MODELING INTERNAL DECISION MAKING PROCESS: AN EXPLANATION OF CONFLICTING EMPIRICAL RESULTS ON BEHAVIOR OF NON-PROFIT AND FOR-PROFIT HOSPITALS

Abstract

I. INTRODUCTION The prevalence of for-profit hospitals has increased in the United States over the past few decades. The emergence of for-profit hospitals that often compete in markets with not-for-profit hospitals has generated vigorous academic and policy debates about the implications of the distinct legal and organizational for-profit and non-profit forms on hospital behavior. The main issues arise from differences in monetary and non-pecuniary incentives for administrators and physicians that are associated with organizational form, including variations in tax treatment, access to capital markets, and governance structures. These debates center on two possible outcomes related to organizational form and ownership mix. One is whether these differences lead to observable differences in revenues and costs as well as the quantity, quality, and mix of services provided by for-profit and non-profit hospitals (NPHs). The other possible outcome is whether competition from for-profit hospitals in the same markets as NPHs coupled with persistent cost-containment efforts of public and private insurers, results in a form of convergence where NPHs behave similarly to for-profit hospitals despite differences in organizational form. This notion of convergence that has arisen in the literature on performance in the hospital industry has called into question both academically and for policy makers the tax benefits that NPHs enjoy on the grounds that NPHs are indistinguishable from for-profit hospitals and fail to provide community benefits at a level sufficient to justify the subsidies. (1) The ongoing debate about whether NPHs behave differently from for-profit hospitals has given rise to a substantial empirical literature seeking to inform the debate. Schlesinger and Gray (2006) and Rosenau (2003) comprehensively reviewed the empirical evidence on the performance differences between for-profit and NPHs. Performance is evaluated along several dimensions including economic measures like cost per admission, revenue or charge per admission, technical efficiency, non-pecuniary measures like quality, provision of care to indigent patients, and trustworthiness of organizational practices. Both reviews conclude that ownership-related differences in hospital behavior and outcomes are mixed. Schlesinger and Gray reviewed 162 empirical studies comparing non-profit and for-profit hospitals and nursing homes along the dimensions of economic performance, quality of care, and indigent patients' access to care. (2) In terms of economic performance, 18 studies found no significant differences between for-profit and NPHs; 10 studies found an advantage for profit-making hospitals (PMHs); and 30 found an advantage for NPHs. Twenty-one studies reported no significant differences between for-profit and NPHs in terms of quality of care, whereas 4 studies found higher quality of care in for-profit hospitals, and 19 studies found higher quality of care in NPHs. Rosenau synthesized the results of approximately 75 peer-reviewed studies published between 1985 and 2001 along quality, cost, access, and charity care dimensions. (3) In terms of cost, 23 studies found NPHs to be superior, 5 studies found for-profit hospitals to be superior, and 9 studies were inconclusive. Twelve studies found that NPHs provided higher quality care, 3 studies found quality of care to be higher in for-profit hospitals, and 9 studies found no differences in quality. Their finding that evidence of ownershiprelated differences in cost, quality, profits, pricing policies, technical efficiency, access to care, and service offerings vary greatly across empirical studies is not surprising because the studies themselves vary in terms of data used and statistical methods. The empirical approach taken in the majority of the studies is to estimate a reduced-form model of hospital characteristics that typically includes an indicator variable for ownership status as a control variable. …