|
A correction has been published: N Engl J Med 1997;337(24):1783.
|
| |||||||||||||||||||||||||||||||||||||||||||||
Background In fiscal year 1990, administration accounted for 24.8 percent of total hospital costs in the United States — nearly twice the share in Canada. Studies from the 1970s and early 1980s found high costs, especially for administration, at for-profit hospitals.
Methods We calculated administrative costs for 6227 nonfederal hospitals and the total costs of inpatient care for 5201 acute care hospitals in the United States for fiscal year 1994 on the basis of data the hospitals submitted to Medicare. We analyzed similar data for fiscal year 1990. Using multivariate analysis, we assessed the effect of hospital ownership (private not-for-profit, for-profit, and public) on administrative costs, controlling for hospital type, census region, hospital size, and the proportion of revenues derived from outpatient services. We adjusted inpatient costs for local wage levels, hospitals' reporting periods, and case mix.
Results Administrative costs accounted for an average of 26.0 percent of total hospital costs in fiscal year 1994, up 1.2 percentage points from 1990. They increased by 2.2 percentage points, to 34.0 percent, for for-profit hospitals; by 1.2 percentage points, to 24.5 percent, for private not-for-profit hospitals; and by 0.6 percentage point, to 22.9 percent, for public hospitals. In 1994, administration accounted for 37.5 percent of total costs at psychiatric hospitals (44.4 percent at for-profit hospitals) and 33.0 percent of total costs at rehabilitation hospitals (37.7 percent at for-profit hospitals). In a multivariate analysis, for-profit ownership was associated with a 7.9 percent absolute (34 percent relative) increase in the proportion of hospital spending devoted to administration as compared with public hospitals and a 5.7 percent absolute (23 percent relative) increase as compared with private not-for-profit hospitals. Among acute care hospitals, for-profit institutions had higher adjusted costs per discharge ($8,115) than did private not-for-profit ($7,490) or public ($6,507) hospitals. Much of the difference was due to higher administrative costs ($2,289, $1,809, and $1,432 per discharge, respectively).
Conclusions Administrative costs as a percentage of total hospital costs increased in the United States between 1990 and 1994 and were particularly high at for-profit hospitals. Overall costs of care were also higher at for-profit hospitals.
Methods
Sources of Data
Each year, hospitals that receive Medicare payments must file a detailed report on their characteristics, revenues, and expenses. In response to a request filed under the Freedom of Information Act, the Health Care Financing Administration (HCFA) supplied us with a data tape of hospital expenses from Worksheet A of the Medicare Cost Reports for 6235 hospitals for the fiscal year beginning after September 30, 1992, and before October 1, 1993 (fiscal year 1994). We linked these data to HCFA's Medicare Minimum Data Set (Prospective Payment System X) for this same period, which details hospital ownership, hospital type (e.g., short-term general, psychiatric), number of beds, revenues, and income.
A total of 6227 hospitals provided data on cost categories in Column 7 of Worksheet A, which we used to calculate administrative costs. We excluded two hospitals (0.03 percent) that were missing data on ownership and hospital type from analyses of these variables.
Classification of Costs as Administrative, Clinical, or Other
We analyzed administrative costs for fiscal 1994 using methods we had previously used to analyze the data from fiscal 1990.1 Table 1 lists the categories hospitals use to tabulate expenses in Medicare's Worksheet A and indicates our classification of these categories into four groups: administrative, clinical, mixed administrative and clinical, and other (e.g., gift shop). We based our interpretation of the categories on the Medicare Provider Reimbursement Manual and supplemented it with information from consultations with the financial officer responsible for our hospital's Medicare Cost Report and with members of the staff of HCFA's Maryland and Boston-area offices. The 1994 Worksheet A subdivided capital costs into more categories than did the 1990 Worksheet A and reported cost figures for different types of intensive care units individually. However, these changes did not shift any expenses among our four groups.
|
We allocated expenses from the mixed administrative and clinical categories — capital spending, interest (most interest costs are reported under "capital" categories), building maintenance, and employee-benefit costs (except benefits-department salaries) — between our administrative and clinical categories. We assumed that the proportion of the costs of the mixed categories that was attributable to administration was the same as the share of overall costs (excluding the mixed categories) devoted to administration. We calculated the proportion of costs consumed by administration in each hospital by adding the total costs in the administrative categories to the allocated share of the costs of the mixed categories and dividing the sum by total hospital costs less costs in the categories classified as "other." To assess the sensitivity of our findings to our methods of allocating the costs of the mixed categories, we recalculated administrative costs, using the extreme assumption that no capital or maintenance costs were attributable to administration.
Analysis of Hospital Ownership
We classified hospital ownership into three categories: private not-for-profit (3382 hospitals), for-profit (1288), and public (1555). We analyzed data according to six hospital types: general short-term care (5201 hospitals), general long-term care (98), cancer (9), psychiatric (627), rehabilitation (207), and other (e.g., pediatric and Christian Science) (83).
We used a linear multivariate model12 to assess the effect of the type of hospital ownership on the proportion of spending devoted to administration, after controlling for hospital type, census region, proportion of revenues derived from outpatient services, and number of beds stratified according to hospital type. The model generates estimates of the percentage of costs devoted to administration for various groups of hospitals with the use of the following equation:
percentage for administration = intercept +O +
where O, T, C, P, and B are, respectively, the beta coefficients for hospital ownership, hospital type, census region, percentage of revenues derived from outpatient services, and number of beds (stratified according to hospital type). The intercept is a constant, and the beta coefficients represent the percent absolute increase or decrease in the proportion of hospital spending devoted to administration related to specific factors, such as ownership.
Analysis of Inpatient Costs at Acute Care Hospitals
In addition, we calculated total hospital inpatient costs and administrative costs per inpatient day and per discharge for acute care hospitals. Since hospitals' fiscal years start on different dates, we adjusted each hospital's cost figures according to the midpoint of its fiscal year, using inflation-adjustment factors supplied by the HCFA. We then adjusted the cost figures using the HCFA's case-mix index, which reflects the average severity of the Medicare cases treated at each hospital. Finally, we adjusted for local variations in labor-related costs by applying the HCFA's wage index to 71.246 percent of hospital costs, as prescribed by the HCFA's adjustment methods.13
Because administrative costs are not reported separately for inpatient and outpatient operations, we calculated inpatient administrative costs by multiplying the hospital's adjusted inpatient costs by the proportion of its total expenses devoted to administration. Because this method assumes that the share of administrative costs is equal for inpatient and outpatient operations, we examined the correlation between the proportion spent for administration and the share of total revenues derived from outpatient operations. Administration's share was higher at hospitals that derived more of their revenues from outpatient services (r = 0.11), suggesting that outpatient services have slightly higher administrative costs. For-profit acute care hospitals derived less of their revenues from outpatient care (29.7 percent) than did not-for-profit (34.1 percent) or public (38.7 percent) hospitals. Hence, our comparisons of administrative costs per inpatient day and per discharge may slightly understate the actual costs at for-profit institutions.
We based our previous analysis of fiscal year 1990 hospital administrative costs1 solely on data from Worksheet A. For the present study, we linked this data with data from the Medicare Minimum Data Set (Prospective Payment System VI) and reanalyzed it using the same methods as for the 1994 data. Finally, we analyzed changes in individual hospitals' administrative costs for the 5780 hospitals that operated under the same Medicare provider number in both years and that provided all relevant data.
Results
Administration accounted for an average of 26.0 percent of total hospital costs in fiscal year 1994, up from 24.8 percent in fiscal year 1990. Administration's share increased by at least 1 percentage point in 25 states and in the District of Columbia and fell by at least that much in 2 states and in Puerto Rico (Table 2). The proportion spent on administration varied according to hospital type, ranging from 24.3 percent at short-term general hospitals to 37.5 percent at psychiatric hospitals (Table 3). Administrative expenditures totaled $61.8 billion in 1994, a figure that excludes administration at Department of Veterans Affairs and other federal hospitals, as well as overhead that cannot be charged to Medicare.
|
|
Alternative methods of allocating capital and maintenance costs had little effect on our finding of higher administrative costs at for-profit hospitals. Even under the extreme assumption that no capital or maintenance costs were attributable to administration, administrative costs would have been 28.6 percent at for-profit hospitals, as compared with 20.5 percent at not-for-profit and 19.3 percent at public hospitals.
The higher administrative costs at for-profit hospitals were not due to a small number of high-cost hospitals; the entire distribution of administrative costs was shifted upward for the for-profit hospitals. For instance, the 25th, 50th, 75th, and 95th percentiles of the percentage of total spending devoted to administration at for-profit hospitals were, respectively, 25.7 percent, 32.1 percent, 40.5 percent, and 55.7 percent. Among not-for-profits, the comparable figures were 20.5 percent, 23.7 percent, 27.5 percent, and 34.8 percent; among public hospitals they were 19.0 percent, 22.2 percent, 25.8 percent, and 33.2 percent.
At short-term general hospitals, 1994 administrative costs averaged, respectively, 28.6 percent, 24.1 percent, and 22.5 percent of total costs at for-profit, not-for-profit, and public institutions. The administrative costs at for-profit hospitals were higher in each census region and within each quartile of hospital size (data not shown). The proportion spent on administration fell as the number of beds increased, from 25.5 percent at short-term general hospitals with fewer than 63 beds (lowest quartile), to 22.7 percent at those with more than 258 beds (highest quartile).
In the multivariate model (Table 4), psychiatric and rehabilitation hospitals had strikingly higher administrative costs, with absolute increases of 14.1 percent and 9.9 percent, respectively, over short-term general hospitals. Hospitals in the Pacific region had the highest administrative costs.
|
Between 1990 and 1994, administrative costs rose, on average, 2.5 percentage points at the 75 hospitals in our data set that switched to for-profit ownership. In contrast, administrative costs rose by only 0.4 percentage point at the 105 hospitals that were for-profit in 1990 but not-for-profit or public in 1994. Among facilities whose ownership status did not change, administrative costs rose 1.2 percentage points at for-profit hospitals, 1.0 percentage point at not-for-profit hospitals, and 0.8 percentage point at public hospitals.
Inpatient costs at short-term general hospitals (adjusted for case mix, local wage levels, and the starting date of each hospital's fiscal year) averaged $1,053 per inpatient day and $7,319 per discharge. For-profit hospitals had higher costs per inpatient day ($1,403) and per discharge ($8,115) than not-for-profit hospitals ($1,040 and $7,490) or public hospitals ($895 and $6,507). The unadjusted cost figures showed even greater disparities between for-profit and other hospitals (data not shown).
Adjusted administrative costs at short-term general hospitals averaged $262 per inpatient day and $1,778 per discharge: $396 per inpatient day and $2,289 per discharge at for-profit hospitals; $255 and $1,809 at private not-for-profit institutions; and $205 and $1,432 at public hospitals. Administrative costs accounted for 76.8 percent of the total cost difference per discharge between for-profit and not-for-profit hospitals and for 53.3 percent of this difference between for-profit and public hospitals.
For-profit short-term general hospitals had lower wage and salary costs (40.9 percent of total costs, vs. 47.6 percent at not-for-profit and 48.7 percent at public hospitals) and lower employee-benefit costs (6.7 percent of total costs, vs. 8.1 percent at not-for-profit and 8.0 percent at public hospitals). The for-profit hospitals' lower compensation costs for clinical personnel (classified according to the method used for Table 1) accounted for virtually all of this difference (data not shown).
Discussion
We found that for-profit hospitals spend 23 percent more on administration than do comparable private not-for-profit hospitals and 34 percent more than public institutions and have higher total costs per inpatient day and per discharge. The percentage of hospital costs devoted to administration increased between 1990 and 1994 for hospitals in all three ownership categories.
The data that we analyzed covered virtually all nonfederal U.S. hospitals. Although hospitals may tailor cost reporting to maximize Medicare reimbursement, these incentives do not generally reward the inflation of administrative categories at the expense of clinical ones. Furthermore, extensive regulations, routine audits, and the threat of penalties limit inaccuracies. Internal consistency checks reveal few errors.14 The standards for classification and reporting of expenses underwent no significant change between 1990 and 1994.
The hospital cost figures we used may understate total overhead. We omitted profits ($123 per inpatient day at hospitals in the largest for-profit chain in the third quarter of 199515), much of the $1.75 billion hospitals spent for marketing in 1994 (according to an estimate based on a survey of 201 hospitals),16 and the entrepreneurial expenses that Medicare excludes from reimbursement.
We found no evidence that tax exemptions for not-for-profit and public hospitals accounted for their lower administrative costs. No property taxes should have been included in our administrative categories. Corporate income taxes cannot be charged to Medicare and hence should not have appeared in the figures we analyzed. Although for-profit hospitals pay sales taxes on supplies and equipment, they pay less in Medicare and Social Security taxes since their payrolls are lower. In any event, most sales and payroll taxes would appear in our clinical categories.
Because the bulk of administrative costs are lumped in the "administrative and general" category in Medicare cost reports, we know little about what was purchased. Without reliable global measures of hospital quality, we cannot assess the relation between administrative costs and quality. Although we used Medicare's standard methods to adjust inpatient costs for case mix and local wage rates, unmeasured differences in the severity of illness or physicians' practice styles could account for some of the patterns we observed.
Most previous studies comparing for-profit and not-for-profit hospitals date from the 1970s and early 1980s,4,5,6,7,8,9,10,11 before diagnosis-related groups (DRGs) signaled the end of cost-based hospital payment. These studies found higher costs and profits but reduced staffing and lower spending for medical education among for-profit institutions.4,5,6,7,10,11 The higher costs were due to more administration and higher charges for pharmaceuticals, supplies, and other inpatient ancillary services.5,6,7,9 More recent studies of psychiatric and rehabilitation hospitals, which are exempt from DRGs, have also found higher costs, higher profits, and reduced staffing among for-profit hospitals.17,18,19 Scant recent data on acute care hospitals, mostly from Florida, suggest that the pattern of higher profits and reduced staffing at for-profit hospitals has persisted20,21; by some technical measures of efficiency, they are even less efficient than they were.22
Why is hospital administration growing, especially at for-profit institutions that are favored by the market? For all hospitals, success increasingly requires business acumen, coordination, and control of the clinical work force. For-profit hospitals have led the way in curtailing lengths of stay, though surprisingly this does not translate into lower costs per admission. The shorter stays and lower occupancy rates at for-profit hospitals23 may increase the share of costs devoted to administration if clinical staffing is cut more than administration when beds are empty.
It may be incorrect to equate market success with efficiency.3 Business strategies unrelated or even detrimental to efficiency can bolster profitability and market share. For instance, hospitals can enrich their mix of payers through selective marketing, providing financial incentives for physicians, or locating facilities in affluent neighborhoods; increase revenues by "upcoding" DRGs, meticulously adhering to managed-care approval procedures, or vigorously pursuing collections; and cut costs by curtailing research and education programs. Being large and having abundant access to capital enhance hospitals' leverage with suppliers, physicians, insurers, and politicians and provide them the wherewithal to weather price wars and recoup losses after competitors fail or are purchased. National hospital chains have fewer local ties that could inhibit their closing unprofitable services and facilities or avoiding financially unattractive patients and physicians.
Contrary to the rhetoric of the market, market forces are apparently "upsizing" administration. Between 1990 and 1994, the overhead of health insurers and HMOs rose $20.1 billion,24 from 12.8 percent25 to 14.8 percent24 of premiums; the number of administrative and clerical personnel in practitioners' offices increased from 613,000 to 819,000 full-time equivalents (Current Population Survey: unpublished data); and hospital administrative staffing soared.26 Even before these recent rises in expenditures for administration, the U.S. General Accounting Office calculated that the administrative savings that would result from a shift to national health insurance would fully fund universal coverage.27 Since the gap between health care costs in the United States and those in the rest of the world is widening28 and the ranks of the uninsured and underinsured are swelling,29,30 perhaps it is time to ask whether our experiment with market medicine has failed.
Source Information
From the Center for National Health Program Studies, Cambridge Hospital, CAMBRIDGE, MASS., and Harvard Medical School, Boston.
Address reprint requests to Dr. Woolhandler at Cambridge Hospital, 1493 Cambridge St., Cambridge, MA 02139.
References
| |||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Costs at For-Profit and Not-for-Profit Hospitals
Miller C. R., Epstein J., Woolhandler S., Himmelstein D. U.
Extract |
Full Text
N Engl J Med 1997;
337:1779-1780, Dec 11, 1997.
Correspondence
When Money Is the Mission — The High Costs of Investor-Owned Care
Greene H. R., Packard B. J., Mick B. A., Menendez R., Zinberg S. S., Silverman E. M., Skinner J. S., Fisher E. S., Himmelstein D. U., Woolhandler S.
Extract |
Full Text
N Engl J Med 1999;
341:1768-1770, Dec 2, 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
