Hospital conversions, margins, and the provision of uncompensated care

Health Affairs

Volumn 19, Issue 6, 2000, Pages 187-194

  Thorpe, Kenneth E ^a   Florence, Curtis S ^b   Seiber, Eric E ^b  

^a EMORY UNIVERSITY   (United States)

^b TULANE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ARTICLE; COMMUNITY CARE; COMPARATIVE STUDY; ECONOMICS; HEALTH CARE FACILITY; HOSPITAL; HOSPITAL ORGANIZATION; HUMAN; NON PROFIT HOSPITAL; ORGANIZATION AND MANAGEMENT; PUBLIC HOSPITAL; REGRESSION ANALYSIS; STATISTICAL MODEL; STATISTICS; UNITED STATES; UTILIZATION REVIEW;

HEALTH FACILITY PLANNING; HOSPITAL RESTRUCTURING; HOSPITALS, PROPRIETARY; HOSPITALS, PUBLIC; HOSPITALS, VOLUNTARY; HUMANS; MODELS, ECONOMIC; OWNERSHIP; REGRESSION ANALYSIS; UNCOMPENSATED CARE; UNITED STATES;

EID: 0005452774     PISSN: 02782715     EISSN: None     Source Type: Journal    
DOI: 10.1377/hlthaff.19.6.187     Document Type: Article

References (24)

1. American Hospital Association, Annual Survey of Hospitals (Chicago: Health Forum LLC, 1990-1998).
2. J.M. Mann et al., "A Profile of Uncompensated Hospital Care, 1983-1995," Health Affairs (July/Aug 1997): 223-232; and I. Fraser, J. Narcross, and P. Kralovec, "Medicaid Shortfalls and Total Unreimbursed Hospital Care for the Poor, 1980-1999," Inquiry 28, no. 4 (1991): 385-392.
3. J.M. Mann et al., "A Profile of Uncompensated Hospital Care, 1983-1995," Health Affairs (July/Aug 1997): 223-232; and I. Fraser, J. Narcross, and P. Kralovec, "Medicaid Shortfalls and Total Unreimbursed Hospital Care for the Poor, 1980-1999," Inquiry 28, no. 4 (1991): 385-392.
4. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
5. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
6. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
7. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
8. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
9. K.R. Desai, C. VanDeusen Lukas, and G.J. Young, "Public Hospitals: Privatization and Uncompensated Care," Health Affairs (Mar/Apr 2000): 167-172; B.H. Gray, "Hospital Ownership Form and Care of the Uninsured," in The Future of the U.S. Health Care System Who Will Care for the Poor and Uninsured? ed. S.H. Altman, U.E. Reinhardt, and A.E. Shields (Chicago: Health Administration Press, 1998), 207-222; G.J. Young, K.R. Desai, and C.V. Lukas, "Does the Sale of Nonprofit Hospitals Threaten Health Care for the Poor?" Health Affairs (Jan/Feb 1997): 137-141; J. Needleman, J. Lamphere, and D. Chollet, "Uncompensated Care and Hospital Conversions in Florida," Health Affairs (July/Aug 1999): 125-133; G.J. Young and K.R. Desai, "Nonprofit Hospital Conversions and Community Benefits: New Evidence from Three States," Health Affairs (Sep/Oct 1999): 146-155; and E.C. Norton and D.O. Staiger, "How Hospital Ownership Affects Access to Care for the Uninsured," RAND Journal of Economics 25, no. 1 (1994): 171-185.
10. See, for example, B.A. Weisbrod, The Nonprofit Economy (Cambridge, Mass.: Harvard University Press, 1988); and J. Newhouse, "Toward a Theory of Nonprofit Institutions: An Economic Model of a Hospital," American Economic Review 60, no. 3 (1970): 64-74. There are also models with different predictions for the differences between for-profit and not-for-profit hospitals than the one outlined here. For example, M. Pauly and M. Redish, "The Not-for-Profit Hospital as a Physician's Cooperative," American Economic Review 63, no. 1 (1973): 87-99, shows that when a hospital's decision making is controlled by physicians who maximize their net income, the not-for-profit hospital will behave like a for-profit firm in the short run. In the end, the differences between not-for-profit and for-profit hospitals are an empirical question.
11. See, for example, B.A. Weisbrod, The Nonprofit Economy (Cambridge, Mass.: Harvard University Press, 1988); and J. Newhouse, "Toward a Theory of Nonprofit Institutions: An Economic Model of a Hospital," American Economic Review 60, no. 3 (1970): 64-74. There are also models with different predictions for the differences between for-profit and not-for-profit hospitals than the one outlined here. For example, M. Pauly and M. Redish, "The Not-for-Profit Hospital as a Physician's Cooperative," American Economic Review 63, no. 1 (1973): 87-99, shows that when a hospital's decision making is controlled by physicians who maximize their net income, the not-for-profit hospital will behave like a for-profit firm in the short run. In the end, the differences between not-for-profit and for-profit hospitals are an empirical question.
12. See, for example, B.A. Weisbrod, The Nonprofit Economy (Cambridge, Mass.: Harvard University Press, 1988); and J. Newhouse, "Toward a Theory of Nonprofit Institutions: An Economic Model of a Hospital," American Economic Review 60, no. 3 (1970): 64-74. There are also models with different predictions for the differences between for-profit and not-for-profit hospitals than the one outlined here. For example, M. Pauly and M. Redish, "The Not-for-Profit Hospital as a Physician's Cooperative," American Economic Review 63, no. 1 (1973): 87-99, shows that when a hospital's decision making is controlled by physicians who maximize their net income, the not-for-profit hospital will behave like a for-profit firm in the short run. In the end, the differences between not-for-profit and for-profit hospitals are an empirical question.
13. We rely on the same data used in previous work by the Prospective Payment Assessment Commission (now the Medicare Payment Advisory Commission), as well as other published research. See, for example, Mann et al., "A Profile of Uncompensated Hospital Care." The AHA data are unique in their comprehensiveness, although the financial data are not audited. Counts of hospitals reporting information to the AHA, as well as imputations of missing data, are described annually in AHA publications such as Hospital Statistics. In general, approximately 90 percent of all hospitals sun-eyed respond to the AHA survey. Our analysis of converting and nonconverting hospitals is limited to those included in the survey. The AHA data also may underrcport information from for-profit hospitals. For example, approximately three-quarters of all for-profit hospitals responded to the AHA survey. Thus, our analysis undercounts the number of forprofit converting hospitals. Moreover, the external validity of the analysis could be compromised if the characteristics of for-profit hospitals that do and do not report data to the AHA differ. To quantify the magnitude and empirical relevance of any undercount, we compared the AHA survey with Medicare hospital cost report data. We found 369 hospitals included the Medicare cost report data that were not included in the AHA survey data. Of this, 113 (31 percent of all missing hospitals) were for-profit hospitals. However, we found no significant differences in the Medicare data concerning cost per admission, revenues per admission, or total beds among for-profit hospitals that did and did not report information in the AHA survey. While our counts may understate the actual number of for-profit conversions, the external validity of the results does not appear to be compromised.
14. Our analysis included any hospital conversion during the years 1990 through 1997. Of the 431 conversions in our data, approximately 20 percent occurred during 1997. On average, however, converting hospitals in the sample have 3.1 years of preconversion data and 2.3 years of postconversion data. Our regression analysis included several measures of hospital conversion. The analysis reported in the text uses a dummy variable for each type of conversion. This measures the average impact of the conversion on each of our dependent variables. In addition, we estimated other regression specifications to determine if the effect of conversions varies over time. We included a series of dummy variables for each year post conversion. This allowed us to examine whether the impact of the conversion varied over time post conversion (that is, did the effects vary in the first year of conversion, second year, and so on?). We could not reject the hypothesis that the effect of the conversions on each of our dependent variables was equal for each year post conversion. We also estimated a regression model using only the conversions that transpired through 1995. We then estimated separate effects on uncompensated care for the year of conversion and one and two years post conversion. The results were nearly identical to those presented in Exhibit 3. That is, the impact of our conversions on the outcome measures of interest did not increase or decrease significantly over time. Thus, we have no reason to expect that the inclusion of conversions with only one year of postconversion experience would bias the coefficient estimates in our regression analysis. To estimate the impact of the imputation of financial information in the AHA data, we conducted a sensitivity analysis by dropping all hospitals that have imputed values for all years of our sample (approximately 8.7 percent of all hospital-year observations) and including an indicator equal to 1 if the dependent variable was imputed for remaining observations. The coefficients on the conversion variables in these regressions were nearly identical to those using the full sample, and the coefficient on the variable marking an imputed value was never significantly different from zero. As a result, the full regression estimates using all observations are used in producing the adjusted means reported in Exhibit 3.
15. Mann et al., "A Profile of Uncompensated Hospital Care"; and Fraser et al., "Medicaid Shortfalls."
16. Mann et al., "A Profile of Uncompensated Hospital Care"; and Fraser et al., "Medicaid Shortfalls."
17. We used both county-level and metropolitan statistical area (MSA)-based definitions of competition. Our measure relied on the Herfindahl index of total adjusted admissions. Our results were the same using both market definitions. The full regression results are available upon request from the authors: e-mail Ken Thorpe, .
18. We conducted a Hausman specification test to determine if the fixed-effects model was appropriate (as opposed to random effects). The null hypothesis that the random-effects model is appropriate was rejected at the 1 percent level. The fixed-effects model allows us to use each hospital as its own control. C. Hsiao, Analysis of Panel Data (Cambridge: Cambridge University Press, 1990), 11-68. Thus, the coefficient estimate reported for the dummy variables for each type of conversion may be interpreted as the change in uncompensated care after conversion compared with hospitals that did not convert. We also tested to detect whatever changes in uncompensated care after conversions were time dependent (that is, did the changes differ in the first, second, or third year after conversion?). No measurable time-dependent effects were evident.
19. These numbers were calculated by regressing the outcome of interest on a vector of indicator variables for ownership type and year. This controls for the changing composition of hospitals by ownership type and makes nominal values such as expenses per admission comparable across years. The predicted values for each ownership type from the regression results are reported in Exhibit 2. The significance test was conducted by testing for equality of the coefficients on ownership type.
20. The independent variables in the regression are the conversion indicator variables described in the methods section, Herfindahl index of the number of beds for the market, number of beds per capita in the market, percentage of the population enrolled in HMOs, per capita income, average hospital wage in the market, percentage under age thirty, percentage uninsured, percentage unemployed, percentage black, percentage Hispanic, percentage high school graduates, percentage college graduates, indicator variables for population size of market, and year indicator variables. The full regression results are available from the authors.
21. Derived from our regression results. We predict uncompensated care as a percentage of total expenses assuming hospitals that did and did not convert. The change in the ratio of uncompensated care to total expenses for each hospital was multiplied by each hospital's total expenses. The $400,000 figure is an average across the hospitals in the sample.
22. These results are similar to those found by Jack Needleman and colleagues in Florida but differ from the three-state sample (California, Texas, and Florida) examined by Kamal Desai and colleagues. The different results are likely an artifact of differences in methodologies used (that is, fixed-effects panel estimates using each hospital as its own control in our case) and the national scope of our study. See, for comparison, Desai et al., "Public Hospitals"; and Needleman et al., "Uncompensated Care and Hospital Conversions in Florida."
23. These results are similar to those found by Jack Needleman and colleagues in Florida but differ from the three-state sample (California, Texas, and Florida) examined by Kamal Desai and colleagues. The different results are likely an artifact of differences in methodologies used (that is, fixed-effects panel estimates using each hospital as its own control in our case) and the national scope of our study. See, for comparison, Desai et al., "Public Hospitals"; and Needleman et al., "Uncompensated Care and Hospital Conversions in Florida."
24. We use the regression results to predict uncompensated care and total adjusted admissions with and without the conversion dummy variable set to 1. For each hospital's predicted level of uncompensated care, we divide it by each hospital's cost per adjusted admission.