Accounting for teaching hospitals' higher costs and what to do about them

Health Affairs

Volumn 22, Issue 6, 2003, Pages 126-129

  Newhouse, Joseph ^a  

^a HARVARD SCHOOL OF PUBLIC HEALTH   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

CLASSIFICATION; ECONOMICS; HOSPITAL COST; MEDICARE; NOTE; ORGANIZATION; STATISTICS; TEACHING HOSPITAL; UNITED STATES; UNIVERSITY HOSPITAL;

ACADEMIC MEDICAL CENTERS; HOSPITAL COSTS; HOSPITALS, TEACHING; MEDICARE; ORGANIZATIONAL OBJECTIVES; UNITED STATES;

EID: 1542644860     PISSN: 02782715     EISSN: 02782715     Source Type: Journal    
DOI: 10.1377/hlthaff.22.6.126     Document Type: Review

References (13)

1. I derive this by taking the difference between the values for cost per case at AHC hospitals and community hospitals in Koenig's Exhibit 3, Model 2 ($3,310), and apportioning it between wages and case-mix ($1,229) and IME and other mission-related costs ($2,100) (base costs are essentially equal). Exhibit 2 shows that the IME and Other categories within teaching hospitals are approximately equally split between IME and standby costs. Although standby costs are somewhat larger, community hospitals appear to have some standby costs as well (the $260 value in Exhibit 3). Research costs appear to be under 10 percent of the difference. That base costs are similar suggests that the institutions are equally efficiently managed, which accords with having managers at both types of institutions face similar incentives to manage efficiently. I thus agree with Koenig and colleagues' implicit conclusion that differences in costs are not attributable to differences in managerial inefficiency or effort.
2. Medicare Payment Advisory Commission, Report to the Congress: Medicare Payment Policy (Washington: MedPAC, March 2003), 279. Although "major teaching hospitals" are roughly the same as Koenig and colleagues' "academic health centers (AHCs)," these margin data are not strictly comparable to Koenig and colleagues' data, for two reasons: (1) The margin data include all parts of the hospital, such as the outpatient department and the skilled nursing facility, whereas Koenig and colleagues' data include only inpatient operations; and (2) the margin data include all hospitals, whereas Koenig and colleagues' comparison group is appropriately urban hospitals with more than 100 beds. Rural hospitals have higher margins, probably because they face less competition. Moreover, both the margin data and Koenig and colleagues' data exclude physicians who are not employees of the hospital, but there may be subsidies of various types in both directions that a full accounting would incorporate. Comparing only private hospitals in each of the three categories somewhat narrows the gap in margins, but a substantial gap remains.
3. There are a number of econometric issues with Koenig and colleagues' methods that I do not take up here, but see J.A. Rogowski and J.P. Newhouse, "Estimating the Indirect Costs of Teaching," Journal of Health Economics 11, no. 2 (1992): 153-171.
4. Koenig and colleagues find that research accounts for 3.6 percent of AHCs' cost (Exhibit 4), or $308. Any true subsidy to research is less than this, and there may be no subsidy at all, because AHCs receive monies for the care of patients in clinical trials, as well as research grants.
5. J.S. Weissman et al., "Market Forces and Unsponsored Research in Academic Health Centers," Journal of the American Medical Association 281, no. 12 (1999): 1093-1098.
6. J.P. Newhouse and G.R. Wilensky, "Paying for Graduate Medical Education: The Debate Goes On," Health Affairs (Mar/Apr 2001): 136-147.
7. G.S. Becker, Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education, 3d ed. (Chicago: University of Chicago Press, 1993).
8. As the number of residents grew, the coefficient on the number of residents in the equation determining the IME adjustment fell, as Koenig and colleagues note. The fall was sufficiently large that the product of the number of residents and the estimated amount each added to per case cost was roughly constant. (This, however, does not consider DME cost.) If residents are not causally related to cost, the IME measure in Koenig and colleagues' cost decomposition should not be identified with teaching. Furthermore, DME costs should not be segregated from other costs in the regression analysis.
9. Such costs were 5.2 percent at major private teaching hospitals, 4.2 percent at other private teaching hospitals, and 4.5 percent at private nonteaching hospitals. MedPAC, A Data Book: Healthcare Spending and the Medicare Program (Washington: MedPAC, 2003), 93. Differences among the three corresponding types of public hospitals are much greater. Because the amount a private hospital spends on uncompensated care is under its control, these differences among types of hospitals do not prove that demands for uncompensated care are greater at teaching hospitals, but they are consistent with that hypothesis.
10. The 1 percent figure comes from expanding the number of categories into which Medicare inpatients are grouped from the current 500 diagnosis-related groups (DRGs) to 1,420. MedPAC, Report to the Congress: Selected Medicare Issues (Washington: MedPAC, June 2000), 77. The coefficient on the resident-to-bed variable is little changed by this expansion.
11. Center for the Evaluative Clinical Sciences, Dartmouth Medical School, The Quality of Medical Care in the United States: A Report on the Medicare Program, The Dartmouth Atlas of Health Care 1999 (Chicago: American Hospital Association Press, 1999);
12. M.R. Chassin et al., "Does Inappropriate Use Explain Geographic Variation in the Use of Health Care Services? A Study of Three Procedures," Journal of the American Medical Association 258, no. 18 (1987): 2533-2537;
13. and M.R. Chassin and R.W. Galvin, "The Urgent Need to Improve Health Care Quality," Journal of the American Medical Association 280, no. 11 (1998): 1000-1005.