|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Previous estimates of administrative costs in U.S. hospitals have been based on figures for California, and nationwide extrapolation has been controversial. If the costs of bureaucracy are high, major policy reforms may yield substantial savings.
Methods We obtained detailed data on hospital expenses for fiscal year 1990 from reports submitted to Medicare by 6400 hospitals. We calculated each hospital's administrative costs by summing expenses in the following Medicare cost-accounting categories: administrative and general, nursing administration, central services and supply (excluding the purchase cost of supplies), medical records and library, utilization review, and the salary costs of the employee benefits department. We classified costs in most other categories as clinical. Some small categories of expenses (e.g., gift shop) were excluded from both our clinical and administrative groupings, and for others (e.g., plant operations), a proportional share was allocated between the two groupings.
Results Nationwide, administration accounted for an average of 24.8 percent of each hospital's spending in fiscal 1990. Average hospital administrative costs ranged from 20.5 percent in Minnesota to 30.6 percent in Hawaii. Administrative salaries accounted for 22.4 percent of the average hospital's salary costs. Administrative costs were similar in states with high and low rates of enrollment in health maintenance organizations (HMOs).
Conclusions Hospital administrative costs in the United States are higher than previous estimates and more than twice as high as those in Canada. Greater enrollment in HMOs, with more competitive bidding by hospitals for managed-care contracts, an important element of proposed managed-competition health care reforms, does not seem to lower hospital administrative costs.
Previous estimates have relied on a few published figures from California, where one in five hospital dollars went for administration in 19871,2,3,4,5. But nationwide extrapolation from these data has been controversial6,7,8. This paper presents data on the administrative costs of virtually all acute care hospitals in the United States.
Methods
Medicare requires that participating hospitals file detailed reports classifying all their expenses into standard categories. In response to a request under the Freedom of Information Act, the Health Care Financing Administration supplied us with a computerized copy of the data from Worksheet A of the Medicare Cost Report for each of 6400 hospitals. Each hospital reported expenses for the fiscal year that began in calendar year 1989.
Table 1 shows the major categories used to classify expenses in the Medicare Cost Report. For most categories, hospitals report total expenses as well as salary expenses. However, Medicare requires hospitals to adjust the total cost figures (but not the salary figures) to reflect true hospital costs more accurately. For our municipal hospital, for instance, city hall rather than a hospital department manages parking and pensions. The parking and pension costs attributable to hospital operations would appear in the adjusted total cost figures on the Medicare Cost Report, but would be excluded from the salary figures. Hence, our analyses of the overall cost of administration are more reliable than those based solely on salary figures.
|
The costs of the hospital's physical plant are not allocated to individual services or hospital functions in the Medicare Cost Report. For example, the reported costs of a coronary care unit or billing department would include salaries and supplies but not the costs of building, maintaining, and equipping the unit. We assumed that the proportion of the hospital's physical plant that houses administration is the same as administration's share of overall costs (excluding the costs of the physical plant). In our analysis of total administrative costs we therefore allocated 24.8 percent of the costs for capital, plant operations, and maintenance and repairs to administration. In analyzing salary costs, we allocated 22.4 percent of the salaries for plant operations and maintenance and repairs to administration (the Medicare Cost Report attributes no salary costs to capital).
The rubric "employee benefits" on the Medicare Cost Report subsumes the expense of administering benefits as well as actual disbursements for health insurance and other fringe benefits. We classified the salary costs of the employee benefits department as entirely administrative. All other employee benefits costs were allocated between the administrative and clinical categories in the same manner as the physical-plant costs.
We calculated the proportion of costs attributable to administration in each hospital by summing the total costs in the administrative categories plus the allocated share of the physical-plant and employee benefits costs and dividing the result by total hospital costs less excluded categories. We then calculated the mean share spent for administration by hospitals in each state, in the District of Columbia and Puerto Rico, and nationwide.
Results
Table 1 shows the major categories used to classify expenses in the Medicare Cost Report and our designation of each cost category as administrative, clinical, mixed administrative and clinical (allocated proportionally as described above), or neither administrative nor clinical (excluded from both numerator and denominator in the analysis). Table 1 also shows the proportion of the average hospital's costs reported within each category.
Administration accounted for an average of 24.8 percent of hospitals' spending nationwide in fiscal 1990. Average hospital administrative costs ranged from 20.5 percent in Minnesota to 30.6 percent in Hawaii (Table 2).
|
|
Discussion
In fiscal 1990 administration accounted for nearly one quarter of U.S. hospital spending, more than the highest previous estimates. In many hospitals, as the number of patients declined, the number of bureaucrats increased to battle with competing hospitals over market share and with insurers over payment. On an average day in 1968, U.S. hospitals employed 435,100 managers and clerks (unpublished data) to assist in the care of 1,378,000 inpatients9. By 1990, the average daily number of patients had fallen to 853,00010; the number of administrators and clerks had risen to 1,221,600 (unpublished data).
Our state-level findings yield no evidence that managed care and competitive bidding, as envisioned under a managed-competition strategy,11 will prune hospital administration. Indeed, hospitals in states with higher HMO enrollments had higher administrative costs, as did those in the two states (California and Arizona) that have experimented most extensively with competitive bidding for hospital services. Although local administrative savings attributable to HMOs or competition might be obscured in statewide figures, this seems unlikely in California, where 80 percent of employees are insured through managed-care plans and where hospital markets in much of the state are fiercely competitive. Moreover, hospital administrative costs in the Boston area, which has a very high level of HMO enrollment and competition for managed-care contracts, are comparable to those in other U.S. hospitals.
Other regulatory reforms have not significantly streamlined hospital administration. Maryland and New Jersey, states with all-payer rate-setting systems, had administrative costs that were somewhat lower than average. But New York, which operated an all-payer system during the 1980s, did not. We found evidence against the claim that Hawaii's hospitals have strikingly low administrative costs12. Certainly, no state had administrative costs nearly as low as those at most Canadian hospitals -- on average between 9 and 11 percent of total hospital expenditures3,13 (and Fortin G and Rehmer LW, Health Information Division, Health and Welfare Canada: personal communication). (The Canadian estimates are based on cost categories similar, although not identical, to those in our analysis3.)
We did not perform statistical tests on our data and we omit confidence intervals, since the figures reflect the costs of virtually all U.S. hospitals rather than a sample. Medicare Cost Reports surely include some inaccuracies, but we see no reason to believe that hospitals systematically overstate their administrative expenses. Although a few clinical nurses may be included in the category of nursing administration, thus falsely inflating our figures for administration, most other Medicare reporting conventions tend to understate the costs of administration. Salary and other expenses for clerical personnel in clinical units (e.g., ward clerks, receptionists, and secretaries) are attributed to the clinical unit and would be counted as clinical costs in our analysis. In addition, most advertising and marketing costs (about 1 percent of total hospital spending14) are not included in the categories we classified as administrative.
Our allocation of capital, interest, and plant-maintenance costs may slightly overstate the share of these items attributable to administration. However, even under the extreme assumption that all costs for physical plant are clinical (i.e., that administration occupies no space and uses no capital equipment), 20.8 percent of total hospital expenses would still be attributable to administration.
In summary, administrative costs account for 24.8 percent of the average hospital's budget in the United States. State reforms, even those incorporating elements of a managed-competition strategy, have not lowered hospital administrative costs. Trimming the hospital bureaucracy to the Canadian level would save about $50 billion annually. A similar amount could be saved on insurance overhead and physicians' paperwork3.
Supported in part by a grant from the Robert Wood Johnson Foundation.
We are indebted to Ms. Lynn Levecque, M.B.A., for assistance in interpreting the accounting categories in the Medicare Cost Report.
Source Information
From the Center for National Health Program Studies, Cambridge Hospital and Harvard Medical School, 1493 Cambridge St., Cambridge, MA 02139, where reprint requests should be addressed to Dr. Himmelstein.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Hospital Administrative Costs
Palsbo S. E., Daube J. R., Lee S., Devitt N., Rosenheck R., Frisman L. K., Lister C., Sweet V., Himmelstein D. U., Woolhandler S.
Extract |
Full Text
N Engl J Med 1993;
329:1654-1656, Nov 25, 1993.
Correspondence
The American Health Care System
Menon M., Vickers M. A., Alpert J. J., Boren S. D., Boren D. M., Anstadt G. W., Leeman C. P., Rosenblatt M. G., Gornick M. E., Maun R. A., LaPorta R. F., Iglehart J. K., Bodenheimer T., Angell M.
Extract |
Full Text
N Engl J Med 1999;
341:917-921, Sep 16, 1999.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
