=ACSDP5Y2015.DP02 (publicly available).

Ethics approval and consent to participate This study was reviewed and approved by the Northeastern University IRB.

Consent for publication Not applicable.

Competing interests GRC declares no competing interests. GJY declares no competing interests. SS declares no competing interests. JM declares no competing interests. DK declares no competing interests.

Author details 1Bouvé College of Health Sciences and the Center for Health Policy and Healthcare Research, Northeastern University, Boston, MA, USA. 2Bouvé College of Health Sciences, D’Amore-McKim School of Business and the Center for Health Policy Healthcare Research, Northeastern University, Boston, MA, USA. 3Department of Health Management and Policy, University of Michigan, Ann Arbor, USA. 4Bouvé College of Health Sciences, Northeastern University, Boston, MA, USA. 5Bouvé College of Health Sciences and the School of Public Policy and Urban Affairs, Northeastern University, Boston, MA, USA.

Received: 22 July 2020 Accepted: 4 December 2020

References 1. Davis K, Stremikis, K, Schoen, C, Squires, D. Mirror, Mirror on the Wall, 2014

Update: How the U.S. Health Care System Compares Internationally. Commonwealth Fund. 2014. Available at: https://www.commonwealthfund. org/publications/fund-reports/2014/jun/mirror-mirror-wall-2014-update- how-us-health-care-system. Accessed on May 18, 2020.

2. National Research Council and Institute of Medicine. U.S. Health in International Perspective: Shorter Lives, Poorer Health. Panel on Understanding Cross-National Health Differences Among High-Income Countries. Washington, DC: The National Academies Press; 2013.

3. Dwyer-Lindgren L, Bertozzi-Villa A, Stubbs RW, et al. Inequalities in life expectancy among US counties, 1980 to 2014: temporal trends and key drivers. JAMA. 2017;177(7):1003–11. https://doi.org/10.1001/jamainternmed. 2017.0918.

4. Robert Wood Johnson Foundation. Culture of health. 2015. Available at: https://www.rwjf.org/en/cultureofhealth.html. Accessed on May 18, 2020.

5. University of Wisconsin Population Health Institute. County Health Rankings Key Findings 2018. 2018. Available at: https://www.countyhealthrankings. org/resources/2018-county-health-rankings-key-findings-report. Accessed on May 18, 2020.

6. Jacobson P, Parment W. Public health and health care: integration, disintegration or eclipse. J Law Med Ethics. 2018;46:940–51. https://doi.org/ 10.1177/1073110518821994.

7. Benton K, Polite, S. The disconnect between public health and health care. Vol march–April. Catholic health Association of the United States; 2016. Available at: www.chausa.org. Accessed on May 18, 2020.

8. Trust for America’s Health. The Impact of Chronic Underfunding of America’s Public Health System: Trends, Risks, and Recommendations, 2019. 2019. Available at: https://www.tfah.org/report-details/2019-funding-report/. Accessed on: May 18, 2020.

9. United States Government. An Act Entitled The Patient Protection and Affordable Care Act. In: Department of Health and Human Services, ed. Obama Care Act. Washington, D.C.: U.S. G.P.O.; 2010.

10. Beatty KE, Wilson KD, Ciecior A, Stringer L. Collaboration among Missouri nonprofit hospitals and local health departments: content analysis of community health needs assessments. Am J Public Health. 2015;105(Suppl 2):S337–44. https://doi.org/10.2105/ajph.2014.302488.

11. Pennel CL, McLeroy KR, Burdine JN, Matarrita-Cascante D. Nonprofit hospitals' approach to community health needs assessment. Am J Public Health. 2015;105(3):e103–13. https://doi.org/10.2105/ajph.2014.302286.

12. Mays GP, Mamaril CB, Timsina LR. Preventable death rates fell where communities expanded population health activities through multisector networks. Health Aff (Project Hope). 2016;35(11):2005–13. https://doi.org/10. 1377/hlthaff.2016.0848.

13. Bay Area Regional Health Inequities Initiative (BARHII). A Public Health Framework for Reducing Health Inequities. 2015; http://barhii.org/ framework/. Accessed August 3, 2018.

14. Tranel M, Gasen K. Community Partnerships: A Sustainable Resource for Nongovernmental Organizations. Federal Reserve Bank; 2003.

15. Mattessich P, Monsey B. Community building: what makes it work. St. Paul: Amherst H. Wilder Foundation; 1997.

16. Putnam RD. Bowling alone: the collapse and revival of American community. New York: Simon & Schuster; 2000.

17. Austin JE, Seitanidi MM. Collaborative value creation: a review of partnering between nonprofits and businesses: part I. value creation Spectrum and collaboration stages. Nonprofit Volunt Sect Q Title. 2012;41(5):726–58. https://doi.org/10.1177/0899764012450777.

Cramer et al. BMC Health Services Research (2021) 21:1 Page 11 of 12

18. Centers for Disease Control and Prevention. 2015 SMART: BRFSS City and County Data and Documentation 2015; https://www.cdc.gov/brfss/smart/ smart_2015.html.

19. Providence Alaska Medical Center. Anchorage Community Health Needs Assessment. 2015. Available at: https://www.psjhealth.org/-/media/files/ providence-st-joseph-health/community-benefit/ak/2015 chnaandchipprovidencealaskamedicalcenter.pdf?la=en. Published: 2015.

20. UPMC Mercy. Community Health Needs Assessment and Community Health Strategic Plan. Available at: https://www.upmc.com/-/media/upmc/ about/community-commitment/documents/mercy-community-health- needs-assessment.pdf. Published: June 30, 2013.

21. National Association of County and City Health Officials (NACCHO). About. 2019; https://www.naccho.org/about. Accessed June 10, 2019.

22. National Association of County and City Health Officials. Forces of Change Survey. 2015. http://nacchoprofilestudy.org/forces-of-change/2015-forces-of- change/.

23. Carlton EL, Singh SR. Joint community health needs assessments as a path for coordinating community-wide health improvement efforts between hospitals and local health departments. Am J Public Health. 2018;108(5): 676–82. https://doi.org/10.2105/AJPH.2018.304339.

24. Prybil LD, Scutchfield FD, Dixon RE. The Evolution of Public Health-Hospital Collaboration in the United States. Public Health Rep. 2016;131(4):522–5.

25. Rupasingha A, Goetz SJ, Freshwater D. The production of social capital in US counties. J Soc Econ Res. 2006;35(1):83–101. https://doi.org/10.1016/j. socec.2005.11.001.

26. United States Government. Area Health Resource File. Department of Health and Human Services HRSA, 2015. https://data.hrsa.gov/topics/health- workforce/ahrf.

27. United States Government. Census Data. 2015. https://www.census.gov/en. html.

28. Hilltop Institute. Community benefit state law profiles. Baltimore: University of Maryland; 2013. https://hilltopinstitute.org/our-work/hospital-community- benefit/hcbp-state-comparison/.

29. Schenck AP, Meyer AM, Kuo T, Cilenti D. Building the evidence for decision- making: the relationship between local public health capacity and community mortality. Am J Public Health. 2015;105(S2):S211–6.

30. Chaudhary KR, Schreiber-Gregory DN. Multiple Imputation for Arbitrary Missing Data: SAS® and R. Available at: https://www.mwsug.org/ proceedings/2015/TT/MWSUG-2015-TT-04.pdf. Published: 2015.

31. Ene ML, Blue EA, Bell BA, et al. Multilevel models for categorical data using SAS® PROC GLIMMIX: the basics. 2014.

32. United States Government. Schedule H (Form 990). In: Service IR, ed; 2014. Available at: https://support.sas.com/resources/papers/proceedings15/343 0-2015.pdf. Published: 2014.

33. Rosen Cramer G, Singh SR, Flaherty S, Young GJ. The Progress of US hospitals in addressing community health needs. Am J Public Health. 2017; 107(2):255–61. https://doi.org/10.2105/ajph.2016.303570.

34. Singh S, Carlton E. Local health departments’ involvement in hospitals’ implementation plans. Front Public Health Serv Sys Re. 2016;5(4):34–9.

35. Northeast Center for Rural Development. County-level measure of social capital. 2018. Available at: https://aese.psu.edu/nercrd/community/social- capital-resources. Accessed April 28, 2018.

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Cramer et al. BMC Health Services Research (2021) 21:1 Page 12 of 12

A Longitudinal Analysis of the Distinction between For-Profit and Not-for-Profit Hospitals in America

Author(s): Sharyn J. Potter

Source: Journal of Health and Social Behavior , Mar., 2001, Vol. 42, No. 1 (Mar., 2001), pp. 17-44

Published by: American Sociological Association

Stable URL: https://www.jstor.org/stable/3090225

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of Health and Social Behavior

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

A Longitudinal Analysis of the Distinction between For-profit

and Not-for-profit Hospitals in America*

SHARYN J. POTTER

University of New Hampshire

Journal of Health and Social Behavior 2001, Vol 42 (March): 17-44

Have changes in the hospital industry forced not-for-profit hospitals to become more like for-profit hospitals in measures of efficiency and community service? As a result, are not-for-profit hospitals moving away from their community ser- vice missions? In recent years researchers have asserted that the once-salient distinctions between not-for-profit and for-profit hospitals are quickly eroding and that this convergence threatens the community service that not-for-profit hospitals have historically provided. Neo-institutionalists explain that regulato- ry changes often force differing organization types to pursue similar strategies (Fligstein 1991, 1985; DiMaggio and Powell 1983). Guided by this theory, the present research analyzes if regulatory changes and the implementation of sim- ilar strategies result in not-for-profit andfor-profit hospitals having similar effi- ciency and community service outcomes.

Researchers have vigorously debated the Association 1992b). These teaching programs implications of the fading distinction between provide a substantial amount of care for the for-profit and not-for-profit hospitals. Policy poor (Reuter and Gaskin 1998). analysts, social scientists, and advocates for Some argue that the for-profit hospitals have the poor fear that the remnants of the safety infused a "corporate ethos" in the healthcare network for the poor, uninsured, and underin- system. This "ethos" is evident in the market- sured provided by some not-for-profit hospi- ing plans, information systems, and manage- tals will erode if not-for-profit hospitals ment techniques now found in not-for-profit become indistinguishable from their for-profit hospitals (Light 1986). Numerous communi- counterparts (Burns 1990; Stevens 1989; Starr ties support not-for-profit hospitals with tax- 1982). Not-for-profit hospitals historically payer dollars, income and property-tax exclu- have offered services that were not traditional- sions, and tax-free financing and contributions ly viewed as revenue producing, including (Kane 1993; Fox and Shaffer 1991). Therefore, burn units, neo-natal intensive care, units and many have been concerned that not-for-profit emergency departments (Altman and hospitals will jettison community service in an Shactman 1997; American Hospital attempt to reduce operating costs (Gentry and Association 1992a). Additionally, not-for-prof- Penrod 1998; Morrisey, Wedig, and Hassan it hospitals have accounted for the majority of 1996). Nonetheless, "two decades of research teaching programs (American Hospital has failed to provide definitive empirical evi-

dence on the differences between for-profit

* The author would like to thank the following indi- and nonprofit health care facilities" viduals for their comments and suggestions: (Blumenthal and Weissman 2000:158). E. Kathleen Adams, Edmund R. Becker, Cliff Brown, The existing research analyzing the question Timothy J. Dowd, Blair Gifford, Karen Hegtvedt, of convergence in the hospital industry has two Alex Hicks, Richard Levinson, Rick Rubinson and methodological shortcomings. First, most of Mike Schwartz. I am also grateful to the reviewers the empirical research examines cross-section- and John Mirowsky for their helpful comments and th e rical res rexamines cros-ection insights. Address correspondence to the author at al data or data representing a limited time the Department of Sociology, University of New frame (Kane 1993; Chang andrTuckman 1988). Hampshire, Horton Social Science Center, Durham, In other words, the researchers attempted to NH 03824; e-mail: sharyn.pottergunh.edu. develop a historical argument, but their assess-

17

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

18 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

ments lacked longitudinal data that allowed for care needs. Since 1913, the federal govern- trend analyses. Second, most of these studies ment has differentiated between not-for-profit ignore environmental pressures that may pro- and for-profit hospitals by exempting not-for- mote convergence (Arrington and Haddock profit hospitals from most revenue and proper- 1990; Chang and Tuckman 1988; Herzlinger ty taxes. In exchange for their tax exemption, and Krasker 1987). Typically, they examined not-for-profit hospitals were required to pro- the hospital's internal structure while overlook- vide free or below-cost medical services. In ing environmental factors such as policy, 1969 an Internal Revenue Service ruling

demand, and supply. Researchers who did con- superseded prior regulations and required sider the environment typically examined only merely that not-for-profit hospitals provide one facet (see Norton and Staiger 1994; services that benefit the community in Hultman 1991). exchange for their tax exemption (Fox and

My research addresses these limitations. Schaffer 1991). Therefore, not-for-profit hos- First, I analyze the claims of a narrowing dis- pitals "have historically had an aura of com- tinction between not-for-profit and for-profit

hospitals over a fifteen-year period with the muniy servic ro s utilization of a latent growth curve model that (Gray 1991 :61)

enables me to capture the development of cer- Not-for-profit hospitals are prohibited from tain trends. Additionally, the present research distributing their profits. Instead, profits must uses data from every short-term general hospi- be reinvested in the hospital (Gray 1991; Light tal in the 48 contiguous states. Therefore, the 1986). Alternatively, for-profit hospitals can results capture national trends rather than just distribute their profits to their owners or share- regional trends. Finally, various environmental holders. Consequently, stockholders demand factors, including the role of the larger regula- that for-profit hospitals should behave in a tory environment and internal factors includ- manner that results in profitable financial ing hospital ownership type are investigated. statements (Homer, Bradham, and Rushefsky

1984). Therefore, potential stockholders do not base their purchasing decisions on how well

ARE NOT-FOR-PROFIT AND FOR-PROFIT the hospitals are meeting community needs HOSPITALS CONVERGING? (Rushing 1976). In fact, some for-profit hospi-

tals avoid high-cost, low-profit services like If, as researchers postulated, the distinction outpatient departments, emergency-rooms,

between for-profit and not-for-profit hospitals and teaching programs in an effort to increase is narrowing, what reasons are given to explain hospital profits and stockholder returns this declining distinction? First, some argue (Ginzberg 1988). that the fading differences in the hospitals Public hospitals are an important compo- legal charter is an important factor in explain- nent in the not-for-profit/for-profit hospital ing the possible convergence between hospital debate. Their presence in a community influ- types. Others explain that the changes in the ences the strategies of the other hospitals, both

regulatory environment may promote a declin- for-profit and not-for-profit. Research indi- ing distinction between hospital types. Some cates that public hospitals often care for hypothesize that the hospital's external envi- . r ronment drives similarities across all hospital patients that other hospitals consider undesir- types. Finally, there are mediating processes able. A disproportionate number of their that need to be analyzed in an effort to under- patients are poor, uninsured or Medicaid recip- stand their effects in the possible convergence ients (Brown 1983). between hospital types. My analysis considers Most research on hospitals neglected to look all four of these explanations. at the hospital's legal charter until 1976, when

Rushing (1976) argued that an organization's profit orientation influences organizational

The Hospital s Legal Charter outcomes. However, in recent years some have argued that hospitals' legal charters no longer

Throughout the twentieth century hospitals provide insights into hospitals' operating in the for-profit, not-for-profit, and public strategies (Kane 1993; Fox and Schaffer 1991; health care sectors have met different health Marmor, Schlesinger, and Smithey 1987).

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 19

The Hospital Regulatory Environment patients (American Hospital Association 1994). Furthermore, private insurers soon

The neo-institutionalists hypothesize that began using the Prospective Payment System new legislation introduced to an organization reimbursement rates for their own enrollees population increases the likelihood that iso- (Reinhardt 1998). DiMaggio and Powell morphism (declining differences between (1983) indicate that increased regulatory con- organizational types) will occur. This enables straints force different types of organizations them to demonstrate that changes in the regu- to adopt strategies that result in similar out- latory environment influences organizational comes. In fact, many researchers pointed out responses to local demand and supply condi- that the scaled back Medicare payments forced tions. Therefore, neo-institutionalists postulate not-for-profit hospitals to behave like busi- that organization leaders consider current pol- nesses (Sorrentino 1989; Coddington, Palm- icy as they develop strategies for managing quist, and Trollinger 1985; Waldholdtz 1982). local demand and supply. Furthermore, the Therefore, the widespread adoption of cost- neo-institutionalists argue that the personnel of saving strategies following the passage of the separate organizations adopt similar strategies Prospective Payment System legislation could to comply with regulatory changes (Dowd and be attributed to a declining distinction between Dobbin 1998; Fligstein 1991, 1985; Baron, hospital types. Dobbin, and Jennings 1986).

Significant federal policy. The passage of the Medicare Prospective Payment System leg- The External Environment islation in 1983 initiated the implementation of highly innovative federal health-care cost con- Some argue that the environments in which tainment policy (Reinhardt 1998; Altman and organizations operate are more definitive than Young 1993; Hultman 1991; Rogers, Draper what the institutions call themselves on their and Kahn 1990; McCarthy 1988). The legal charters (Marmor, Schlesinger and Prospective Payment System legislation, Smithey 1987). That is, the hypothetical con- whose purpose is to regulate Medicare reim- vergence between for-profit and not-for-profit bursement marked the beginning of the use of hospitals may actually result from both types price controls at the national level in the hospi- of hospitals facing similar environmental fac- tal industry (Flood and Fennel 1995). tors. For instance, in highly competitive mar-

Until the passage of the Prospective Payment kets, Shortell and colleagues (1986) find that System legislation, hospitals had been using both for-profit and not-for-profit hospitals their own discretion in pricing their services. offer a larger number of alternative hospital The Prospective Payment System legislation services including geriatric care, health pro- eliminated such discretion by authorizing the motion and outpatient diagnostic services. government to establish uniform prices for all Alternately, when physicians are plentiful in a hospital services for Medicare patients, whether given area, hospitals compete for physicians' the treatment was provided at a for-profit or not- patients by increasing amenities for physicians for-profit hospital.' Therefore, the Prospective (e.g., acquiring sophisticated equipment) and Payment System legislation forced both hospital may duplicate services in an effort to attract types to incur costs lower than the reimburse- referrals (Menke 1997; Hadley and Swartz ment amount to make a profit or else suffer the 1989). Finally, research indicates that commu- consequences of a loss (Altman and Young nity wealth can create similarities in different 1993). Additionally, the "reasonable costs" for hospital types in the same communities (see which hospitals were able to bill Medicare prior Becker and Sloan 1985; Becker and Steinwald to the passage of Prospective Payment System 1981; Sloan and Becker 1981). Therefore, I allowed them to offset costly services via analyze those environmental factors that previ- Medicare reimbursement. The Prospective ous research has found to be significant deter- Payment System legislation forced some hospi- minants of hospital strategies. tals to discontinue some of their more costly services (Kauer, Silvers, and Teplensky 1995).

Since Medicare accounts for approximately Mediating Factors 40 percent of all hospital discharges, most hos- pitals were not in a position to refuse Medicare Many studies have analyzed internal factors

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

20 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

to determine whether the distinction between staff is also widely accepted as a proxy for hos- hospital types is narrowing (see Arrington and pital efficiency (Arrington and Haddock 1990). Haddock 1990; Herzlinger and Krasker 1987). Hospitals spend approximately 80 percent of I argue that the internal hospital characteristics their costs on labor (Anderson and Kohn 1996) act as mediators in that they are influenced by and can realize more cost savings by eliminat- the hospital type and ultimately influence hos- ing staff positions than by any other mechanism pital outcomes. Thus, for-profit and not-for- (Barrett 1995). Therefore, it is difficult for hos- profit hospitals that share the same internal pitals to earn profits when they have a large hos- characteristics may have similar community pital staff. For-profit hospitals have fewer per- service and efficiency outcomes. Therefore, I sonnel than their not-for-profit counterparts analyze three groups of mediating factors that (Woolhandler and Himmelstein 1997; Sorren- previous research has found to be significant; tino 1989; Light 1986; Watt et al. 1986). How- hospital facility characteristics, hospital prac- ever, as hospitals face an increasingly cost-con- tice characteristics and hospital payor mix. science environment, researchers hypothesize

that both hospital types will use staff reduction as a means of reducing costs.

EFFICIENCY AND COMMUNITY SERVICE AS INDICATORS OF

CONVERGENCE Community-Service Outcomes

In previous studies, researchers have used Compared to other hospital departments, similarities in hospital outcomes to demon- emergency departments are rarely successful strate a convergence between for-profit and profit centers, and hospitals with teaching not-for-profit hospitals (see Herzlinger and commitments generally care for disproportion- Krasker 1987), but they have not simultane- ate numbers of uninsured patients and patients ously analyzed efficiency and community-ser- with serious conditions (Blumenthal and vice outcomes. The present research addresses Weissman 2000; Reuter 1999). Therefore, both this gap by analyzing both efficiency and com- endeavors are seen as community service by munity-service outcomes. hospitals and their communities.

Emergency room utilization. Many see the emergency room as the cornerstone of a hospi-

Efficiency Outcomes tal's provision of community care: Emergency departments employ a specialized medical

Two of the most visible hospital efficiency staff capable of dealing with any medical prob- strategies are reducing costs and reducing the lem 24 hours a day, 7 days a week (Albrecht, size of the hospital staff. Slobodkin, and Rydman 1996). In the past few

Hospital Expenses. Total hospital expenses years, hospitals have seen an increase in are the annual costs incurred to operate the patient demand for emergency-room services hospital. Examples include payroll, adminis- (Williams 1996; General Accounting Office tration, cafeteria, supplies, operating-room, 1993). More than half of the emergency and hospital lab expenses (Morey et al. 1995). department visits made in 1992 were for Hospitals can implement various strategies in nonurgent care (Baker, Stevens, and Brook order to reduce costs. Examples include dis- 1995). The competitive healthcare market has continuing unprofitable services, reducing the forced many community clinics to close, there- average length of stay, and substituting lower- by increasing the number of people seeking skilled hospital personnel. Two prior studies nonurgent care in hospital emergency rooms examined the relationship between hospital (Gordon 2000). In addition to the demand for costs and hospital performance. The first illus- nonurgent care, a high proportion of emer- trated a strong relationship between hospitals gency-department patients are uninsured acci- with low costs and an increased risk of negli- dent and trauma victims. These patients fre- gent injury (Burstin et al. 1993). In the second quently require extensive and costly inpatient study, Hartz and colleagues (1989) found a services, for which hospitals are often unable relationship between a hospital's financial sta- to collect payment (Arrington and Haddock bility and a lower mortality rate. 1990; Committee on Implications of For-prof-

Size of hospital stafT The size of the hospital it Enterprise in Health Care 1986).

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOsPITALs 21

Hospital teaching commitment. The avail- H2: Over the pastfifteen years, all things being ability of training opportunities for medical equal, there has been a convergence in the students, residents, physicians, and other med- nation 's short-term general hospitals as ical professionals is used to proxy a hospital's measured by community service outcomes.

provision of community service. Research has In recent years, however, many researchers shown that hospital training programs are cost- have argued that the cost-containment focus of ly to provide for several reasons (Duffy, the Prospective Payment System legislation

Ruseski, and Cavanaugh 2000; Kuttner 1999; has forced not-for-profit hospitals, like their Reuter 1999; GAO 1989; Morey et al. 1995; for-profit counterparts, to focus on cost-saving Grannemann, Brown, and Pauly 1986; Sloan, strategies in an effort to become more efficient Feldman, and Steinwald 1983). Hospitals with (Altman and Young 1993; Sorrentino 1989).

teaching programs treat a more costly mix of Particular concerns are that these hospitals will patients, maintain larger reserve margins, have cut back on unprofitable services including larger staffs, and offer more extensive treat- emergency rooms and teaching programs. ment options than do nonteaching hospitals Recently scholars have raised concern that not- (Thorpe 1988a; 1988b). These hospitals are for-profit hospitals have begun to discontinue typically not reimbursed the full cost of such these types of services in an effort to reduce care. Furthermore, the cost of attracting and their expenses (Morrisey, Wedig, and Hassan retaining high-quality physicians (Morey et al. 1996). Therefore, I hypothesize that the impact 1995; Custer and Wilke 1991) and maintaining of ownership type decreases as organizations the latest in medical technology (Friedman and adopt similar outcomes to meet regulatory Jorgensen 1994) accounts for higher costs at changes: teaching hospitals. Researchers find that these higher costs often act as a deterrent in hospi- H3: Changes in the regulatory environment

tahs' efforts to attract managed care patients have made hospital ownership less signifi- (Reuter 1999; Reuter and Gaskin 1998). cant in explaining the variation in efficien- cy outcomes in recent years compared with

previous years;

HYPOTHESES H4: Changes in the regulatory environment In recent years some have argued that have made hospital ownership less signifi-

changes in the health care arena have forced cant in explaining the variation in commu- not-for-profit hospitals to become indistin- nity service outcomes in recent years com- guishable from their for-profit counterparts. pared with previous years. Much of the health care literature suggests that when hospitals concentrate on efficiency they do so at the expense of community care. If METHODS indeed this is true, we should expect not-for-

profit hospitals to cut their costs by reducing Models expenses and cutting hospital staffing ratios. We should also expect not-for-profit hospitals Ordinary least squares regression. The data to reduce their provision of community service analysis occurred in two steps. First, for each by, for example, scaling back or closing their cross sectional data set I performed an ordi- emergency departments or reducing the num- nary least squares (OLS) regression with the ber of available teaching opportunities. In two efficiency and two community service other words, knowing the hospital's legal char- outcomes. Each cross-sectional data set differ- ter will not provide insight into the hospital's entiated between the three major hospital efficiency and community service operations. types, for-profit, not-for-profit, and public Therefore, I hypothesize that, hospitals. The OLS regression identified sig-

Therefore, I hypothesize that, nificant environmental and internal character- H1: Over the pastfifteen years, all things being istics that were included in the second step of

equal, there has been a convergence in the my analysis that utilized latent growth curve nation s short-term general hospitals as models (see Wright, Gronfein, and Owens measured by efficiency outcomes, 2000).

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

22 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Latent growth curve model. Using structural YU = aio + adt + ai2 (t2 / 00) + eit, equation models with AMOS 4, 1 examined the relationship between hospital outcomes and where hospital type over time (Arbuckle and Wothke

1997). In particular, I used a latent growth t = Year – 1980 curve model to model changes in the effect of hospital type (e.g., not-for-profit, public) on I analyzed five different models for each the efficiency and community service out- efficiency and community service outcome. comes over the research time frame The first model (see Figure 1) estimated the (1980-1994). The analysis used a merged data association between hospital type and hospital set that combined the four cross sectional data outcome at the baseline of the study, 1980. sets used in the OLS regression analysis. The Additionally, this model estimated the change hospitals that were included in this consolidat- in the association between hospital type and ed data were merged using a unique identifica- the hospital outcome over the fifteen year tion number. As a result of consolidating the research period. Model 2 builds upon model 1 cross sectional data sets, the new data set has a by including the environmental variables to fourth hospital type, those hospitals that estimate the total effect and the change in total change ownership during the study time frame effect with time. The model includes two types (e.g., from not-for-profit in 1980 to for-profit Of environmental variables, population charac- in 1990). teristics and market characteristics. Population In the model the intercept represents the c haracteristics include the unemployment rate,

research baseline period, and it is constant for the per capita income, and the proportion of each hospital across time (Duncan, Duncan, the population that is age 65 and over. Market and Li 1998). The slope term illustrates the characteristics include the proportion of the shape of the hospital's growth and is deter- population that is HMO members, the mined by repeated measures (Duncan et al. Herfindahl index, and the number of hospitals 1998). To complement the linear slope factor, in the state involved in mergers and consolida- each model includes an added growth factor to tions (e.g.with values from 1980, the study account for nonlinear growth. The present baseline).2 models utilize a quadratic growth term to cap- In models 3-5 (see table 2), I include vari- ture differences between the four groups of ables that are affected by hospital type and in hospital types (e.g., for-profit, not-for-profit, turn influenced the hospital efficiency and public, and hospitals that change ownership community service outcomes. In model 3, hos- type): "The added growth factor approach pro- pital facility characteristic variables, including vides a test of the difference in the growth rate hospital size and the ratio of full-time licensed between the [four] groups without having to practical nurses to full-time registered nurses, resort to the use of Lagrange Multiplier or are added to model 2. In model 4, hospital other model modification tests" (Duncan et al. practice characteristics including the surgical 1999:59; Muthen and Curran 1997). operations per adjusted inpatient day, the ratio Additionally, the inclusion of the quadratic of technology services and the average length growth factor improves the root mean square of stay-are added to model 2. Finally, model error of approximation (RMSEA) goodness of 5 combines model 2 and two variables repre- fit index for each of the latent growth curve senting hospital payor mix, Medicare and models in the analysis. Medicaid inpatient days as a proportion of

Each hospital in the analysis implicitly has adjusted inpatient days. its own intercept, linear coefficient, and qua- dratic coefficient, measured from the average one among the consistently for-profit hospi- Sample tals. Those implicit coefficients are taken as latent factors, indicated by the observed The analysis utilizes two types of secondary scores. The fixed factor loadings shown in data: (1) data from the American Hospital Figure 1 correspond to to, t', and t2 respective- Association and (2) data from the Area ly. The loadings on the latent quadratic coeffi- Resource Files. The American Hospital cient are divided by 100 to make the results Association and Area Resource File data sets readable. are two of the most commonly utilized sources

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 23

FIGURE 1. The Latent Growth Curve Model

0, 0, 0, 0,

1 ~~ ~~1 11 0 0 0 0

1980 1985 1990 1994 Expenses Expenses Expenses Expenses per an d Ser Adjusteinal And pe r Adjusted per Adjusted 11Te Admission Admission Admission thissuho

dn al al ps a96

Not-For-Profit Public Hospitalsa /

Note: heireerece, caeoyisfrpoit slospetals.

of hoptladdmgahcifrain(e ht fe eea eia n ugclsrie

1981). The American ~~~~HospitalsAscainptet hti es hntit as uhhs

defne aut-crehospitals astoehospiaspitals aritingisedfometlhahan

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

24 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

specialty hospitals (American Hospital were found. Finally, in the analysis utilizing Association 1992b). These acute-care hospitals the latent growth curve model, the missing are used in the analysis. Table 1 presents a cases were handled using a maximum likeli- breakdown by type of all U.S. hospitals in each hood estimation procedure with the AMOS 4 of the years covered by this research. program (Katula, Blissmer, and McAuley

The Area Resource File data, compiled by 1999; Arbuckle and Wothke 1997). The maxi- the Bureau of Health Professions, provide mum likelihood estimation procedure allows demographic data for each of the 3,248 coun- missing data to be accounted for in the analy- ties in the contiguous United States. Linking sis by calculating reliable standard errors internal hospital data from the American (Katula, Blissmer, and Mcauley 1999; Muthen, Hospital Association with demographic data Kaplan, and Hollis 1987). from the Area Resource File enables me to analyze the claims of a declining distinction between hospital types in light of data about Measures the hospitals' local environments. The means and standard deviations for the variables used Hospital expenses. Total expenses per in the analysis are presented in the appendix. adjusted admission is the first proxy for hospi-

The merge of the American Hospital tal efficiency. For purposes of standardization, Association and the Area Resource File data I divided the total hospital expense by the files created four separate files of internal and adjusted admissions (an aggregate figure environmental hospital information for every reflecting inpatient and outpatient days). short-term for-profit, not-for-profit, and public Size of hospital staff The number of full- general hospital at four points in time: 1980, time equivalent employees per adjusted daily 1985, 1990, 1994. The variables were speci- census is used as the second proxy for hospital fied identically for each year in the study to efficiency. Full-time equivalents employees ensure an accurate comparison of the relation- include all full and part time hospital personnel ships between the dependent variables and the with the exception of medical residents and independent variables for each of the years in other trainees. The number of full-time equiv- the study. alents employees is standardized using the

Missing data. Since the American Hospital adjusted census, that is, the average number of Association compiles self-reported hospital inpatients and outpatients receiving care dur- data on a voluntary basis from U.S. hospitals, ing a given twelve-month reporting period hospital officials have the option of not pro- (American Hospital Association 1994). viding the requested information. In the OLS Emergency room utilization. The ratio of regression models, those hospitals that did not emergency room visits per adjusted inpatient provide the appropriate data were left out of days is the first community service variable. In the analysis. However, the means and standard order to standardize emergency-room utiliza- deviations of the cases that were not included tion, the number of emergency-room visits was in the analysis were compared with the cases divided by adjusted inpatient days-an aggre- that were retained in the analysis, and no sig- gate figure reflecting both hospital inpatient nificant differences between the two groups and outpatient days.

TABLE 1. Number and Percent of U.S. Hospitals by Type and Year

Entire 1980 1985 1990 1994 Research Period

# % # % # % # % # %

Public Hospital 1,531 30% 1,433 28% 1,333 28% 1,193 27% 1,415 24% Not-for-profit 3,030 59% 3,063 60% 2,909 60% 2,706 62% 3,194 53% Hospital

For-profit hospital 556 11% 620 12% 590 12% 489 11% 697 12% Hospitals Changing 654 11% Ownership Type (e.g., from not-for- profit in 1980 to for-profit in 1990)

Total 5,117 100% 5,116 100% 4,832 100% 4,388 100% 5,960 100%

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 25

Hospital teaching commitment. The extent RESULTS of a hospital's teaching commitment is the sec-

ond proxy for community service. Research The Hospital 's Legal Charter

that focuses on hospitals' teaching status usu-

ally differentiates three levels of teaching corm- What effect does the hospital's legal charter mitment. Therefore, this study uses a weighted have on hospitals' efficiency and community ratio of the three levels of hospital teaching service outcomes? The hospital type intercept

status.3 coefficients in model 1 of Table 2 indicate that,

TABLE 2. Latent Growth Curve Model Regression Weights-Expenses per Adjusted Admission

Model 1 Model 2 Model 3 Model 4 Model 5

Intercept-Not-for-profit Hospital -.028 .021 -.065*** -.085*** .007

(.02) (.02) (.01) (.01) (.01) Intercept-Public Hospital -.258*** .122*** -.076*** -.106*** -.071***

(.02) (.02) (.02) (.01) (.02) Intercept-Hospitals Changing Ownership Type -.212*** .049* -.069*** -.091*** -.053**

(.02) (.03) (.02) (.02) (.02)

Yearly Slope-Not-for-profit Hospital -.014*** -.015*** -.015*** -.014*** -.016***

(.00) (.00) (.00) (.00) (.00) Yearly Slope-Public Hospital -.011*** .006*** -.012*** -.011*** -.013***

(.00) (.00) (.00) (.00) (.00) Yearly Slope-Hospitals Changing Ownership Type -.002 .563*** -.002 -.002 -.003

(.00) (.03) (.00) (.00) (.00) Quadratic Growth-Not-for-profit Hospital .119*** .138*** .122*** .122*** .122***

(.02) (.02) (.02) (.02) (.02) Quadratic Growth-Public Hospital .136*** -.003 .138*** .138*** .138***

(.02) (.00) (.02) (.02) (.02) Quadratic Growth-Hospitals Changing .047 .287*** .049* .048* .049* Ownership Type (.03) (.05) (.03) (.03) (.03)

Intercept-Unemployment Rate -.039* .004** .007*** .002

(.02) (.00) (.00) (.00)

Intercept-Proportion of Population Age 65 and Over -.033 -.156 -.381*** .310*

(.02) (.12) (.10) (.12)

Intercept-Number of Hospital Mergers in the State .004*** .004*** .004*** .003***

(.00) (.00) (.00) (.00) Intercept-Per Capita Income -.299*** .392*** .425*** .545***

(.02) (.03) (.02) (.03) Intercept-Proportion of Population HMO Members -.012*** .382*** .251*** .280***

(.00) (.05) (.04) (.05) Intercept-Hefindahl Index -.544*** -.186*** -.160*** -.298***

(.13) (.02) (.01) (.02) Intercept-Number of Hospital Beds (size) .001***

(.00) Intercept-Ratio of Full Time LPNs to Full Time RNs -.256***

(.03)

Intercept-Surgical Operations per Adjusted Inpatient Day -.424*** (.11)

Intercept-Average Length of Stay .022*** (.00)

Intercept-Ratio of Technology Services .547*** (.01)

Intercept-Medicare Inpatient Days as a Proportion -.364*** of Adjusted Inpatient Days (.04)

Intercept-Medicaid Inpatient Days as a Proportion .645*** of Adjusted Inpatient Days (.06) RMSEA .03 .04 .10 .09 .09 CFI 1.00 1.00 .99 .99 .99 N 5,960 5,960 5,960 5,960 5,960

*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

* p <.05 ** p< .01 *** p< .001

Note: The reference category is for-profit hospitals. Standardized errors are in parenthesis.

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

26 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

in 1980, not-for-profit (b = -.028) and public tors and hospital practice characteristics. The (b = -.258, p < .001) hospitals had lower for-profit hospitals are the omitted category expenses than their for-profit counterparts, the and have a value of zero for each time period. reference category. The slope coefficients for From Figure 2, we see that hospital type is a the hospital type indicate that, over time, not- more important predictor of hospital expenses for-profit (b = -.014, p < .001), and public in 1980 than in 1994 when we compare the hospitals (b = -.011, p < .00 1) have signifi- three hospital types against for-profit hospi- cantly lower expenses than their for-profit tals. This finding supports my first hypothesis; counterparts. Furthermore, the significant over the past fifteen years, all things being coefficients indicate that not-for-profit and equal, there has been a convergence in the public hospitals are reducing their expenses at nation's short-term general hospitals as mea- a more rapid growth rate than their for-profit sured by efficiency outcomes. counterparts. This illustrates the not-for-profit Table 3 presents the results of the five latent and public hospitals' focus on efficiency. growth curve models for the second efficiency

In model 2 in Table 2, I add baseline hospi- outcome, full-time equivalents employees per tal environmental factors to model 1. In model adjusted census. Once again, all five models 3, hospital facility characteristics are added to have relatively good fit indicators as indicated model 2. In model 4, hospital practice charac- by the root mean square error of approxima- teristics are added to model 2, and in model 5 tion and comparative fit index (Duncan et al. hospital payor mix indicators are added to 1999; Bentler 1990; Browne and Cudeck model 2. Hospital type remains a significant 1989). The not-for-profit hospital positive predictor of expenses per adjusted admission slope coefficients in models 1-5 (b = .017, b = with the inclusion of both environmental vari- .015, b = .007, b = .022, b = .0 14) indicate that ables and mediating factors. Based on the root over the research time frame, not-for-profit mean square error of approximation, and the hospitals had higher ratios of full-time equiva- comparative fit index, we find that all five lents employees per adjusted census than for- models have relatively good fit indexes profit hospitals. Furthermore the non-signifi- (Duncan et al. 1999; Bentler 1990; Browne cant slope coefficients in all five models indi- and Cudeck 1989). cate that over time not-for-profit and public

In Figure 2, I graphed the predicted values hospitals are experiencing parallel growth for the expenses per adjusted admission by rates in their staffing ratios. hospital type, adjusting for environmental fac- In Figure 3, I graphed the predicted values

FIGURE 2. The Predicted Values of Expenses Per Adjusted Admission by Hospital Type, Adjusting for Environmental and Hospital Practice Characteristics

0.05

160 0.001- C* CO 1 1982 1984 1986 1988 1990 1 -0.05

A -0.10

< L -0.15 CDE 0 0

-0.20

= F -0.25

-0.30 * not-for-profit – – public changed

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 27

FIGURE 3. The Predicted Values of Full-time Equivalent Employees per Adjusted Census by Hospital Type, Adjusting for Environmental and Hospital Payor Mix

L? s 0.50 0.40

0.30

8 W . 0.20

?co'5 010

= _ O.1d9 1982 1984 1986 1988 1990 1992 1

* not-for-prof ft * public A changed

for the full-time equivalent employees per adjusting for environmental factors and hospi- adjusted census by hospital type, adjusting for tal practice characteristics. The for-profit hos- environmental factors and hospital payor char- pitals are the omitted category and have a acteristics. From Figure 3, we see that over the value of zero for each time period. While not- research time frame, not-for-profit and for- for-profit hospitals provided a higher ratio of profit hospitals have similar ratios of full-time emergency-room care than their for-profit equivalent employees per adjusted census. counterparts in 1980, this difference narrows Again, we see that ownership is a more signif- by 1994. From Figure 4, we see that the ratio icant predictor of hospital efficiency outcomes of emergency room visits per adjusted inpa- in 1980 compared to 1994, providing evidence tient days for not-for-profit and for-profit hos- of the convergence between not-for-profit hos- pitals have begun to converge. While not-for- pitals and their for-profit counterparts. profit hospitals are still providing more emer- Therefore, these findings also support my first gency care, ownership is a more significant hypothesis regarding the claims of conver- predictor of this community service outcome gence among short-term general hospitals. in 1980 compared to 1994.

In Table 4 the results of the five latent The hospital ownership regression coeffi- growth curve models for the first community cients from the latent growth curve models for service outcome, ratio of emergency-room vis- the second community-service outcome, the its per adjusted inpatient day, are presented. hospital's teaching commitment, are shown in Analyzing the intercept coefficients in model 1 Table 5. These five models also have good fit of Table 4, we see that in 1980 not-for-profit indicators, as indicated by the root mean (b = .01) and public (b = .042, p < .001) hospi- square error of approximation and comparative tals had a higher ratio of emergency-room vis- fit index (Duncan et al. 1999; Bentler 1990; its per adjusted inpatient day than their for- Browne and Cudeck 1989). The not-for-profit profit counterparts. The difference is signifi- intercept coefficients in models 1 through 5 cant in models 3-5 for not-for-profit hospitals, (b = .124,p < .001, b = .143,p < .001; b =.045, and the difference is significant for public hos- p < .001; b = .070,p < .001; b = .137,p < .001) pitals in all five models. All five models have indicate that, in 1980, the not-for-profit hospi- relatively good fit indicators, as indicated by tals had a higher teaching commitment than the root mean square error of approximation their for-profit hospital counterparts. and comparative fit index (Duncan et. al. 1999; Analyzing the slope coefficients for the not- Bentler 1990; Browne and Cudeck 1989). for-profit hospitals in models 1 through 5, (b =

In Figure 4, I graphed the predicted values .003, p < .05; b = .003, p < .05; b = .003, p < for the ratio of emergency room visits by .01; b = .003,p < .01; b =.003,p < .05) we see adjusted inpatient days by hospital type, that, over the research time frame, the not-for-

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

28 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 3. Latent Growth Curve Model Regression Weights-Full Time Equivalent Employees per Adjusted Census

Model 1 Model 2 Model 3 Model 4 Model 5

Intercept-Not-for-profit Hospital -.016 .068 -.016 .122** .150** (.05) (.05) (.05) (.05) (.05)

Intercept-Public Hospital .041 .314*** .265*** .426*** .414*** (.06) (.06) (.06) (.05) (.06)

Intercept-Hospitals Changing Ownership Type .029 .247*** .207** .302*** .303*** (.07) (.07) (.06) (.06) (.06)

Yearly Slope-Not-for-profit Hospital .017 .015 .007 .022 .014 (.02) (.02) (.02) (.02) (.02)

Yearly Slope-Public Hospital .021 .019 .022 .041* .019 (.02) (.02) (.02) (.02) (.02)

Yearly Slope-Hospitals Changing Ownership Type .020 .018 .019 .036 .019 (.02) (.02) (.02) (.02) (.02)

Quadratic Growth-Not-for-profit Hospital -.214 -.199 -.197 -.202 -.195 (.11) (.11) (.11) (.11) (.11)

Quadratic Growth-Public Hospital -.276* -.261* -.258* -.263* -.254* (.12) (.12) (.12) (.12) (.12)

Quadratic Growth-Hospitals Changing -.299* -.285* -.287* -.292* -.283* Ownership Type (.14) (.14) (.14) (.14) (.14)

Intercept-Unemployment Rate -.002 -.004 -.006 .008 (.01) (.01) (.01) (.01)

Intercept-Proportion of Population Age 65 and Over -3.687*** -3.249*** -1.097** -7.358*** (.43) (.42) (.38) (.41)

Intercept-Number of Hospital Mergers in the State .001 .001 .001 -.003* (.00) (.00) (.00) (.00)

Intercept-Per Capita Income .632*** .379*** .433*** .926*** (.09) (.09) (.08) (.09)

Intercept-Proportion of Population HMO Members .850*** .948*** .701 *** .704*** (.18) (.18) (.15) (.17)

Intercept-Hefindahl Index -.237*** -.100 -.134** -.155** (.05) (.05) (.05) (.05)

Intercept-Number of Hospital Beds (size) .001*** (.00)

Intercept- Ratio of Full Time LPNs to Full Time RNs -.499*** (.09)

Intercept-Surgical Operations per Adjusted Inpatient Day 2.911 * (.43)

Intercept-Average Length of Stay -.058*** (.00)

Intercept-Ratio of Technology Services .464*** (.05)

Intercept-Medicare Inpatient Days as a Proportion 2.592*** of Adjusted Inpatient Days (.11)

Intercept-Medicaid Inpatient Days as a Proportion 1.345*** of Adjusted Inpatient Days (.19) RMSEA .06 .04 .09 .06 .09 CFI 1.00 1.00 .99 1.00 .99 N 5,960 5,960 5,960 5,960 5,960

* <.05 **p <.01 ***p <.001 Note: The reference category is for-profit hospitals. Standardized errors are in parenthesis.

profit hospitals are continuing to provide sub- dent that not-for-profit hospitals are not aban- stantially more teaching service than their for- doning their community-service mission as profit counterparts. they pursue efficiency strategies. Therefore,

Finally, Figure 5, the graph of the predicted these findings do not support my second values of the ratio of the hospital's teaching hypothesis. commitment by hospital type, adjusting for environmental and mediating factors illus- trates the absence of a convergence between The Hospital Regulatory Environment not-for-profit and for-profit hospitals over the research period. From these results it is evi- Do changes in the regulatory environment

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 29

FIGURE 4. The Predicted Values of the Ratio of Emergency Room Visits by Hospital Type, Adjusting

for Environmental and Hospital Practice Characteristics

*.I 0.12 E 0>

CL csX0.10

008

*u 0.02

> 0.00

1980 1982 1984 1986 1988 1990 1992 1994

* not-for-profit * public * changed

affect hospital efficiency and community ser- incurring these extra costs, they were increas-

vice outcomes? In model 1 of Table 2, the neg- ing their bottom line.

ative coefficients for not-for-profit (b = -.028) From the predicted values of expenses per

and public hospitals (b = -.258) from the latent adjusted admission by hospital type graphed in growth curve model indicate that in 1980 the Figure 2, we see that following the 1983 pas- for-profit hospitals, the reference category, sage of the Prospective Payment System legis- were benefiting more from the reimburse- lation the expenses per adjusted admission for

each hospital type converged with for-profit ments than their counterparts; that is, the for- hospitals. That is the various hospital types profit hospitals were incurring higher expens implemented similar strategies in an attempt to es. Prior to the passage of the Prospective reduce their expenses in order to work under

Payment System legislation, the administrators the newly imposed price controls. These find-

at for-profit hospitals knew that they would be ings support my third hypothesis: Changes in

reimbursed for these higher expenses. By the regulatory environment have made hospital

FIGURE 5. The Predicted Values of the Ratio of the Hospital's Teaching Commitment by Hospital Type, Adjusting for Environmental and Hospital Practice Characteristics

C I- 0.12

M2 E0.10

_o 0.08

C0.O .6

0.04

0.02

0.00 I I l I l

8 1980 1982 1984 1986 1988 1990 1992 1994

– not-for-proft * public A changed

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

30 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 4. Latent Growth Curve Model Regression Weights-Ratio of Emergency Room Visits per Adjusted Inpatient Day

Model 1 Model 2 Model 3 Model 4 Model 5

Intercept-Not-for-profit Hospital .010 .015 .035*** .047*** .020* (.01) (.01) (.01) (.01) (.01)

Intercept-Public Hospital .042*** .058*** .074*** .083*** .065*** (.01) (.01) (.01) (.01) (.01)

Intercept-Hospitals Changing Ownership Type .021 .036** .052*** .055*** .040*** (.01) (.01) (.01) (.01) (.01)

Yearly Slope-Not-for-profit Hospital .000 -.001 .001 .001 .000 (.00) (.00) (.00) (.00) (.00)

Yearly Slope-Public Hospital .005 .005 .004 .004 .005 (.00) (.00) (.00) (.00) (.00)

Yearly Slope-Hospitals Changing Ownership Type .012*** .012*** .011** .012*** .012*** (.00) (.00) (.00) (.00) (.00)

Quadratic Growth-Not-for-profit Hospital -.029 -.027 -.027 -.027 -.027 (.02) (.02) (.02) (.02) (.02)

Quadratic Growth-Public Hospital -.040 -.040 -.040 -.040 -.040 (.02) (.02) (.02) (.02) (.02)

Quadratic Growth-Hospitals Changing -.074** -.073** -.071* -.073** -.072** Ownership Type (.03) (.03) (.03) (.03) (.03)

Intercept-Unemployment Rate .008*** .009*** .008*** .009*** (.00) (.00) (.00) (.00)

Intercept-Proportion of Population Age 65 and Over -.933*** -1.056*** -.905*** -1.193*** (.08) (.08) (.08) (.08)

Intercept-Number of Hospital Mergers in the State .000 .000 .001 .000 (.00) (.00) (.00) (.00)

Intercept-Per Capita Income .008 .028 .034 .020 (.02) (.02) (.02) (.02)

Intercept-Proportion of Population HMO Members .177*** .149*** .182*** .173*** (.04) (.03) (.03) (.04)

Intercept-Hefindahl Index .025* -.012 -.017 .028** (.01) (.01) (.01) (.01)

Intercept-Number of Hospital Beds (size) .000*** (.00)

Intercept-Ratio of Full Time LPNs to Full Time RNs -.069*** (.02)

Intercept-Surgical Operations per Adjusted Inpatient Day .182* (.08)

Intercept-Average Length of Stay -.009*** (.00)

Intercept-Ratio of Technology Services -.127*** (.01)

Intercept-Medicare Inpatient Days as a Proportion .127*** of Adjusted Inpatient Days (.02)

Intercept-Medicaid Inpatient Days as a Proportion -.030 of Adjusted Inpatient Days (.04) RMSEA .01 .03 .09 .06 .09 CFI 1.00 1.00 .98 1.00 .99 N 5,960 5,960 5,960 5,960 5,960

*p <.05 **p< .01 *** <.001 Note: The reference category is for-profit hospitals. Standardized errors are in parenthesis.

ownership less significant in explaining the full-time equivalents employees per adjusted variation in efficiency outcomes. census. These findings also support my third

Figure 3, the graph of the predicted values hypothesis regarding regulatory change and of the full-time equivalent employees per convergence, as the hospitals exhibit similar adjusted census, indicates that following the growth rates in the wake of the Prospective 1983 passage of the Prospective Payment Payment System legislation. System legislation hospitals adopted similar When we examine the effect of regulatory strategies that resulted in not-for-profit and change on hospital community service out- for-profit hospitals having similar ratios of comes we find mixed results. Figure 4 is the

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 31

graph of the predicted values of the ratio of community service outcomes in recent years emergency room visits per adjusted inpatient compared with previous years.

day (the graph of model 3 in Table 4). In 1985,

following the 1983 implementation of the Prospective Payment System legislation, we The External Environment and Mediating see that the not-for-profit hospitals' ratio of Factors (Internal Hospital Characteristics) emergency room services began to decline. By 1994 the not-for-profit hospitals were provid- What effects do the external environment ing slightly more emergency room care than and hospital internal characteristics have on

their for-profit counterparts. These findings hospital efficiency and community service

support my fourth hypothesis: Changes in the outcomes? Figure 6 illustrates that much of the regulatory environment have made hospital differences in expenses per adjusted admission ownership less significant in explaining the between not-for-profit and for-profit hospitals variation in community service outcomes in are attributable to baseline differences in the recent years compared with previous years. demographics of the service population and

When we analyze the effect of regulatory differences in the hospital payor mix. By con- change on the not-for-profit hospitals' teach-

ing commitment we see different results. From .troln 'for these actors, ft d ifens Figure 5, the graph of the predicted values of between not-for-profit and for-profit hospitals Fiur . th grap of th prdce vlue of in this efficiency measure decreased over the the hospitals' teaching commitment (the graph intch of model 4 in Table 5), we can see that since period. 1980 not-for-profit hospitals have had a larger In Figure 7, we compare the effects of con- teaching commitment than their for-profit trolling for baseline differences in the hospi-

counterparts. Furthermore, the not-for-profit tal'sexternal environmentcand hospitalepractice hospitals did not reduce their teaching com- characteristics including the number of surgi- mitments following the passage of the cal operations per adjusted inpatient day, the Prospective Payment System legislation. average length of stay and the ratio of technol- Therefore, these findings do not support my ogy services on the hospitals' community fourth hypothesis: Changes in the regulatory service outcome. Figure 7 indicates that differ- environment have made hospital ownership ences in the for-profit and not-for-profit hospi- less significant in explaining the variation in tals' provision of emergency care increased

FIGURE 6. Not-for-profit Hospitals Compared to For-profit Hospitals

0.04

T X 0.02 –

< 0.00 -|||I|/

] as 1 980 1982 1984 1986 1988 1990 2 1994 ? -0.02- ~0

-0.04

X 0 -0-06

Mb S -0.08

* not-for-profit without adjustments

U not-for-profit adjusting for baseline demographics and hospital payor ray

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

32 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR TABLE 5. Latent Growth Curve Model Regression Weights-Ratio of Hospital's Teaching

Commitment

Model 1 Model 2 Model 3 Model 4 Model 5

Intercept-Not-for-profit Hospital .124*** .143*** .045*** .070*** .137*** (.01) (.01) (.01) (.01) (.01)

Intercept-Public Hospital .063*** .139*** .092*** .104*** .123*** (.01) (.01) (.01) (.01) (.01)

Intercept-Hospitals Changing Ownership Type .048*** .000*** .058*** .063*** .094*** (.01) (.01) (.01) (.01) (.01)

Yearly Slope-Not-for-profit Hospital .003* .003* .003** .003** .003* (.00) (.00) (.00) (.00) (.00)

Yearly Slope-Public Hospital .002 .002 .002 .002 .002 (.00) (.00) (.00) (.00) (.00)

Yearly Slope-Hospitals Changing Ownership Type .000 .000 .001 .001 .000 (.00) (.00) (.00) (.00) (.00)

Quadratic Growth-Not-for-profit Hospital -.015* -.016* -.015 -.015 -.016* (.01) (.01) (.01) (.01) (.01)

Quadratic Growth-Public Hospital -.014 -.014 -.013 -.014 -.014 (.01) (.01) (.01) (.01) (.01)

Quadratic Growth-Hospitals Changing -.006 -.006 -.007 -.007 -.007 Ownership Type (.01) (.01) (.01) (.01) (.01)

Intercept-Unemployment Rate -.001 -.003*** -.001 -.003* (.00) (.00) (.00) (.00)

Intercept-Proportion of Population Age 65 and Over -.509*** -.066 -.099 -.190* (.10) (.07) (.08) (.09)

Intercept-Number of Hospital Mergers in the State .000 .001* .000 .000 (.00) (.00) (.00) (.00)

Intercept-Per Capita Income .110*** -.038* .014 .114*** (.02) (.02) (.02) (.02)

Intercept-Proportion of Population HMO Members -.126** -.025 -.155*** -. 136*** (.04) (.03) (.03) (.04)

Intercept-Hefindahl Index -.151 -.015 -.039*** -.146*** (.01) (.01) (.01) (.01)

Intercept-Number of Hospital Beds (size) .001 *** (.00)

Intercept-Ratio of Full Time LPNs to Full Time RNs -.086*** (.02)

Intercept-Surgical Operations per Adjusted Inpatient Day -.720*** (.08)

Intercept-Average Length of Stay .001** (.00)

Intercept-Ratio of Technology Services .507*** (.01)

Intercept-Medicare Inpatient Days as a Proportion -.167*** of Adjusted Inpatient Days (.03)

Intercept-Medicaid Inpatient Days as a Proportion .385*** of Adjusted Inpatient Days (.04) RMSEA .03 .02 .09 .06 .09 CFI 1.00 1.00 .99 1.00 .99 N 5,960 5,960 5,960 5,960 5,960

* p< .05 ** p< .01 *** p <.001 Note: The reference category is for-profit hospitals. Standardized errors are in parenthesis.

when I controlled for hospital baseline demo- that hospitals in areas with lower concentration graphics and hospital practice characteristics. (generally areas with large numbers of hospi-

Table 6 provides the OLS regression coeffi- tals) have higher expenses than those hospitals cients for both efficiency outcomes, the in areas with higher hospital concentration (see expenses per adjusted admission, and the full- Menke 1997). The positive and significant time equivalent employees per adjusted cen- coefficients for the unemployment rate (equa- sus. A negative coefficient on the Herfindahl tions 1-4, b = 23, p < .001; b = 21, p < .001; Index for all four study years (equations 1-4, b = 41,p < .001; b = 63,p < .001) indicate that b = -384, p < .001; b = -397 p < .001; b = hospitals in areas with high unemployment are -561; p < .001; b = -522; p < .001) indicates more likely to incur the higher expenses per

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 33

FIGURE 7. Not-for-profit Hospitals Compared to For-profit Hospitals

0.06

0.04-

0.02 –

0 *-

C) C a

0.00 –

CD (~1982 1984 1986 198899

-*- not-for-pofit hos ptds without austmens

-U not-for-pofit hos piths alus tin for to; hine cbtu~ ics hos pith

ltice dwra~tis tac

adjusted admission for treating patients with- p < .001; b = .l128, p < .001; b = .O64, p < .001; out insurance. b = .056, p < .00 1). Finally, my results indicate

In Table 7, the OLS regression coefficients that the higher the hospitals teaching commit- for the community service outcome, the ratio ment, the higher the proportion of Medicaid of emergency-room visits per adjusted inpa- patients (b = .294, p < .001; b = .069, p < .05; tient day are provided. In equations 1-4, we b = .150, p< .001; b= .084, p< .01). see that an increase in the proportion of the population aged 65 and over has a negative effect on emergency-room utilization (b = DISCUSSION

-.35,p< .01;b= -1.21,p< .001,b=-1.11; p < .001; b = -1.70, p < .00 1). Equations 14 The External Environment and Mediating indicate a positive relationship between the Factors UInternal Hospital Characteristics) unemployment rate and the ratio of emer- gency-room visits per adjusted inpatient day. My research demonstrates the importance of In other words, an increase in the unemploy- controlling for the external hospital environ- ment rate increases emergency room utiliza- ment and mediating factors. Without control- tion (b =.Ol, p < .001; b = .Ol, p < .001; b = ling for these factors we lose sight of how they .02, p < .001; b =.01, p < .00 1). influence the community service and efficien-

In equations 14 in Table 6 we see that hos- cy outcomes. For example, when we analyze pitals with higher ratios of licensed practical efficiency outcomes we see that the variation nurses to registered nurses had lower expenses between not-for-profit and for-profit hospitals per adjusted admission (b =-8 18, p <.001; b= decreases over the research period when we -1 l29,p <.001; b=-lO87,p <.001; b=-322, control for the environmental and mediating p < .001). Hospitals with longer lengths of stay factors. Alternatively, when looking at commu- have higher expenses per adjusted admission nity service outcomes we see that when con- (b = 85.l7,p < .001; b = i3.38,p < .001; b = trolling for these factors the difference 74.25, p < .001; b = 75.55, p < .001). Finally, a between not-for-profit and for-profit hospitals' higher ratio of technology services available at provision of community service is more pro- the hospital results in higher total expenses per nounced-not-for-profit hospitals are providing adjusted admission (b = 1,009, p < .001; b = more community service than their for profit

1,326,p <.001; b= 1,185,p <.001; b =1,116, counterparts. p <. .01). Additionally, the results of controlling for

In Table 7, equations 5-8 we see that hospi- the external environment and mediating fac- tals with a high ratio of technology services tors are consistent with prior research and thus have a higher teaching commitment (b = .142, substantiate my models. For instance, my

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

34 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

00

0)

0 ) -00ONV C >' 00 C 'I0 >W) ) O 00 r- 0 — – — – > C ooo C> 0 – -" – o o ei er-e 0V

0 C ) * n * * * *- *f I

~~ CO 0~ 0~0 t0-r–00 0 cf) ICf)er ei0r- 0 Ne-0CC)- 0 en -N en 0 >0

C) ~~~ – -eI0~~ 00 000e el- — – –O i- "ei 0ri – 00 – -7 > oO 0 )e > – t ol~r- r- f .- q0

en C> 00 00 00 0 CCfC) r 00e r-O N ~ ' – CO 0 0~' t ~~ o o- en 0 6n~eO ~ ~

– ) – ) – WC

0 rm c n 0 1 1 1 I t m0-

C~~~*) * * * * *~~~~~~0 * *t* *

0-e~~000cN 0~0 0~0~ e~0~ 0o co 00c co1 0

M>1 zC~ 0- –

0) * ~~'~~* ,-.* ,-.*A * * * *co~ '~ 400C- Nco ~0 ~ C)0- N c

0~0 00.00 0~0 – e.e N 0tN Ne C 0 ~ 0 t0- I)0W C) N 0 00 0- 0WC~' O W~ CO

~~~~~~~~ – ~~~~~~~~~~~~~~~~~~~~~~Z – ' -~~~~~~~ – – TC -~~~~~~~~~~~~~~~~~~4 = C' 0~~~~~~~~~~~~~C Ot

F-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C * * * * * * * * *~~~~~~~~~~~~~c co*.

– ~~~~0* * * * * * * * *~~~~~~~~~~~~~C *4tl) a

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 35

0 0

< b ~~~* c .o 0 * O. x, C) O In

*0 * 0 00

E a m * ~* ~-* * rA 0000-~0

0 *

00 ~~~~~~~ * * ~ ~ ~ ~ 0 C)0 ~00 000 0C0-

g-i 0,0. -0.0. b 00 , 000.W

* * * *

E x, | o 0- 0000 oo 6 o C)- 11Orl~t n 0 C) el 00

, o o. ? m m m .t ~~~.non o N~~~~~~C oc _c < ll2t C N4 ooCo 2) o) I.-

e~~~Q~~W -00 0

O wi o o o x x 0 ? 0 x 00 o oly v) lt 11 W) r

<~~~~~~~~0 I 4 06 ci40 r * * * * * * e N * * * * * *

= .-o 0 0. O. ? ~N — 0 0 cE 0- O~~ 00 00~00 C)0 rC0 W) W 0 0

S .O~~~~~~~~~~~. * * * * *O M 0 -00~ C -rr4

_00 M C < 2 <}Nt- 11 *~t ?) *t ??0 00 . .

WQ – C1 – 0 –

*

1 7 * * c E – * .c * ,y t x – B v' _ 3 00 R ?~ o 0 = 0 co~~~~~~~~~~~~~-

u~~~~~~~~~~~ 4 .; U E ~~~~~~~~~~~~~~~o .tE4vS ~

N ~ ~~ U .3 0 o-os i;t

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

36 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

I oo ~* * * * * * -o t- * * * * * * * * * *

o ) 0ec -00c000C Q ~ Ne c 00 c C C) I0D – ~ 00 W C 000

r-00M 0 – 0 00 00 -0 00 00

A co * * * * * ** * * * 0n . * * * * * * * * ***

* a – * ~-* ~* *~- * * * *

00000

* 00 C -0c – 000 0 0 -0 0

In M

. 4 o m * * * * * * * * .v .~ 4 o * * * * * * * ***

o m ~ -* ~-* ~-* , * *~- *~- * ~ -* * *

~~ 00000000 C C> 0 -0M 00 00 0- M 0 ,0 0* CIO

o o * * * * * * * * * * 0 _o * * * * * * * * * *

0 ~ 0 00 s~~~~~~~) M co ooooo C oo

* * * * * * * *

' X : . * * * *' * -.1 * * *

* ~ -* * ** * * *~ *~- ~ * ~-*

-o *~~~~~~~~~~~~~~~~~oC

* * * * * * * * * *

<SO~ m0.0.0.00,0. NO. 00. n0. 000.00 – oo0-

0~~~~~~~~~~~~~~~

* * * * * * * * * *

0 – * * * * * * * * * *

~~~ .~~~~~~~-~4

'-4 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~-

d~~~~~~~~~~~~~~~~~t >

0~~~~~~~~~~~~~~~~~~~~~~~~~~~0

– 04 W)

Q~~~ E g~ ;

O~~~~' *t .l O 0. O O OOO O 0. .OOOOO

*E~~~ ;

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 37

00

00 (C cq W) C) r__ m W) C C C> C> In W

to r O m C) C) Q Q Q ON m Q Q rn

cq 00

W) 00

C) C) >rn cq Q C14 00 C14 m C) C) C) C> C> C> C> C> C> C> C> C> W)

W)

C>

00 C> rn C> rn C> C> rn rn C> C> C> rn C> C> C) C) C) 0 C> C> C> C> w) C14

W)

o 00 C> tn C14 C> r__ W) C14 C> C> C) C) C) C) C) C) C C C> C> 00

cd

C> m c> G c, W) (& m m C> C> C> W) C43 aN C> C> C> C> C> C> C> C> C> C> C> C> C14

C14 00

IN

r__ aN C> W) M C> 00 m M 00 W) 00 C C C> C,

C14

4S

tj 0

O rn aN C> C> c", -t C> '-t C14 C> C> rn (:=> 0 0 C> 00 rn C> C> C> C) C) C> C> C> C> C> cq 0 – 00

W) cr

N

> 40. C>

C th

to 0

0 W)

to CI,

to= r. g 0 0-4 9 >_. r. C) = 0 0 u u E

Cd Q 0 0

W)

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

38 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

results indicate an increase in per capita The 1983 Prospective Payment System leg- income is also associated with higher expenses islation marked the first time in the history of per adjusted admission. This finding is consis- the hospital industry that hospitals were forced tent with previous findings that for-profit hos- to accept price controls. Prior to the passage of pitals choose to locate in affluent areas this legislation, not-for-profit hospitals had lit- (Norton and Staiger 1994) and that hospital tle incentive to look for more efficient ways to offerings are often a reflection of the wealth in provide patient care. Medicare and other third the hospital community (Sloan and Becker party insurance payors were reimbursing hos- 1981). Consistent with previous studies, my pitals for virtually all the costs incurred for an results indicate that people aged 65 and over episode of patient care. Therefore, one could seek emergency-room care at a lower rate than conclude that passage of the Prospective other sectors of the population (Tyrance, Payment System legislation prompted the not- Himmelstein, and Woolhandler 1996). for-profit hospitals to pursue efficiency strate- Alternatively, my findings show that emer- gies. My research supports these assertions as gency department utilization increases when we see not-for-profit and for-profit hospitals there is an increase in the unemployment rate converging in terms of their efficiency out- in the hospital community. For most people, comes following the passage of the health insurance is coupled with their employ- Prospective Payment System legislation. These ment (Altman, Reinhardt, and Shields 1998). findings are consistent with recent work by Previous research has established that emer- neo-institutionalists who suggest that efficien- gency rooms are sources of care for people cy is shaped by the regulatory environment; without regular health care access (Albrecht et that is, organizations do not entertain certain al. 1996). efficiency solutions until policy change occurs

My research substantiates prior studies that (see Scott 1995). indicate substituting lower paid licensed prac- Alternately, using the latent growth curve tical nurses for higher paid registered nurses model results, an assessment of the conver- enables hospitals to decrease their labor costs gence in hospital types for the community-ser- (Woolhandler and Himmelstein 1997). vice outcomes provides mixed evidence for the Additionally, the results corroborate prior claims that hospitals are reducing community research, as my findings indicate that it is care in an effort to reduce their expenses. In expensive for hospitals to acquire and maintain fact, over the fifteen-year period in this analy- the latest technology (Steiner et al. 1997; sis, not-for-profit hospitals continued to pro- Newhouse 1993). However, teaching hospitals vide substantially more community service as want to offer the latest in medical technology measured by the not-for-profit hospitals' to train new physicians (Friedman and teaching commitment compared to their for- Jorgensen 1994). Finally, my research supports profit counterparts, and they continue to pro- prior research that indicates the higher the hos- vide slightly more emergency care. These find- pitals' teaching commitment the higher the ings are not consistent with neo-institutional proportion of Medicaid patients (Morey et al. theory and leave many questions unanswered. 1995; Thorpe 1988a, 1988b). Why would convergence between these two

hospital types occur in terms of efficiency but not in terms of community service? Some may

The Hospital Regulatory Environment hypothesize that community service is tightly coupled with a hospital's legal charter. In fact,

Neo-institutionalists assert that isomor- the mission statements of many not-for-profit phism occurs when regulatory changes force hospitals specifically mention the hospital's different types of organizations to pursue sim- responsibility to its community. The findings ilar strategies (Fligstein 1991, 1985). The pre- suggest that not-for-profit hospitals can simul- sent research analyzes if the regulatory taneously pursue both efficiency and commu- changes generate similar strategies that result nity service strategies. in similar organizational outcomes. If indeed neo-institutionalists' predictions are accurate, the hospitals' legal charters would be a less The Hospital a Legal Charter significant predictor of hospital outcomes fol- lowing legislative change in the industry. Prior research indicates that when hospitals

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 39

behave in a manner that does not coincide with their legal charters there are many policy implications. Furthermore, researchers empha- size that hospital efficiency and community service are inversely related and assert that hospital efficiency comes at the price of com-

munity care (Kane 1993). As explained earlier, all not-for-profit hospitals receive the same property and revenue tax exemptions, regard- less of their location. Critics question whether it is cost effective for communities and all lev- els of government to continue to support hos- pitals through tax exemptions, which are esti-

mated to be in the billions of dollars (Kane 1993; Fox and Schaffer 1991). Many critics insist that not-for-profit hospitals should not be eligible for tax exemptions if indeed a con- vergence is occurring in the hospital industry. Others argue that we should not discontinue the exemption; instead we should hold hospi- tals accountable for their actions and penalize hospitals that are not acting responsibly.

The present research empirically analyzes the hypotheses of convergence between not- for-profit and for-profit hospitals in the con- text of both internal and environmental factors. The research covers a fifteen-year period. Its scope thus enables me to more accurately assess the presence or absence of convergence in the hospital industry. Prior research was lim- ited in both time frame and environmental fac- tors. My findings clearly demonstrate that not- for-profit and for-profit hospitals are converg- ing in terms of efficiency outcomes. However, I find mixed evidence that not-for-profit hos- pitals are abandoning their community-service missions as they become more efficient. In fact, I find evidence that not-for-profit hospi- tals are simultaneously pursuing efficiency and community service. We should find comfort that, to date, not-for-profit hospitals are giving back to their communities. However, further research is necessary to find incentives to ensure not-for-profit hospitals do not slip off the so-called "community service path" as they face increasing cost-containment pressures from their environments.

Finally, we have begun to see an increase in

the sale of not-for-profit hospitals to for-profit

entities (Shactman and Altman 1998; Claxton et al. 1997). As not-for-profit hospitals become more efficient, as evident in this research, they also become more attractive to for-profit own- ers. Additionally, as this research indicates,

for-profit hospitals are more likely to provide

less community service. Policies must be in

place to ensure that communities receive bene- fits as their "more attractive" not-for-profit

hospitals are purchased by for-profit organiza- tions and the community service mission is replaced by a profit motive.

NOTES

1. Small variations in Medicare capital pay- ments per case accounted for hospital out- liers including geographic locations, case

mix, larger teaching hospital, and whether the hospital is the sole community facility (Kauer, Silvers, and Teplensky 1995).

2. I use a Herfindahl Index calculated by uti- lizing inpatient days. This enables me to

capture the hospital's actual market share

rather than Hospital A's available beds as a share of all beds in Hospital A's market (Zwanziger, Melnick, and Eyre 1994). In this analysis, the Herfindahl Index ranges from .02 to 1. Hospitals in areas with low market concentration (a large number of hospitals) have a Herfindahl Index approaching .02; those that have no local competitors have a Herfindahl Index of 1.

3. This study uses a weighted ratio of the three levels of hospital teaching status. Membership in the Council of Teaching Hospitals signifies the highest level of teaching commitment, medical-school affiliation represents the second-highest commitment to medical training, and the presence of an internship or residency pro- gram represents the lowest level of teaching commitment. Hospitals may offer all three, two, one or none of the specified teaching opportunities.

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

40 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

~~~~~~~~~~~W, o W) oo~ Cq r- en cl W) 0 – Co ,o "t o = oo W oo Tt

I te

I11 C1 W) C1 b o tob^

m ~~~~~~~~~~~~~~~~~~t

W) C1 cq "t 0 C oNO O N

an 0 ::: ::0:::

~~~~~~~~~~~~~~~~~~~~~~W N en O 0 00N " o Oc 't t Z q CO en rq I ^t en O O= "00t O

N tt00.m 000^ en ^

Z = CN 0 cn – C1 b, O r- C n 1 t en O C> C> O t –

r- C1 W) 90 CNo M0 e 00 r- C1 O co 9c "t rl 00 W0 00 b ,t t 1 't 't m co O. O = O C On O O> O C't O> 00 0 m O

cn e QON 00 sW

00 ^ i

T 11 F XXebObN C Z v- O. – CN O> C N O. 00 O. cn r- C O. O.4 C1 O q W .t

_ CZ C n C> C> C> O OCO en CO O) O O> O- O O> O> O> C O > O- O> en

o ,.NE

: 11 | = s e g o o t 2 t o > 2 = c | o~~~W

X Ut l_ V *E = ? v E Q Q e Wo = ? =~~~~~~~~~~~~~~~~~~~~~~~~C

m < .t E tq vo X | U; < 8; t m t S e~~~~~~~~~~~~~~~~~~~~~~~(

= X W E X-* j ; > = * = ; = t~~~~~~~~~~~~~~~~~~~~~~~~~~C

a ;.EO O~~~~~~o?Q < <.O<.O?Q??U~~~~~~~~~~t v vtgQ-O ?~~~~~C

t;~~~~~~~~C U O to- M? Ov? AeeSeoS

v~~~~o n v :v0 ~ EsEs:< ~ ~ mt~ X; ? Ne m?b oo tm bX > 4-

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 41

REFERENCES

Albrecht, Gary L., David Slobodkin, and Robert J.

Rydman. 1996. "The Role of Emergency

Departments in American Health Care." pp. 289-318 in Research in the Sociology of Health

Care, edited by Jennie J. Kronenfeld. Greenwich,

CT: JAI Press.

Altman, Stuart H., Uwe E. Reinhardt, and Alexandra E. Shields. 1998. "Healthcare for the

Poor and Uninsured: An Uncertain Future." Pp. 1-24 in The Future US. Healthcare System:

Who Will Care for the Poor and Uninsured?, edited by Stuart H. Altman, Uwe E. Reinhardt,

and Alexandra E. Shields. Chicago: Health Administration Press.

Altman, Stuart H. and David Shactman. 1997.

"Why Should We Worry about Hospitals' High

Administrative Costs?" The New England

Journal ofMedicine 336:798-99. Altman, Stuart H. and Donald A. Young. 1993. "A

Decade of Medicare's Prospective Payment

System-success or Failure?" Journal of American Health Policy: 1 1-19.

American Hospital Association. 1992a. Hospital

Mergers and Consolidations: 1980-1991.

Chicago: American Hospital Association.

American Hospital Association. 1992b. 1992 AHA Guide. Chicago: American Hospital Association.

American Hospital Association. 1994. 1994 AHA Guide. Chicago: American Hospital Association.

Anderson, Gerard F and Linda T. Kohn. 1996. "Hospital Employment Trends in California,

1982-1994." Health Affairs 15:152-158. Arbuckle, James L. and Werner Wothke. 1997.

Amos 4.0 User's Guide. Chicago: SmallWaters Corporation.

Arrington, B. and C. C. Haddock. 1990. "Who Really Profits from Not-for-profits?" Health Services Research 25:291-304.

Baker, David W, Carl D. Stevens, and Robert H. Brook. 1995. "Determinants of Emergency Department Use by Ambulatory Patients at an

Urban Public Hospital." Annals of Emergency Medicine 25:311-16.

Baron, J.N., FR. Dobbin, and PD. Jennings. 1986. "War and Peace: The Evolution of Modern Personnel Administration in U.S. Industry." American Journal of Sociology 92:350-83.

Barrett, Marilyn W 1995. "Downsizing: Doing it Rationally." Nursing Management 26:24-29.

Becker, Edmund R. and Frank Sloan. 1985. "Hospital Ownership and Performance." Economic Inquiry 23:21-35.

Becker, Edmund R. and Bruce Steinwald. 1981. "Determinants of Hospital Casemix Complexity." Health Services Research 16:439-58.

Bentler, Peter. 1990. "Comparative Fit Indexes in

Structural Models. Psychological Bulletin 107:238-46.

Blumenthal, David and Joel S. Weissman. 2000. "Welling Teaching Hospitals To Investor-owned Hospital Chains: Three Case Studies: Sales to For-profit Chains May Strengthen Financial Viability Without Harming Social Missions." Health Affairs 19:158-66.

Brown, E. Richard 1983. "Public Hospitals on the Brink: Their Problems and Their Options." Journal of Health Politics, Policy and Law 7:927-44.

Browne, Michael W. and Robert A. Cudeck. 1989. "Single Sample Cross-validation Indicies for Covariance Structures." Multivariate Behavioral Research 24:445-55.

Burns, Lawton R. 1990. "The Transformation of the American Hospital: From Community Institution toward Business Enterprise." Comparative Social Research 12:77-112.

Burstin, Helen R., Stuart R. Lipsitz, I. Steven Udvarhelyi, and Troyen A. Brennan. 1993. "The Effect of Hospital Financial Characteristics on Quality of Care." Journal of the American Medical Association 270:845-49.

Chang, Cyril F and Howard P. Tuckman. 1988. "The Profits of Not-for-profit Hospitals." Journal of Health Politics, Policy and Law 13:547-64.

Claxton, Gary, Judith Feder, David Shactman, and Stuart Altman. 1997. "Public Policy Issues in Nonprofit Conversions: An Overview." Health Affairs 16(2):9-28.

Coddington, Dean C., Lowell E. Palmquist, and William V Trollinger. 1985. "Strategies for Survival in the Hospital Industry." Harvard Business Review: 129-38.

Committee on Implications of For-profit Enterprise in Health Care. 1986. "Access to Care." Pp. 97-126 in For-Profit Enterprise in Health Care, edited by D. H. Gray. Washington, DC: National Academy Press.

Custer, WS. and R.J. Wilke. 1991. "Teaching Hospital Costs: The Effect of Medical Staff Characteristics." Health Services Research 25:831-57.

DiMaggio, Paul J. and Walter W Powell. 1983. "The Iron Cage Revisited: Institutional Isomorphism and Collecive Rationality in Organizational Fields." American Sociological Review 48:147-60.

Dowd, Timothy J. and Frank Dobbin. Forthcoming. "Was There a Market Before Antitrust?: Public Policy and Railroad Strategy in Early America." in Constructing Markets and Industries, edited by J. P. and. M. Ventresa. New York: Pergamon Press.

Duffy, Sarah Q., Jane E. Ruseski, and Sean Cavanaugh. 2000. "Graduate Medical Education Costs in Nonacademic Health Center Teaching Hospitals : Evidence from Maryland." Medical Care Research and Review 57:3-23.

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

42 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Duncan, Terry E., Susan C. Duncan, and Fuzhong Li. 1998. "A Comparison of Model- and Multiple Imputation-based Approaches to Longitudinal Analyses with Partial Missingness." Structural Equation Modeling 5:1-21.

Duncan, Terry E., Susan C. Duncan, Lisa A. Strycker, Fuzhong Li, and Anthony Alpert. 1999. An Introduction to Latent Variable Growth Curve Modeling: Concepts, Issues and Applications. London: Lawrence Erlbaum Associates.

Fligstein, Neil. 1985. "The Spread of the Multidivsional Firm, 1919-79." American Sociological Review 50:377-91.

Fligstein, Neil. 1991. "The Structural Transformation of American Industry: An Institutional Account of the Causes of Diversification in the Largest Firms, 1919-1979." Pp. 311-36 in The New Institutionalism in OrganizationalAnalysis, edit- ed by Walter W. Powell and Paul J. DiMaggio. Chicago: The University of Chicago Press.

Flood, Ann Barry and Mary L. Fennel. 1995. "Through the Lenses of Organizational Sociology: The Role of Organizational Theory and Research in Conceptualizing and Examining Our Health Care System." Journal of Health and Social Behavior (Extra Issue): 154-69.

Fox, Daniel M. and Daniel C. Schaffer. 1991. "Tax Administration as Health Policy." Journal of Health Policy, Politics, and Law 16:251-63.

Friedman, Leonard H. and Jane Jorgensen. 1994. "Physician's Influence on the Decision to Acquire Magnetic Resonance Imagers in Acute Care Hospitals." International Journal of Technology Assessment in Health Care 10:667-74.

General Accounting Office. 1989. "Medicare: Indirect Medical Payments are Too High." Washington, DC: Government Printing Office.

General Accounting Office. 1993. "Report to the chairman, Subcommittee on Health for Families and the Uninsured, Committee on Finance, US. Senate: Emergency Departments Unevenly Affected by Growth and Change in Patient Use." Washington DC: Government Printing Office.

Gentry, William M. and John R. Penrod. 1998. "The Tax Benefits of Not-For-Profit Hospitals." Working Paper No. 6435 National Bureau of Economic Research, Cambridge, MA.

Ginzberg, Eli. 1988. "For-profit Medicine: A Reassessment." The New England Journal of Medicine 319:757-61.

Gordon, James. 2000. "Emergency Care As Safety Net." Letter to the editor in Health Affairs 19:277.

Grannemann, Thomas W. and Randall S. Brown. 1986. "Estimating Hospital Costs: A Multiple- output Analysis." Journal of Health Economics 5:107-127.

Gray, Bradford H. 1991. The Profit Motive and

Patient Care: The Changing Accountability of

Doctors and Hospitals. Cambridge, MA: Harvard University Press.

Hadley, Jack and Katherine Swartz. 1989. "The

Impact of Hospital Costs Between 1980 and 1984 of Hospital Rate Regulation, Competition, and Changes in Health Insurance Coverage." Inquiry 26:35-47.

Hartz, Arthur J., Henry Krakauer, Evelyn M. Kuhn, Mark Young, Steven J. Jacobsen, Greer Gay, Larry Muenz, Myron Katzoff, R. Clifton Bailey,

and Alfred A. Rimm. 1989. "Hospital Characteristics and Mortality Rates." The New England Journal of Medicine 321:1720-25.

Herzlinger, Regina E. and William S. Krasker. 1987. "The Profits from Nonprofits?" Harvard Business Review:93-105.

Homer, Carl G., Douglas D. Bradham, and Mark Rushefsky. 1984. "To the Editor, Investor-owned and Not-for-profit Hospitals: Beyond the Cost

and Revenue Debate." Health Affairs: 133-36. Hultman, Cheryl I. 1991. "Uncompensated Care

before and after Prospective Payment: The Role of Hospital Location and Ownership." Health Services Research 26: 614-22.

Kane, Nancy. 1993. "Report on the Financial Resources of Major Hospitals in Boston." Department of Health and Hospitals, Boston MA.

Katula, Jeffrey A., Bryan J. Blissmer, and Edward McAuley. 1999. "Exercise Intensity and Self- efficacy Effects on Anxiety Reduction in Healthy, Older Adults." Journal of Behavioral Medicine 22:223-47.

Kauer, Robert T., J.B. Silvers, and Jill Teplensky. 1995. "The Effect of Fixed Payment on Hospital Costs." Journal of Health Politics, Policy and Law 20:303-27.

Kuttner, Robert. 1999. "Managed Care and Medical Education." The New England Journal of Medicine 341:1092-96.

Light, Donald W 1986. "Corporate Medicine for Profit." Scientific American 255:38-45.

Marmor, Theodore R., Mark Schlesinger, and Richard W.Smithey. 1987. "Nonprofit Organizations and Health Care." Pp. 221-40 in The Nonprofit Sector: A Research Handbook, edited by W W Powell. New Haven, CT: Yale University Press.

McCarthy, C. 1988. "DRGs-Five Years Later." New England Journal of Medicine 318:1683-86.

Menke, Terri J. 1997. "The Effect of Chain Membership on Hospital Costs." Health Services Research 32:177-96.

Morey, Richard C., Yasar Ozcan, Donna L. Retzlaff- Roberts, and David J. Fine. 1995. "Estimating

the Hospital-wide Cost Differentials Warranted for Teaching Hospitals." Medical Care 33:531-32.

Morrisey, Michael A., Wedig Gerald J., and

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

THE DISTINCTION BETWEEN FOR-PROFIT AND NOT-FOR-PROFIT HOSPITALS 43

Mahmud Hassan. 1996. "Do Nonprofit Hospital

Pay Their Way?" Health Affairs 15:132-144. Muthen, B. and P. Curran. 1997. "General Growth

Modeling of Individual Differences in Experimental Designs: A Latent Variable Framework for Analysis and Power Estimation." Psychological Methods 2:371-402.

Muthen, B., D. Kaplan, and M. Hollis. 1987. "On Structural Equation Modeling with Data that are not Missing Completely at Random." Psychometrika 2:431-62.

Newhouse, Joseph P. 1993. "An Iconoclastic View of Health Cost Containment." Health Affairs 12:152.

Norton, Edward C. and Douglas 0. Staiger. 1994. "How Hospitals Ownership Affects Access to Care for the Uninsured." RAND Journal of Economics 25:171-85.

Reinhardt, Uwe E. 1998. "Employer-based Health Insurance: R.I.P." Pp. 325-73 in The Future

US. Healthcare System: Who Will Carefor the Poor and Uninsured?, edited by Stuart H.

Altman, Uwe E. Reinhardt, and Alexandra E. Shields. Chicago: Health Administration Press.

Reuter, James A. 1999. "Patters of Specialty Care: Academic Health Centers and the Patient Care Mission." Report for Task Force on Academic Health Centers. The Commonwealth Fund, New York, NY.

Reuter, James A. and Darrell J. Gaskin. 1998. "The Role of Academic Health Centers and Teaching Hospitals in Providing Care for the Poor." Pp. 151-65 in The Future US. Healthcare System:

Who Will Carefor the Poor and Uninsured? edit- ed by Stuart H. Altman, Uwe E. Reinhardt, and Alexandra E. Shields. Chicago: Health Administration Press.

Rogers WH., D. Draper, K.L. Kahn, E.B. Keeler, L.V Rubenstein, J. Kosecoff, and R.H. Brook. 1990. "Quality of Care Before and After Implementation of the DRG-based Prospective Payment System. A Summary of Effects." Journal of the American Medical Association 15:1989-94.

Rushing, William A. 1976. "Profit and Nonprofit Orientations and the Differentiations- Coordination Hypothesis for Organizations: A Study of Small General Hospitals." American Sociological Review 41:676-91.

Scott, W Richard. 1995. "Introduction: Institutional Theory and Organizations." Pp. xi-xxiii in The Institutional Construction of Organizations: International and Longitudinal Studies, edited by W Richard Scott and Soren Christensen. Thousand Oaks, CA: Sage Publications.

Shactman, David and Stuart H. Altman. 1998. "The Impact of Hospital Conversions on the Healthcare Safety Net." Pp. 189-203 in The Future US. Healthcare System: Who Will Care for the Poor and Uninsured, edited by Stuart H.

Altman, Uwe E. Reinhardt, and Alexandra E. Shields. Chicago: Health Administration Press.

Shortell, Stephen M., Ellen M. Morrison, Susan L. Hughes, Bernard Friedman, James Coverdill, and Lee Berg. 1986. "The Effects of Hospital Ownership on Nontraditional Services." Health Affairs (Winter):97-111.

Sloan, Frank A. and Edmund R. Becker. 1981. "Internal Organization of Hospitals and Hospital Costs." Inquiry 18:224-239.

Sloan, Frank A., Roger D. Feldman, and A. Bruce Steinwald. 1983. "Effects of Teaching on Hospital Costs." Journal of Health Economics 2:1-28.

Sorrentino, Elizabeth A. 1989. "Hospital Mission and Cost Differences." Hospital Topics 67:22-25.

Starr, Paul. 1982. The Social Transformation of American Medicine USA: Basic Books.

Steiner, Claudia A., Neil R. Powe, Gerard F. Anderson, and Abhik Das. 1997. "Technology Coverage Decisions by Health Care Plans and Considerations by Medical Directors." Medical Care 35:472-89.

Stevens, Rosemary. 1989. In Sickness and in Wealth: American Hospitals in the Twentieth Century USA: Basic Books.

Thorpe, Kenneth E. 1988a. "The Use of Regression Analysis to Determine Hospital Payment: The Case of Medicare's Indirect Teaching

Adjustment." Inquiry 25:219-31. Thorpe, Kenneth E. 1988b. "Why are Urban

Hospital Costs So High? The Relative Importance of Patient Source of Admission, Teaching, Competition, and Case Mix." Health Services Research 22:821-36.

Tyrance, Patrick H., David U. Himmelstein, and Steffie Woolhandler. 1996. "US Emergency Department Costs: No Emergency." American Journal ofPublic Health 86:1527-31.

Waldholdtz, Michael 1982. "To Keep Doors Open, Non-profit Hospitals Act Like Businesses." in The Wall Street Journal Volume cc. N0127 Page 1, col. 6.

Watt, J. Michael, Robert A. Derzon, Steven C. Renn, Carl J. Schramm, James S. Hahn, and George D. Pillari. 1986. "The Comparative Economic Performance of Investor-owned Chains and Not- for-profit Hospitals." The New England Journal ofMedicine 314:89-96.

Williams, Robert M. 1996. "The Costs of Visits to Emergency Departments." The New England Journal of Medicine 334:642-46.

Woolhandler, Steffie and David U. Himmelstein. 1997. "Costs of Care and Administration at For- profit and Other Hospitals in the United States." The New England Journal of Medicine 336:769-74.

Wright, Eric R., William P. Gronfein, and Timothy J. Owens. 2000. "Deinstitutionalization, Social Rejection, and the Self-esteem of Former Mental

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms

44 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Patients." Journal of Health and Social Behavior 41:68-90.

Zwanziger, Jack, Glenn A. Melnick and Lisa

Simonson. 1996. "Differentiation and Specialization in the California Hospital Industry." Medical Care 34:361-372.

Sharyn J. Potter is Assistant Professor of Sociology at the University of New Hampshire. Her research focuses on the hospital industry, health care policy and social inequalities in health care. She is currently working on several research projects. One study analyzes the role of for-profit hospitals in rural areas. Another longitudinal study examines the careers of women hospital administrators.

This content downloaded from ������������199.94.80.42 on Sun, 24 Sep 2023 07:59:25 +00:00������������

All use subject to https://about.jstor.org/terms