National Academies Press: OpenBook

For-Profit Enterprise in Health Care (1986)

Chapter: 4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance

« Previous: 3 Trends in the Growth of the Major Investor-Owned Hospital Companies
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 260
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 261
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 262
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 263
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 264
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 265
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 266
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 267
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 268
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 269
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 270
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 271
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 272
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 273
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 274
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 275
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 276
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 277
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 278
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 279
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 280
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 281
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 282
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 283
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 284
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 285
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 286
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 287
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 288
Suggested Citation:"4 The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance." Institute of Medicine. 1986. For-Profit Enterprise in Health Care. Washington, DC: The National Academies Press. doi: 10.17226/653.
×
Page 289

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

INTRODUCTION The hospital industry is undergoing major structural change. Shifts in Medicare and Medicaid reimbursement policies, increasing private concern about health care costs, tech- nological change, and the growth of organized delivery systems such as health maintenance organizations (EIMOs) and preferred provider organizations (PPOs) have radically altered the environment facing hospitals. Hospitals have responded in two primary ways: (1) by diver- si*ing away from inpatient acute care services to find new revenues Tom services such as ambulatory care, home care, and skilled nurs- ing care (vertical integration); and (2) by join- ing existing multihospital systems or forming new ones (horizontal integration) or starting non-health-care businesses such as construc- tion (diversification). Both of these strategies aim to improve the potential service benefits and economics of hospital operations, the first primarily by rais- ing service volumes by assuring continuity in referrals, and the second primarily by achiev- ing administrative economies of scale and bet- ter access to capital. These Rends have resulted in the large numbers of corporate reorgani- zations of hospitals and the steady growth in *Mr. Watt is a principal, Lewin and Associates, Inc., San Francisco, California. Mr. Renn is a research associate, lhe Johns Hopldns Center for Hospital Fi- nance and Management, and the Department of Health Policy and Management, The Johns Hopkins Univer- sity, Baltimore, Maryland. Mr. Hahn is a research as- sistant, Lewin and Associates, Inc. Mr. Derzon is vice president, Lewis and Associates, Inc. Dr. Schra~nm is director, The Johns Hopldus Center for Hospital Fi- nance and Management, and associate professor, the Department of Health Policy and Management, The Johns Hopldns University, Baltimore, Maryland. For-Profit Enterprise in Health Care. 1986. National Academy Press, Washington, D.C. The Effects of Ownership and Multihospital System Membership on Hospital Functional Strategies and Economic Performance I. Michael Watt, Steven C. Renn, James S. Hahn, Robert A. Derzon, and CarIJ. Schramm numbers of hospitals in multihospital systems that have been seen over recent years. This paper focuses on the second of these phenomena the behavior of multihospital systems-and examines differences in the lo- cation, strategies, and economic performance of five classes of community hospitals: inves- tor-owned multihospital system (IOMS) mem- bers, not-for-profit multihospital system (NFPMS) members, and investor-owned (IO) freestanding, NFP freestanding, and govern- ment hospitals. This study attempts to sepa- rate differences associated with system membership from those associated with own- ership form and to develop answers to the fol- lowing questions: · Are there differences in demographics, reimbursement systems, and regulatory struc- tures between areas where multihospital sys- tem members are located and those in which freestanding hospitals are located? Are there differences between the areas where IO and NFP hospitals locate? · Do multihospital systems produce econ- omies in patient care or employ different h- nancial strategies compared with freestanding hospitals? Are there differences in the advan- tages multihospital system membership gives IO or NEP hospitals? · Are the inter-hospital differences that may be attributed to ownership-IO versus NFP- greater than the differences that may be due to organizational form chain-affiliated versus freestanding? ~ What does Me historical record say about the potential of each type of hospital for suc- cess in a world in which purchasers of care are increasingly price-conscious? Issues of comparative quality of care and access, while also important, were beyond the scope of our analysis of economic performance. 260

OWNERSHIP AND ECONOMIC PERFORMANCE METHODS We analyzed a sample of hospitals that is national in scope and includes about 10 per- cent of all U. S. community hospitals, ranging from urban academic health centers to small rural facilities. The Sample The sample was chosen by a stratified ran- dom sampling technique from the universe of 4,491 Medicare-certified, nonfederal, and nonstate general acute care hospitals with lengths of stay between 3 and 13 days in 1980. Using the Guicle to Multihospital Systems of the American Health Association (AMA) and the Directory of the Federation of American Hospitals (FAH), each of the hospitals was classified based on ownership and system af- filiation. We selected from this universe a sam- ple of hospitals totaling 800 in aggregate. We originally hypothesized that there would be little difference (other than greater direct re- ceipt of public funds) between voluntary hos- pitals and those organized as county and hospital district entities. However, in the models pre- sented here, to control for differences such as sunshine laws, limitations on outside contract- ing, and restrictions on governing board com- position that may remain even after adjusting for public subsidies, we included a category for government hospitals. The original sample of 800 was reduced to 561 due to cost reports missing important schedules or covering periods of less than 9 or more than 15 months. The number of sample hospitals in each of the five strata are as fol- lows: IOMS members-122; NFPMS- 114; IO freestanding hospitals 148; NFP free- standing hospitals (voluntary or religious- sponsored) 93; and government (county or hospital district) hospitals-84. We tested each of the samples against the eligible universe from which it was drawn, us- ing two-tailed l-tests of the difference between means. We found the samples to represent their classes well in hospital size, occupancy, length of stay, and the ratio of outpatient to total gross revenue, as shown in Table 1, with two exceptions. First, the freestanding IO hos- pitals in our sample are slightly smaller and 261 the sampled chain IOs slightly larger than their classes overall. We do not believe these dif- ferences have important consequences for the results that follow because we explicitly con- trol for number of beds in each equation. Sec- ond, the inpatient proportion of total revenue differs between the sampled freestanding NFPs and their class. However, a variable to control for this factor is also included in the equations. There is little evidence of "response bias"- differences between the chosen random sam- ple and the final usable sample. The sample has a good distribution of hospitals from the major muldhospital systems. However, sev- eral large public system teaching hospitals in New York and California could not be included in the sample due to our inability to disaggre- gate some key financial data for the individual hospitals. Kaiser hospitals were excluded for slml. ar reasons. Several measures of hospital costs and rev- enues, utilization, and balance sheet strength were taken Tom audited Medicare cost re- ports. Data on the number of births, surgical admissions, urban/rural location of the hospi- tal, and contract management and member- ship status in NFP systems in 1980 were taken from the AMA's Annual Survey. Demographic and physician supply information from the county where each hospital is located was taken from the U.S. Bureau of Health Professions' Area Resource File. Figures for the percentage of admissions covered by each type of third- party payer were constructed from the U.S. Office for Civil Rights' 1980 Survey of Hos- pitals.2 Finally, the Health Care Financing Administration (HCFA) supplied wage and case- mix indices and lists of hospitals that were sole community providers. This data base was checked for consistency among sources and within each hospital record. Techniques Financial data were standardized to a com- mon, 12-month fiscal period ending Decem- ber 31, 1980, using trend factors based on monthly expense data from the AHA s Panel Survey. Expense and statistical data from hos

262 TABLE 1 Sample Compare] to Universe, 1980 l FOR-PROFIT ENTERPRISE IN HEALTH CARE Beds per Hospital Occupancy Rate Average Length of Stay Ratio of Inpatient Gross Revenue to Total Gross Revenue N Mean N Mean N Mean N Mean Chain-afEiliated investor-owned Sample 122 167.48 122 0.62 122 6.17 122 0.91 Universe 310 146.19* 307 0.63 307 6.37 306 0.90 Chains liated not-for-profit Sample 114 220.77 114 0.69 114 6.58 114 0.88 Universe 436 223.32 416 0.70 418 6.80 413 0.87 Freestanding investor-owned Sample 148 94.01 148 0.63 148 6.80 142 0.91 Universe 213 111.63* 213 0.64 213 6.57 209 0.91 Freestanding not-for-profit Sample 93 220.73 93 0.70 93 6.94 93 0.88 Universe 2,397 221.01 2,306 0.72 2,306 7.24 2,287 0.87 Government Sample 84 103.56 84 0.60 84 5.96 82 0.87 Universe 1,264 108.41 1,244 0.61 1,247 6.16 1,221 0.87 NOTE: Levels of statistical significance are denoted by asterisks as follow: *probability <.10; **probability <.05; ***probability <.01 that observed difference is due to chance. SOURCE: Sample: Lewin and Associates and The Johns Hopkins University; Universe: American Hospital Association Annual Survey of Hospitals. pitals with reporting periods of other than 366 days were adjusted by the ratio of the days in the period to 366. Edits were used to check the consistency of the measures calculated, and all outliers were verified. We then analyzed this data set through or- dinary least squares multiple regression on the models discussed below. Multiple regression was chosen to overcome two limits of the matched-pairs methodology we have applied in other studies. First, regression permits larger sample sizes, because there is no need to match hospitals explicitly in the respective samples under study. This allows the study to be na- tional in scope, rather than limited only to states in which the hospital classes under study are located, and increases one's confidence that the measures calculated represent average re- lationships nationwide. Second, because the regression technique is not limited in its ap- plication to two equal-sized samples, differ- ences among IO and NAP system hospitals and IO and NFP freestanding hospitals can be ex- amined explicitly at the same time. Dependent Variables The 28 dependent variables included mea- sures of case mix, operating revenues and costs, capital structure, markups and profitability, and activity and productivity. With few excep- tions, these measures were taken from the Medicare cost reports directly or calculated from them, as we felt this was the most reliable and consistent source. Table 2 shows the de- pendent variables, their sources, and Weir mean values for each of the hospital classes. Severity and mix of cases are indicators of patient selection strategies by the hospitals. Those operate through decisions on siting, ser- vice offenngs, and recruitment of different types of physicians to the medical staff. We mea- sured severity and mix of cases by the Medi- care case-mix index, inpatient surgeries per

OVKNERSHIP AND ECONOMIC PERFORMANCE TABLE 2 Descriptive Statistics: Dependent Vanables 263 Standard Variable Source: Definition NMeanDeviation Case mix Case-mix index (1980) HCFA, Federal Register 5610.98940.0822 Inpatient surgeries per 100 admissions AHA Survey 56037.348117.3455 Births per 100 admissions AHA Survey 5606.62766.6465 Outpatient revenue to total gross revenue (ratio) Medicare Cost Report (MCR) 5590.10920.0659 Revenues and costs Total patient care revenue MCR, adjusted for case mix, per adjusted admission ($) wage rate, and outpatient differences 4901,888.6383607.3580 Net patient care revenue per adjusted admission ($) Total operating expenses per adjusted admission ($) Total patient care expenses per adjusted admission ($) General and administrative costs per adjusted admission ($) Home office costs per adjusted admission ($) Capital structure Debt-to-asset ratio Current ratio Capital cost percent A (excluding ROE, percent of operating costs) Capital cost percent B (including BOE, percent of operating costs) Accounting average age of plant Total fixed assets per bed ($) Markups and profitability Gross patient care markup ratio Revenue deduction ratio Nonoperating revenue to total gross revenue Total markup ratio Return on total assets Return on equity or fund balance MCR, as above MCR, as above MCR, as above MCR, as above MCR, as above MCR: total liabilities/total assets MCR: current assets/current liabilities MCR: depreciation, interest, and capital leases/total operating costs MCR: depreciation, interest, capital leases and Medicare ROE/total operating costs MCR: accumulated depreciation plant/depreciation expense MCR: fixed assets/beds MCR: gross patient revenues/ operating expenses MCR: (gross patient revenues - net patient revenues)/gross patient revenues MCR: nonoperating revenue/ (operating revenue + nonoperating revenue) MCR: (gross patient revenue + net nonpatient revenue)/total operating expense MCR: total net income/total assets MCR: total net income/equity or fund balance 495 1,589.7865 543 1,598.0947 510 1,440.0125 510 495 548 540 550 455.5115 467.9824 399.4624 696.3724 230.4583 19.0471 40.1001 0.5315 2.7231 7.1173 7.8796 0.3193 2.2946 4.2766 4.5876 4424.97815.7084 49735,690.504225,489.4713 536 538 531 537 529 517 1.1935 0.1455 0.0301 1.2299 0.0573 0.1861 0.1654 0.0823 0.0324 0.1595 0.1067 0.3550

264 TABLE 2 Continued FOR-PROFIT ENTERPRISE IN HEALTH CARE Standard Variable Source Definition N Mean Deviation Activity and productivity ratios Total asset turnover ratio MCR: gross revenue/assets 537 1.5215 0.9312 Current asset turnover ratio MCR: gross revenue/current assets 533 4.2029 1.9565 Case flow MCR: admissions/beds 561 37.2416 11.0268 F1Es per adjusted daily MCR, supplemented by AHA: census full-time equivalent Occupancy rate Medical education personnel/adjusted average daily census MCR: inpatient days/total days available 518 2.3923 0.8455 561 64.7465 17.0570 Interns and residents total MCR: interns and residents expense per bed (S) expense/beds 561 278.1697 1,097.9548 100 admissions, births per 100 admissions, and outpatient proportion of total gross revenue (a proxy for outpatient volumes). Operating revenues and costs per case were examined as indicators of differences in hos- pital pricing strategies, productive efficiency, and the costs imposed by the home offices of system members. Our measures for operating revenues and costs were total patient care rev- enue (gross charges), net patient care revenue, total operating expense, total patient care ex- pense, general and administrative costs, and home office costs, all per admission and ad- justed for outpatient volume, case mix, re- porting period length and end, and wage rate differences. We also measured total intern and resident costs per bed. Financial ratios addressing the capital struc- tore of hospitals yield insight into the financial strategies followed by hospitals in the different classes as well as their differential access to capital. These were measured by total fixed assets per bed, debt-to-asset and current ra- tios, capital costs as a percentage of operating costs, and accounting age of plant (defined as accumulated depreciation divided by depre- ciation expense). Markups and profitability measures reflect other aspects of managerial strategy as well as the relative success of the hospitals' policies. The patient care markup ratio (gross patient revenues divided by patient care expenses), revenue deduction ratio (gross patient reve- nues minus net patient revenues, divided by gross patient revenues), ratio of nonoperating revenue to total net revenue, total markup ratio (total gross revenue divided by total op- erating cost), return on total assets, and return on equity (ROE) were used as measures of markups and profitability. Finally, measures of activity and productiv- ity were included as dependent variables to show differences that might exist in the effi- ciency of use of assets across the hospital class- es. These measures were total asset turnover, current asset turnover, case flow (admissions per bed), and full-time equivalent staff (FlEs) per adjusted average census. We controlled for case mix and wage dif- ferences from the outset using the methods employed by HCFA to standardize hospital data for these factors under the Prospective Payment System (PPS). Additionally, depen- dent variables measuring revenues and costs were adjusted for outpatient volume using the technique of the AHA. Finally, as described above, all financial data were standardized to a 12-month reporting period ending Decem- ber 31, 1980. Thus, all revenue and cost-de- pendent variables were standardized for case- mrx differences, wage rate differences, and dif- ferences in outpatient volume before they en- tered the regressions.

OWNERSHIP AND ECONOMIC PERFORMANCE Independent Variables and Hypotheses The independent variables of primary in- terest were those representing ownership and system membership, specified as a construct of four dummy variables, with NFPF being the omitted reference group; and length of affiliation and contract management, which were specified as a construct of three dummy variables. Other independent variables con- sisted of five tvDes that mav also affect hocnitn1 of, economic performance: 1. Competition and regulation were mea- sured by two proxies sole community pro- vider status under Medicare (a measure of lack of competition) and percentage of admissions under state rate-setting programs. 2. Case mix and patient selection was mea- sured by the Medicare case-mix index, sur- geries per 100 admissions, births per 100 admissions, and percent charge-based payers. We expected both higher case-mix indices and percentages of surgical admissions to be as- sociated with higher revenues and costs per case. 3. Input costs and regional practice differ- ences were examined through three measures, the HCFA area wage index, a dummy equal to 1 if the hospital was in a Standard Metro- politan Statistical Area (SMSA), and a second dummy equal to 1 if the hospital was in Census Region nine the West Coast. This last vari- able was specified following preliminary regressions that found only this region to show different hospital performance on many of our measures. Hider wage indices and urban lo- cation were expected to be associates} with higher revenues and costs. 4. Productive capacity and utilization were measured by the hospital's number of beds, ratio of outpatient to total gross revenue, oc- cupancy rate, and, in some models, length of stay. Higher numbers of beds and lower oc- cupancy rates were expected to be associated with higher costs per stay. We entered the linear form of the bed variable because pre- liminary regressions showed it fit as well as quadratic and other formulations. 5. Finally, medical education commitment was measured by the medical resident and in- tern expense per bed from the Medicare cost 265 report, and was expected to be associated with higher costs. The sources and means of the independent variables are shown in Table 3. We developed initial hypotheses about the ejects of the independent variables on the de- pendent variables and tested them through the regressions. From our previous studies,2 we hypothesized that IO system hospitals would tend to be located in the sunbelt states char- acterized by faster population growth, higher proportions of charge-based reimbursement, endless regulation. They would showcase mixes about equal in total (Medicare case-mix index) and surgical proportions, and lower in obstet- rical and outpatient proportions than NFP freestanding ones. We hypothesized that IOMS revenues and administrative and home office costs per case would be higher than either freestanding or system NFPs, but their total operating and patient care expenses per case would be equiv- alent to the NFPs. We believed they would show lower asset values per bed, higher debt- to-asset ratios, higher capital costs as a per- centage of operating costs and younger plants (by accounting-based measures) than NFPs, and that these strategies would result in higher markups and profitability. Finally, we ex- pected IO system hospitals to show faster asset turnover on revenue-based measures, and greater productivity measured by full-time equivalent staff per occupied bed. Our hypotheses regarding NFP system hos- pitals were that they would exhibit similarities to bow NEP freestanding hospitals (which most of them once were) and IO system hospitals (due to similar opportunities to achieve sys- tem-derived economies and access to capital). Therefore, for many relationships, our hy- potheses were indeterminate. However, we ejected their case mixes to be somewhat higher overall and relatively richer in obstetri- cal and outpatient services than IO system hos- pitals. Like the IO systems, we expected that their home office costs would increase admin- istrative and general costs relative to NFP freestanding facilities. Because one of the rea- sons for forming a multihospital system is the ability to pool resources to improve access to capital, we expected to see relatively higher

266 TA;BLE 3 Descriptive Statistics: Independent Enables FOR-PROFIT ENTERTR1SE IN HEALTH CARE Standard Variable Source N Mean Deviation Ownership and affiliation Investor-owned chain Not-for-profit chain Investor-owned freestanding Government Not-for-profit freestanding (omitted reference group) Contract management and length of affiliation Contract managed in 1980 Chain affiliated in 1977 and 1980 Contract managed in 1977 and 1980 Competition and regulation (D) ~ `` 561 ( ) | American Hospital Association | (D) (AMA) Guide, Federation of 561 (D) American Hospitals (FAH) 561 Directory, and Medicare (D) ~Cost Reports (MCRs) 561 (D) (D) (D) AHA Special Run 0.2175 0.2032 0.2638 0.1497 0.1658 0.4129 0.4027 0.4411 0.3571 0.3722 561 0.0963 0.2952 561 0.2175 0.4129 561 0.02S0 0.1561 _ Sole community provider (D) Health Care Financing Administration (HCFA) 561 0.0303 0.1716 Percent of admissions Office for Civil Rights (OCR) covered by rate setting survey and state rate setting agencies 559 5.8147 21.0005 Case mix and patient selection Case-mix index Inpatient surgeries per 100 admissions Births per 100 admissions Percent of admissions covered by charge-based payers Association HCFA, Federal Register AHA Annual Survey AHA Annual Survey OCR survey and Blue Cross 561 560 560 0.9894 37.3481 6.6276 0.0822 17.3455 6.6465 539 41.3264 18.0965 Input cost and regional practice differences Wage index HCFA, Federal Register 5610.99270.1595 Located in Census Region 9 (D) AHA 5610.15690.3640 Capacity and facility utilization Number of beds MCR (available beds) 561158.1854173.2655 Occupancy rate (percent) MCR 56164.746517.0570 Outpatient revenue to total MCR (includes outpatient gross revenue (ratio) ancillary and ER) 5590.10920.0659 Average length of stay MCR 5616.51471.6238 Medical education Interns and residents: total expense per bed ($) MCR 561278.16971,097.9548 NOTE: D = dummy variable: value equals 1 if hospital has attribute, 0 otherwise.

OWNERSHIP AND ECONOMIC PERFORMANCE debt-to-asset ratios and newer physical plants among system NFPs than freestanding NFPs. Based on an earlier Lewin and Associates study,3 we expected many similarities be- tween IO freestanding hospitals and IO system members. However, we expected higher pro- portions of surgical admissions and a lower overall case-mix index at freestanding IOs than freestanding or system NFPs, lower overhead costs due to the absence of home office costs, and lower fixed assets per bed and debt-to- asset ratios than system IOs. We hypothesized that in many ways gov- ernment hospitals would resemble freestand- ing NFPs. However, they would have higher outpatient volumes, older plants, lower ratios of capital to operating costs, and higher FTEs per occupied bed than NFPs (due to the public service employer role of some of these facili- ties). After adjusting to control for public sub- sidies, we expected little difference in the markups or proportion of nonoperating to total revenue between government and NFP hos- pitals. Several studies have shown that it often takes several years for the advantages of multihos- pital system membership 'to be realized. Therefore, we expected that hospitals affiliated with systems for more than 3 years would show better results than hospitals with shorter his- tories. We hypothesized similar relationships for contract-managed hospitals. Hospitals managed for less Man 3 years would show poorer financial performance (because correcting such situations is'' a primary reason for becoming managed), but performance would be better in hospitals managed for a longer time. Regression Models The regression models were developed by including, for each dependent variable,'the independent variables which we hypothesized would have an effect. To ease discussion of the results and to provide a more complete expla- nation of the effects and relationships between variables, we included the same set of inde- pendent variables in almost all the equations. Two sets of regression equations were cal- culated. First, freestanding NFP hospitals were used as the reference group against which other classes of hospitals' were compared. These re 267 suits are presented in the tables of regression results (Tables 5-9), which show the freestand- ing NFPs as the omitted reference group. Iben the model was recalculated with the chain- alliliated NFP hospitals as the omitted refer- ence group in order to explicitly measure the magnitude and statistical significance of the differences between NFP and IO system hos- pitals. This result is shown near the bottom of He tables as "IONIS compared to NFPMS.', The changed reference group does not affect the coefficients or significance level values of any of the independent variables except the ownership construct, due to the mathematical relationships involved. Standard tests eliminated variables suffi- ciently collinear to bias the results. However, our research has four limitations. First, in some cases the data we were able to obtain may not represent completely the effects we were trying to capture. For example, our payer-mix data are based on anticipated source of payment, not actual payer source. Second, several factors that affect a hospi- tal's performance are inadequately repre- sented due to poor data with which to construct measures. An example is that our measure of the competitive environment is whether the hospital was a Medicare-designated sole com- munity provider a negative measure that does not distinguish between cases in which a hos- pital may be ringed by five larger ones and cases in which the sample hospital is the dom- inant one in a three-hospital town.4 Third, other factors that may be important to strategies and performance, such as the way in which strategies are formulated and com- municated, the nature oftheincentives placed on management, and management skill, could not be included in the current study. Finally, our mode! is cross-sectional and, thus, limits the inferences that can be made about cau- sality rather than association. RESULTS Our results include findings on SLY groups of issues: siting and service area characteris- tics, case mix, operating revenues and ex- penses, markups and profitability, capital structure, and productivity in the use of assets.

268 Siting and Service Area Characteristics Hospitals of the five classes considered in this study differ in their geographic distribu- tion as a result of historical and economic fac- tors. Table 4 shows the distribution of hospitals across U.S. Census regions. IO chain facilities are concentrated in the South Atlantic region, including Florida (re- gion 3), the East South Central region, in- cluding Tennessee (region 5), the West South Central region, including Texas (region 7), and the Pacific region, including California (region 9) essentialRy the "sunbelt," characterized by faster population growth than other areas of the country. NFP systems are also somewhat concentrated, but in different regions: the East North Central region (Midwest region 4), West North Central region (the Plains states region 6), and the Mountain states (region 81. They overlap the IO systems in the Pacific and West South Central regions. Freestanding IO hospitals are highly concentrated in the West South Central and Pacific regions, but also re- main on the Eastern Seaboard. NFP free- standing hospitals are well distributed across census regions. The government hospitals in our sample tend to be the county and district- sponsored facilities prominent in the South. Since the geographic regions differ in demo- graphics, income, and health care industry regulation, as well as other factors, the broad differences in the locations of hospitals of the five classes influence our findings on five types of siting characteristics. Urban, Suburban, and Rural Overall, about 60 percent of the hospitals in our sample were located in counties or ag- gregates of counties defined as SMSAs, com- pared with about 51 percent of U. S. hospitals overall. A higher proportion (74 percent) of IOMS hospitals, and a significantly lower pro- portion of government hospitals (27 percent), were urban.5 Differences were sharper when subregions of SMSAs were examined. About 30 percent of the hospitals in the sample over- all were located in suburban areas (non-central cities) of the SMSAs. However, about 40 per- cent of the chain IOs and 42 percent of free- standing IOs were located outside central city FOR-PROFIT ENTERPRISE IN HEALTH CARE city areas-significantly higher than the per- centages of the freestanding NFP hospitals that were located in suburban areas. NFPMS and government hospitals were about as likely as freestanding NFPs to be located in suburban metropolitan areas, but less likely than IO chains to be in the suburbs. Wage Index and Sole Community Provider Status As a consequence of their differing patterns of urban-rural location, the wage indices de- veloped by BCFA were slightly higher for the investor-owned hospitals (both those in mul- tihospital systems and freestanding ones) and lower for the government hospitals in our sam- ple than for the NFPs. No IO hospitals in our sample had been given Medicare sole community provider sta- tus, a designation indicating that a hospital lacks local competition from other hospitals. Hospitals must apply for this designation, and the benefit of doing so in 1980 was to exempt the hospital from Section 223 controls on rou- tine costs under Medicare. Therefore, fewer sole community providers among the IOs could indicate either fewer rural hospitals among the IOs (supported by our findings above) or fewer hospitals needing relief from the cost limits. Of the 17 sole community providers in our sample, 6 were government hospitals, 3 were freestanding NFPs, and 8 were members of NFP-systems. Area Demographics On average, about 14 percent of the pop- ulation in the counties in which the sample hospitals were located were below the poverty level in 1980. The figure was highest for gov- ernment hospitals (about 16 percent). IO hos- pitals, both Eeestanding and chain-a~liated, also had slightly higher rates of poverty in their home counties than did the average freestand- ing or multihospital system NFP hospital in our sample, but when the broad census re- gions of the hospitals were controlled for (e.g., New England and Pacific), none of these dif- ferences was significant. The hospitals' home counties showed essentially no difference in the percentage of their populations over 65

269 o - C~ - ._ C~ o ._ o ~n - 6, ~o ~ 5 ~, Ct o C) - ~, ~ cn ~_ CC - C~ ~, ._ C~ o o o · _ ~5 L, U. ·_ C) ~_ - o C) - ~3 m -1 ~ ~o, t E ,o C~ ._ - .= C) Ct 3 C~ ~, ._ _ ~7 a~ 3 O - C' 3 ~ O =, o ~ C~ C~ ~ _ _ o _ ~,~ L' ~ .= _ C~ Z ~ - 3 ~ ._ O ~ I q C~ ~ ~V ~s ~ C) "o~ o~ ,o =5 ~ C.) U) ~ CO U: ~ ~ _ a~ ~ C) C~ U. o ~o ~ ~o o' o o o o C ~C ~_ _ o ~o o o o _ co ~O ~_ ~' O - 8 o o o o o o o o _ 8 o ~ o o o o o ~ ~o 8 o ~ o o o. o o o o C~ ° o ° 0 o o o o o 8 o o ~o_ _ _ - o o o o o oo o o o _ 8 o ~o ~ ° ~ 8 o o o o oo o o o _ o ~o ~o _ 8 o o o o oo o o o _ CO CO Ct e~ . - o z .. _ . Z ~ . . _ ~ z ~ . . C~ C~ C) , C: CO ^ <~s c) ~ z 3 . ~ . . U, U, ._ . ^ '3;, Ct t~ . . tC ~ ~ U: C~ Ce Z ~ ^a ~ 0 3 ~ Z o . . au Ct _ _ . O o^Z ~ _ <`s C°: ^ ~ ^ o .. .. ._ ._ Ct ~ ~¢ _ ;> CC C: ~ ^ e, e~ O - . . c) 3 8) C~ .~ C~ C~ .- ~ ~o o. co) ._ _ o ~o s:: c) ~ ~ ,., Ct ~ o ~o ~ ~ . - = Ct ~ om _ S~ .= _ o~ c: Cn ._ C) ~ ~, C. 3 ~ C~ '~ I, Ct 3 - O C~ > O _ s~ ._ --o ~ ~ ._ C) U, ~ . ~ C~ ~ 4;) O _ S~ {C ,3 C) 3 Ct C~ 30- ~ ~ o . O '== ~ o O ~ ~ - 4.~ ~ `~ ~ _ ~ ._ t4- O C: ~ -0 G ~_ S~ CO o :t C~ _3 G :5 S: o ._ - .o o - ._ o _~ Ct C~ ~_ - ._ C~ C) ._ o C~ - o _. - S: C~ C~ ._ CD o :~ e~ ._ 3 . . C) ;~ o

270 years of age. While the home counties of the IO hospitals showed higher percentages of nonwhite population (about 20 percent versus 16 percent in the sample overall), these dif- ferences were primarily related to the differ- ences in the census regions in which the hospitals were located. Area Health Care Capacity and Utilization The home counties of NFP chain hospitals showed higher ratios of beds per 1,000 persons than the sample as a whole (about 5.3 com- pared with 4.8 overall). The counties in which the government hospitals in our sample were located showed lower rates of hospitalization as measured by admissions per 1,000 persons (156 versus 175 overall) and surgical operations per 1,000 persons (65 versus 80 overall) com- pared to freestanding NFPs. These differences are measured after con- trolling for regional differences in capacity and use, which other studies have found to be large and significant. When these regional differ- ences in practice were not controlled for, the sampled NFP chain hospitals, located primar- ily in the Midwest and Rocky Mountain states, show higher rates of admission per 1,000 per- sons than the other classes of hospitals in the sample. Third-Party Coverage and Rate Setting We found essentially no difference in the percentage of admissions covered by Medicare and Medicaid across the hospitals. However, the shares of the insurance market held by Blue Cross and the bases for Blue Cross pay- ment differ across geographic regions and show in lower proportions of Blue Cross admissions at the IO and government hospitals than at the NFPs of both classes. IO chain hospitals had a significantly higher proportion (35 percent versus 31 percent overall) of their admissions covered by insurance. Coupled with the fact that Blue Cross more often paid charges in their service areas, this led the IO system sam- ple to have higher proportions of charge-based payers (48 percent versus 41 percent overall) than any other group. Virtually none of the IO FOR-PROFIT ENTERPRISE IN HEALTH CARE chain hospitals and few freestanding IOs were located in rate-setting states in 1980. Discussion We believe these findings, taken together, suggest several important differences in the siting strategies of IO and NFP hospitals. First, IO hospitals in 1980 preferred to locate in areas with rising populations-the sunbelt and the suburban areas, rather than the Midwest, cen- tral cities, and rural areas which were losing population. As a consequence of this ~er- ence, IO hospitals had higher wage indices than the average NFP hospital in our sample or the government hospitals, which were pri- marily rural. Also, the lower proportion of ru- ral hospitals among the IOMS systems in 1980 suggests they may have favored areas in which they could be the seconcl provider in town, where they could differentiate the services they offered from those of the existing provider and become the hospital of choice for scheduled medical-surgical patients.6 Second, at least at the county level, the ser- vice areas of the IO hospitals do not show ev- idence of attempts to "cream-skim." IO hospitals did not avoid counties with relatively higher rates of poverty, proportions of nonwhite pop- ulation, or high proportions of elderly. Below the county level, however, we have no data, except for whether the hospital is located in the suburban area or the central city of an SMSA. Their more frequent location in sub- urban areas may afford the IO hospitals im- med~iate service areas relatively richer in charge- paying patients. A third difference concerns per capita hos- pital capacity. Even after controlling for re- gional differences, NFPMS members are located in areas of higher hospital capacity (measured in beds per 1,000), which is one measure of the extent of hospital competition in an area. IOMS multihospital system mem- bers, on the other hand, have not located in areas of high bed-to-population ratios. NFP system hospitals may believe that they will be the survivors in competition in overbedded areas, or may have joined the systems to im- prove their chances. Fourth, the lower use of health services in the home counties of the government hospitals

OWNERSHIP AND ECONOMIC PERFORMANCE may indicate access barriers in rural areas pri- marily served by government hospitals. How- ever, we cannot conclude from our sample that the presence of NFP chain or IO chain hos- pitals either drives up use rates or is associated with higher rates of hospital use. Finally, in 1980, IO hospitals of both classes avoided rate-setting states and ones in which Blue Cross penetration was high or its discount deep. Case Mix Table 5 presents the results of our models of four indicators of case-mix differences: the Medicare case-mix index, inpatient surgeries per 100 admissions, births per 100 admissions, and outpatient revenue to total gross revenue. IOMS hospitals, when compared to NFP\IS facilities, had slightly lower Medicare case-mix indices (0.999 versus 1.017), fewer births per 100 admissions (4.79 versus 7.36), and lower proportions of outpatient to total revenue (0.099 compared to 0.119~. We did not find any sig- r~ificant difference between the two groups on the measure of inpatient surgeries per 100 ad . . missions. NFPMS members were not significantly dif- ferent from freestanding NFP hospitals on any of the measures, but showed slightly higher case-mix indices and outpatient proportions, and lower surgeries and births per 100 ad- missions than freestanding NFPs. IO freestanding hospitals had the lowest case- mix indices (0.982) of any of the groups in our sample. Both they and government hospitals had significantly lower numbers of surgeries per 100 admissions than did freestanding NFPs. Freestanding IOs also had lower numbers of births per 100 admissions than did freestand- ing NFPs about halfofthe freestanding NFP value. Whether a hospital was contract managed, how long it was contract managed, and how long a hospital had been part of a system had little effect on any of the measures of case mix. Independent variables from each of the other categories had important effects on hospital case mix. Hospitals in regulated rate-setting states (i.e., Maryland, Massachusetts, New Jersey, New York, Connecticut, and Wash- ington) had slightly lower case-mix indices 271 Hospitals having 100 percent admissions un- der rate setting would have had indices av- eraging about 4 percentage points less than the sample overall); more surgeries per 100 admissions, and higher proportions of outpa- tient to total revenue. The variables showing interrelationships among the case-mix measures suggest that higher case-mix indices are associated with higher values for births or surgeries per 100 admissions; higher proportions of births are associated with lower proportions of surgeries and vice versa, and higher proportions of charge-based payers are associated with higher proportions of surgical and obstetrical admis- sions, but lower proportions of outpatient to total revenue. Three proxies for differences in input prices and regional medical practices also were as- sociated with differences in the case-mix mea- sures. The wage index, a direct measure of unit labor cost levels and a proxy for urban location, was related to greater numbers of surgeries per 100 admissions. The more direct measure of urban location location in an SMSA showed that urban hospitals had higher case-mix indices than rural ones, whether due to more co-morbidities in urban areas, better medical record reporting, or other factors. Hospitals located in the Pacific region also had higher case-mix indices and higher outpatient proportions of total revenue, suggesting that in the Pacific region less acutely ill patients who in other areas may be admitted (reducing the overall case-mix index) were instead treated as outpatients. Variables measuring hospital capacity and utilization have statistically significant but practically negligible effects on the case-mix measures over reasonable ranges of differences in the independent variables. Discussion In our view, three of these findings on case mix are most important. First, the differences among the case-mix indices of these classes of hospitals are small, even though they are sta- tistically significant. Second, within each own- ership type, the hospitals affiliated with multihospital systems have indices slightly higher than those ofthe freestanding hospitals.

2 72 TAiBLE 5 Regression Results: Case Max Measures FOR-PROFIT ENTERPRISE lN HEALTH CARE Inpatient Outpatient Revenue Case Mix Surgeries per Births per to Total Variable Index 100 Admissions 100 Admissions Gross Revenue Dependent variable mean (whole sample)0.990 37.3416.632 0.109 Not-for-profit freestanding (reference group) dependent mean1.010 41.0528.027 0.115 Ownership and affiliation Investor-owned chain- 0.011 - 1.449- 3.237*** - 0.016* Not-for-profit chain0.007 - 2.007- 0.666 0.004 Investor-owned freestanding- 0.028*** - 4.865**- 3.707*** - 0.011 Government- 0.002 -6.913***0.096 0.011 Contract management and length of affiliation Contract managed in 1980 Chain affiliated in 1977 and 1980 Contract managed in 1977 and 1980-0.013 0.003 0.011 1.486 0.680 0.010 - 0.264 - 0.521 - 0.004 -3.401 - 1.244 0.002 Competition and regulation Sole community provider - 0.014 -3.834 1.513 0.020 Percent of admissions covered by rate setting Case mix and patient selection Case-mix index Inpatient surgeries per 100 admissions Births per 100 ~ . . admissions Percent of admissions covered by charge based payers Input cost and regional practice differences Wage index Located in an SMSA Located in Census Region 9 - 3.836E - 4*** 0.052* 0.013 2.148E - 4* 68.970*** 12.133*** 0.137*** 0.001*** 0.001 - 0.060*** - 4.278E - 4** - 0.341*** 4.854E - 4 - 1.729E - 40.215*** 0.072***- 0.001*** 2s.306*** 1.6660.021 0.055***- 1.585 - 0.38s0.032*** Capacity and facility utilization Number of beds 6.126E - s*** o.oos**0.006*** - 4.260E - 6 Occupancy rate (percent) 0.001*** 0.0640.001 - 7.431E-s Outpatient revenue to total gross revenue (rabo) 0.090* - 21.804**4.417 Average length of stay -- 1.639*** - 0.013***

OWNERSHIP AND ECONOMIC PERFORMANCE TABLE 5 Continued 273 Inpatient Outpatient Revenue Case Mix Surgeries per Births per to Total Variable Index 100 Admissions 100 Admissions Gross Revenue Medical education Interns and residents: total expense per bed 9.423E-6*** - 0.001 0.001*** 3.273E-6 Intercept 0.845*** - 62.497*** 3.007 0.079** N 535 535 535 535 Adjusted R2 IOMS compared to NFPMS 0.4335 - 0.018* 0.3656 0.3391 0.2413 0.558 -2.571*** - 0.020** NOTE: Levels of statistical significance are denoted by asterisks as follow: * probability <.10, ** probability <.05; ***probability <.01 that observed difference is due to chance. Although these differences were measured be- fore "diagnosis-related group (DRG) maximi- zation," they suggest that hospitals in multihospital systems may have had tougher cases or better medical records systems in these years. Clearly, medical records is an area where economies of scale could allow systems to at- tract better staff and purchase software to im- prove the positions of their hospitals. Finally, our results confirm others' findings that IO hospitals of both classes have been consider- ably less active than NFPs in obstetrics, a ser- vice traditionally considered unprofitable. Operating Revenues and Expenses Table 6 presents the results of regression models explaining differences in per-case rev- enue and expense measures. All these depen- dent variables were adjusted to control for case mix, wage index, and outpatient volume dif- ferences. Chain-affiliated IO hospitals had total pa- tient care revenues per adjusted admission $378 (21 percent) higher than NFP system hospi- tals, net patient care revenues per adjusted admission $185 (12 percent) higher, and home office costs per adjusted admission $57 (377 percent) higher. IO and NFP system facilities were much more similar on measures of op- erating, patient care, and administrative costs per case; these differences were not statisti- cally significant. NFP system hospitals showed no statisti- cally significant differences from freestanding NFPs in revenues or expenses, except of course in their home office costs. Freestanding IO hospitals also showed no statistically significant differences from freestanding NFPs on any revenue or cost measure. Unlike the system IOs, however, the patient care costs per case of the Freestanding IO hospitals appeared lower than those of the freestanding NFPs. Finally, government hospitals had lower revenues and costs per case than freestanding NFPs, espe- cially lower administrative costs. Hospitals contract managed for less than 3 years, whether by IO or NFP managers, in 1980 had home office costs per adjusted ad- mission $18 higher than hospitals that were not contract managed, but hospitals that had been managed for 3 years or more showed only slightly higher home office costs than those not managed ($17.72 - $15.46 = $2.26). Simi- larly, hospitals Foliated with systems for longer periods had lower home office costs than those recently joining systems. The presence of a mandatory state rate-set- ting program was associated with lower net patient care revenue per adjusted admission ($205 lower, if all admissions were covered under the program) of hospitals operating un- der it, as well as lower total operating expenses ($154 lower). Sole community provider status was significantly associated only with higher home office costs.

274 FOR-PROFIT ENTE~MSE IN HEALTH CAM TABLED Regression Results: Case-Mix-Adjusted Revenues and Operating Costs per Adjusted Admission _ Total Patient Net Padent Total Operating Total Patient Variable Care Revenue Care Revenue Expenses Care Expenses _ . _ . . Dependent variable mean (whole sample)$1,882.64$1,586.15$1,593.32$1,436.19 Not-for-profit freestanding (reference group) dependent mean1,926.191,611.631,630.141,515.91 Ownership and affiliation Investor-owned chain289.98***137.94**82.74-8.84 Not-for-profit chain-87.85- 46.604.02- 35.27 Investor-owned freestanding104.4422.3744.47-73.86 Government- 65.98- 80.09- 66.93- 87.56 Contract management and length of affiliation Contract managed in 198042.8639.6889.9454.76 Chain affiliated in 1977 and 1980- 58.48- 47.17- 73.13- 59.35 Contract managed in 1977 and 1980196.84164.57115.49135.35 Competition and regulation Sole community provider119.7956.6431.2625.11 Percent of admissions covered by rate setting- 1.51-2.05**- 1.54*- 1.15 Case mix and patient selection Case-mix index Inpatient surgeries per 100 admissions6.09***3.55***3.71***2.91 Births per 100 admissions-21.56***- 19.65***- 18.34***- 16.14 Percent of admissions covered by charge-based payers- 5.46***- 1.18- 2.43**- 2.55 Input cost and regional practice differences Wage index Located in an SMSA140.72**132.33***128.39***113.57 Located in Census Region 9245.66***185.28***210.49***81.55 Capacity and facility utilization Number of beds0.91***0.63***0.62***0.49 Occupancy rate (percent)3.70**0.57- 2; 13*- 2.37 Outpatient revenue to total gross revenue (ratio) Average length of stay Medical education Interns and residents: total expense per bed Intercept N Adjusted R2 IOMS compared to NFPMS 0.08***0.09*** 1,442.16***1,364.18*** 469473 520 0.10***0.09*** 1,568.09*$*1,561.87*** 4~87 0.4146 0.3814 0.3402 0.31 377.83*** 184.55*** 78.72 26.43

OWNERSHIP AND ECONOMIC PERFORMANCE TABLE 6 Continued 275 General and Interns and Residents: Variable Administrative Costs Home Office Costs Total Expense per Bed Dependent variable mean (whole sample)$694.48$19.10$ 276.87 Not-for-profit freestanding (reference group) dependent mean738.450.68547.56 Ownership and affiliation Investor-owned chain8.3471.38***-65.32 Not-for-profit chain- 34.7414.43***336.89 Investor-owned freestanding- 23.78- 0.788.69 Government- 77.04**- 1.8386.53 Contract management and length of affiliation Contract managed in 198024.1417.72***-134.31 Chain affiliated in 1977 and 1980-30.82- 8.75**- 154.22 Contract managed in 1977 and 1980109.69*- 15.46131.08 Competition and regulation Sole community provider1.9414.10*-60.33 Percent of admissions covered by rate setting0.090.053.53 Case mix and patient selection Case-mix index 1,807.76 Inpatient surgeries per 100 admissions1.18*- 0.18*-2.78 Births per 100 admissions-8.24***-0.3719.47 Percent of admissions covered by charge-based payers-1.68***0.12-6.88 Input cost and regional practice differences Wage index 920.11 Located in an SMSA50.68**2.34 Located in Census Region 9-14.38- 7.60*- 397.32 Capacity and facility utilization Number of beds0.26***0.012.22 Occupancy rate (percent)-3.13***- 0.141.38 Outpatient revenue to total gross revenue (ratio)--859.81 Average length of stay -39.38 Medical education interns and residents total expense per bed0.06***8.41E-4 Intercept916.65***12.07-2,935.69 N487472535 Adjusted R20.28640.46840.3168 IOMS compared to NFPMS43.0856.95***- 402.20 NOTE: Levels of statistical significance are denoted by asterisks as follow: *probability <.10; **probability <.05; ***probability <.01 that observed difference is due to chance.

276 As expected, case-mix differences were as- sociated with important differences in most measures of hospital revenues and costs. The results confirm the hypothesis that surgery is a profitable service for hospitals and obstetrics an unprofitable one. Each additional inpatient surgery per 100 admissions raised the hospi- tals' gross patient revenues by about $6 per case, net revenues by almost $4 per case, and total patient care expenses by less than $3 per case. On the other hand, each additional birth per 100 admissions reduced average gross pa- tient care revenue per case by about $22, net patient care revenues by about $20, but ex- penses by only $16. Higher percentages of charge-based payers also were associated with lower revenues and costs per case. The reason is not entirely clear, but may be related to fewer co-morbidities and shorter lengths of stay among the younger pa- tients most often covered by charge-based in- surance (an explanation of the cost difference) and a decreased need to "cost shift" in order to earn a target net revenue figure (a possible explanation for the relationship between higher charge-based percentages and lower gross rev- enues). This second hypothesis is supported by the fact that gross revenues declined more than five times faster than net revenues for each percentage increase in charge-based pay- ers. Urban hospitals and those located in the Pa- cific region also had higher revenues and costs than rural hospitals and ones located else- where in the country, perhaps due to regional practice differences. The only exceptions to these patterns were (1) in home office costs, where we found no difference between urban and rural hospitals, and (2) in general and ad- ministrative and home office costs, where hos- pitals in the Pacific region showed somewhat lower costs. Hospitals' capacity and utilization of staged beds, and their financial commitment to grad- uate medical education affected their charges and costs to a much lesser degree than many of the other factors in the equations. Each ad- ditional bed above the mean of the sample added less than $1 to the hospitals' revenues and costs per case, but more to revenues than to costs. Each additional percentage point of occupancy increased the hospital's total pa FOR-PROFIT ENTERPRISE IN HEALTH CARE tient care revenue per case by about $4, but reduced total operating expenses per case by about $2, total patient care expenses per case by about $2.50, and the general and admin- istrative component of total operating costs per case by a little over $3. Finally, in 1980, before the Medicare revenue enhancements for med- ical education, each additional dollar of ex- pense for residents and interns per bed added only about 9 cents to the total patient care expense per admission and about 8 cents to total patient care revenues per admission. The financial commitment to graduate med- ical education varied across the classes of hos- pitals. In the raw data, NFP chain-affiliated hospitals averaged about $733 in expenses for interns and residents per bed per year. NFP freestanding hospitals averaged $543; govern- ment hospitals $199; but IO freestanding hos- pitals only $34 and chain-alPliated IOs less than $3. Controlling for a wide range offactors other than hospital class, the regression model pre- dicts that the average differences among the classes would be much less. The highest pre- dicted value is that for members of NFPMS intern and resident expenses per bed per year of about $884; government hospitals were the second highest ($634), followed by NFP and IO freestanding hospitals ($548 and $556, re- spectively). IO chain hospitals in 1980 were predicted to have the lowest financial com- mitment to medical education in our sample ($482~. Discussion In 1980, hospitals in IO systems charged significantly more per case, even after adjust- ing for case mix differences, relative input costs, and other factors, than did freestanding or sys- tem-nf~liated NFP hospitals. Freestanding IOs appear to have charged somewhat more per case as well. Even after discounts, allowances, and reserves for bad debt, the net patient rev- enues of IO system hospitals one measure of their costs to their communities were higher than those of NFP hospitals. However, we found no significant differ- ences in patient care expenses or total oper- ating expenses per case across the classes of hospitals. Indeed, the ownership and affilia- tion dummy variables as a group did not to

OWNERSHIP AND ECONOMIC PERFORMANCE "ether explain a significant part of the variance in hospital costs per case. These findings reinforce the hypothesis that, during the era of cost-based reimbursement, IO hospitals earned higher profits by charging more rather than by being more efficient. Mile this strategy appears to have been followed by both classes of IO hospitals, it was stronger among those in multihospital systems. Home office costs per case were consider- ably higher in IO systems than NFP systems. This may indicate that the IO systems were more centrally organized than NFPs and pro- vided more services centrally. It also, how- ever, may reflect the more active building and acquisition efforts undertaken by the IOs to support their faster rates of growth. Markups and Profitability Table 7 presents the results for the markup and profitability dependent variables. Gross patient care markup, defined as total patient charges divided by total operating costs, is a conservative measure of the profit margins hospitals built into their charges. The revenue deduction ratio measures contractual allow- ances, free care, discounts, and bad debts as a proportion of gross patient revenues. The third measure shows the proportion of the hos- pitals' total gross revenues derived from non- operating sources. The total markup ratio, total revenue divided by total expense, measures the overall profitability of the hospital. Return on total assets (total net income divided by assets) and return on equity (net income di- vided by owner's equity not to be mistaken for the add-on capital payment that Medicare pays to IO hospitals) are measures that relate profitability to the investment required to gen- erate it. Chain-a~liated IO hospitals took higher markups than NFP multihospital system mem- bers, both in total (12.8 percentage points higher) and on their patient care services (14.2 percentage points higher), but had higher de- ductions from Moss revenues due to their higher charges, and lower ratios of nonoperating rev- enue to total revenue. These pricing strategies allowed IO system hospitals to generate better profitability as measured by return on assets (5.5 percentage points higher) and return on 277 equity (34 percentage points higher) than NFP system hospitals in 1980. NFP system members, somewhat surpris- ingly, had significantly lower markups on both their patient care services and in total than Freestanding NFPs, and therefore had lower operating surpluses than the freestanding NFPs. Freestanding and system NFPs were virtually identical, however, on the proportions of their revenues derived from nonoperating sources, and higher on this measure than either class of IO hospital. Nonoperating revenue was a larger portion of the total gross revenues of government hospitals than of freestanding NFPs, but these classes of hospitals differed little on markups and profitability. Freestanding IO hospitals achieved a better return on equity (ROE) than freestanding NFPs on only slightly higher markup ratios and while earning less nonoperating revenue as a per- centage of total revenue. Thus, their higher ROE was due largely to the higher proportion of debt financing among the IO hospitals than the NFPs. That IO hospitals of both classes had lower proportions of nonoperating reve- nue to total revenue than the NFP or govern- ment hospitals probably reflects the fact that IO hospitals are rarely the beneficiaries of ei- ther private philanthropy or tax revenues, im- portant sources of nonoperating revenues for some NFP and government hospitals, respec- tively. The length of affiliation with chains or uncler contract management had no significant rela- tionships with hospital markups and profita- bility. We found no evidence that sole community providers take advantage oftheir status to Large higher markups or generate "monopoly" prof- its. We found no statistically significant rela- tionships between the presence of a state rate- sethng program and levels of markups or de- ductions from gross revenues. However, hos- pitals in rate-setting states had slightly lower returns on assets than hospitals elsewhere. We found no relationship between either the overall case-mix index of a hospital or its proportion of surgical cases and its markups or profitability. However, hospitals with higher proportions of maternity cases had lower prof- itability. Hospitals with higher percentages of charge-based payers showed lower revenue

278 C~ o o ~ = .- ~ 5 U. O C':l c: ~ L' ~ o 3 C~ - o ~ ._ ~ _ o ~ :t eC o ~ E~ ~ 3_, '~ ~ p: ~ ~ C~ o ~ ° ._ - ._ D o C~ U. e~ . . - o ._ V) U) C) ~0 a; o ._ C~ ~ o , ._ Q) - ~; =; ~ D~ ·_ C~ Ct P~ ~ ·V' o ~ o C' C) ~ Ct ._ C~ * * * * * * oou: IO oo ~ CD C]CD ~C~- 00 00 ~_ ~ C ~__CDOO U) _o CO o _ o ooooo o oo o o o o ooooo o 1 111 * 1 0 ~ o ~ _ ooC~o o C~ u=,cO O O 0 0 8800` ~0 0 0 0 0 0 0 0 0 0 0 ~0 1 1 1 1 1 * 1 * * * 0 c ~ 0 0 ~oo 0 c~ oo ~_ c ~_ ~0 c ~c ~ ~00 0 0 0 0 0 0 _ 0 _ _00 0 0 0 0 0 0 c ~0 1 '1 1 1 ** 1 ** * _4 ~0 ~c~ ~~ ~~ 0 c ~cc c~ 0 _ _ 0 0 _ _ ~ 0 0 0 0 00 0 0 0 0 _ 0 0 0 0 0 00 0 0 0 0 _ 0 1 1 11 1 1 * 1 ** . ~ =4 ~ _ c~ oo 0 c ~c ~0 cs ~ ~0 0 _ 0 _ c ~ _ _ 0 0 0 0 0 0 0 0 0 0 . . . . . . . . . . . . O 0 0 0 0 0 0 0 0 0 In 0 1 1 * u, 1 * * _ ~3 c ~0 ~C53 ~0 ~0 0 c ~c~ c ~ _ 0 c ~c ~_ ~ _ _ 0 0 0 0 0 0 0 ~0 . . . . . . . . . . . ~4 0 0 0 0 0 0 0 c ~0 1 1 1 c; ~. ~ 7 s e ~ ~e ~ ~ a, E ''3 ~ ·-~ce~ -~: ~ . &- ~ ~ a ~· x e C O O _ q _ _ - · ~ ~e c~ t8 ~= e 0 ~ ~: O e ~ 0 ~ 7 0 7 E ~ ~ ° e ~ :1 ~Z 0 ~ C) C)

279 ~ 8 8 ~o ~ 8 ° o ~ - = ~ ~ C~ o o o o _ o o o _ o 0 o o V, C~ o V ~ 8 8 ~o o g 8 ~o o ai ° ~ o ~ ~ _ o 0 0 0 ~ o 0 0 ~ o ~ o0 * ~U:V o o 8 - o o 8 ~ Oo o C ~o o o _ o o o _ o ~ oo ~-o ~o ~o uD * ~ o Co * * I ** Il~ ~Cq o o ~ X ~o O X o ~ ~ CD ~O O _ ~O ~ O ~O O ~ * * ~ 8 ° o -o U) g X o ~ o ~ o C~ o o o o ~ o o o 0 o ~o o 1 1 1 U. U' ~8 ~o ~ 8 ~ oE ° ° o 0 0 ~ o o 0 _ 0 ~o 0 1 1 1 1 1 ,,, U3 ;~~e0~ =138= (~z=~-~t ~3 a j I 3 o C~ Ct ~3 " U. ~Q C) - o C. s: C~ C .= U) e~ ._ . U) ._ ~ ~ c~s C~ ~ o o~ cO (D ~ - .. } o ~ z~ ~ - ~5

280 deductions, and required lower markup ratios (measured as gross revenues to costs) to achieve higher profitability. Hospitals in areas with higher wage indices took higher total markups of revenues over costs. Pacific Coast hospitals earned less of their total revenue from nonoperating sources than hospitals in the remainder of the county. Capacity and utilization variables were sta- tistically significant, but explained only small proportions of the differences in markups and profitability across the classes of hospitals. In- creased occupancy allowed hospitals to earn higher markups on patient services and in to- tal, and was therefore associated with in- creased revenue deductions and higher return on assets and return on equity. Increased out- patient proportions of total revenue were as- sociated with lower patient service markups and deductions, but with higher proportions of the hospital's revenue coming from non- operating sources. Longer average lengths of stay were associated with lower return on as- sets (1.2 percent decrease for each additional day). Finally, higher medical education costs per bed were associated with slightly lower patient and total markups in the hospitals. Discussion These results support our earlier findings on revenues and costs: the IO hospitals in mul- tihospital chains took higher markups and earned higher returns than did NFP hospitals of either type. Freestanding IO facilities were somewhat more moderate in their pricing practices, but both classes of IO hospitals earned higher returns on equity than Me NFP hos- pitals. NFP system hospitals took relatively lower markups than freestanding NFPs, and earned only modest returns. All hospitals, regardless of ownership, must earn operating surpluses in order to remain attractive to lenders of debt capital and to be able to replace their plant and equipment. All five classes of hospitals in our sample, on av- erage, were able to do so in 1980. However, the higher surpluses of the IO hospitals, cou- pled with their ability to issue stock, give them a flexibility in developing capital that is not matched by the ability of the NFP hospitals FOR-PROFIT ENTERPRISE IN HEALTH CARE to attract and accept donations or government grants. Capital Structure Table 8 shows the results of equations ex- amining capital structure variables. Chain-af- filiated IOs had higher debt-to-asset ratios than NFP system members (about 0.66 versus 0.461; higher capital costs as a percentage of total operating costs, whether excluding or includ- ing Medicare's return-on-equity payment; younger plants, as measured by the ratio of accumulated depreciation to depreciation ex- pense (2.22 versus 6.50 years); and lowerfixed- asset values per bed ($6,533 less). While we found no statistically significant differences be- tween the NFP hospitals in our sample in sys- tems and those that were freestanding, the NFP system sample showed slightly lower debt- to-asset ratios, current ratios, and capital costs as percentages of total operating costs; older plants; and lower fixed-asset values per bed than freestanding NFPs. Freestanding IOs had higher debt-to-asset ratios (0.15 higher), younger plants (1.95 years younger, and considerably lower fixed-asset values per bed ($22,386 less) than freestanding NFPs. On the fixed-asset value measure, the freestanding IO figure is about half the NFP freestanding one. Government hospitals had lower debt-to-asset ratios (0.13 lowers, lower capital costs as a percentage of operating costs, and lower fixed-asset values per bed ($9,559 less) than freestanding NFPs. Hospitals contract-managed for less than 3 years had higher debt-to-asset ratios (0.19 higher) than freestanding NFPs that were not contract-managed. Our results also showed contract-managed hospitals as having less de- preciated plants (about 2 years lower in ac- counting age of plant than freestanding NFPs). Variables measuring the length of affiliation with a chain or under contract management had no significant effects on capital structure. Our variables reflecting competition and regulation in the hospitals' environments yielded only two strong relationships. First, sole community providers appear to not have capital structures significantly different from those of the average hospital in our sample. Second, hospitals operating under rate-setting

281 s~ C~ - C~ ._ - . . U' - U) o ._ V, U) - U. U. ·c _ Ct C) ~o ,; ~S _ ~ , 1 ~ b4 ~ O J t~ ~) ~ ~ O - C;) _ :) o- ~C~ - O ~ _ ~ '- C.) C) ~ o ._ - C~ ~; C O ._ _ D; ~a' < CD ,CI-' C~ ;> C~ Q) au ~ £ ~: s" ~ o ~~ ~ 3 c~Q _o .. _ ooC~ ~_ CO CO C~ C~0 .. ~h~ C ~_ C~ U. __C~ oo o O 1 * * _ C~ C~ C~ oo U: ot _ 1 1 * * * * * * O _ =) C~ U. ~ oo ~n C~ U: C] 0 c~ 1 1 ~_ ... CD00 "4C~C~ CD ~O _ ~ 1 1 * * * co ~oo .. 1CO 1~ ~ . c~ ~ ~ 1 * * o, c~ o c~ - * * * * * * co ~'~ 0 oo cO ~oo ~cO ) ~o u, o ~ C~ O 0) - CS) C~CD _ O U) O O O c~ - - - - oc ~o o - o o o 1 1 1 1 1 1 ] - ~ CD~DO ... ooo 1 ** * ** * ~0~ - ooo ~c~ ~ ~oc~o o ~. ~. . oo ~o o 11 1 * * c ~co c ~oo ~o, 0 ~n ~ ~O O C~ CD o cO ~c ~- ~o o o o o - o o o o o ~o o 1 1 1 1 1 1 1 , ~'o c] oo_ c ~ ~CD - - ~- o - ~- ~o o o o o o o o 1 1 1 1 1 1 1 u) ~ . . o o ~4 : . to ~ ~ ~ o o ~ ~ s~ =~ ~ ~ cD - ~ ~ o 3 Z O * * * * * * * * * o C~U) C~ o~ _ ... . o o o o 1 1 ooU, ~CO ooo ... ooo 11 o oo C: _ ~_ _ ~ ~ ~ ~ ~ 0 ·- _ ._ 3 ~ c; ' . ~ ~ 0 E =° Y q) ~ 0 Y " '- =- o ~ ._ _ ._ _ ~ ° o ~ ~ ~ U, ~ ·- ~ ~ ~ ~ ~ ;- Ct ~ ~ ~ ~ ££ ~o ~5 ._ _ s~ .- o ~ ,~ C~ o o C~ ~ C: * * _ o~ oCO .. oo 1 * C~ o o _ C~ C ~o o . . . _ o o 1 1 1 1 ** C~ CO 8 0 . . o C~ o ._ ~ c) t: ,o _ ~._ CO ~. - ~ ._ ~ c: ~ _ ~ ~s ~ ~ _ ~ ='- ~ £ 3-' es .X ~ ~ ~ _ =~ ~ m~ C: * C~ C~ o o 1 00 oo 1 CD o J C~ ~ S~ C) ~ V' O

262 U) U. U) X =, _ CC et s~ o C' E~ t4 ~o . c: ~ C: ~ Ct Ct C~ C) ~ C_ << o ~o ~ 't ,:;, ~ C) C: p4 - _ Ct ·= C~ C: ~ o := - o . - o ~ _ _ ~ au Q~ o e~ . :> * * * * CO * * * ~* CD~ O ~ ~_ c ~_c~ ~O ~a, _ ~_ oo 0 e ~0 _ ~0 c~ ' ~ ~C~ _ C! ~C~ _ ~C ~ ~ 1 1 1 ~o CD U. * 1 * * ~ * * ~o _ oo _ o COC~ co _ o ooo ~oC ~_ C~ . . .. . .. . . o ~o oC~ o ~oo 1 1 1 1 1 C~1 * * 1* * _ ~c~ u: ~c ~m Cl ~0 Ut~ ~_0 0 ~ _ ~_ C ~_ Cl 0 0 0 ~0 ~U, 0 C~ 1 1 11 1 1 ,,,~ ** 1 ** C;3 * ~* ~ O CD 00 C ~- C ~=) oo oo 0 ~ ~oo C ~C~ 0 0 CO _ ~oo ·_ t 0 0 ~o 0 ~ 0 C] CC 1 1 1 1 1 1 \t, 1 1 * * ~n ~oo CD O ~ Oo ~=n CO O _ O . . . . . . . . . . _ 0 C ~0 _ 0 _ ~CS 0 0 1 1 ~1 U ~U: * 1 1 * * ~_ ** 0O ao ~ _ C] ~0 ~C 2 ~_ ~ 0 C~ _ 0 ~C ~0 CO 0 C~ 0 0 0 C ~C ~C~ 0 0 C ~0 0 0 - 0 ~0 0 1 1 1 1 ,,d, c) ~ ._ _ J ~ ~ _ ~ O ·_ C) - C~) Ce Ct 3 ~_ _ ,, Z ; ,, C) - "CS ~ {_ ._ ~ O ~ U) C) s~ o4 C) C~ _ _ ~ C~ C~ ~ ~. - Q) ~ C~ z - J C~ C. =; £ ° U, C~ ~ O o Ct _ _ o V ._ "Q * * * .^ o v ` - D. * * o - v ~ - - 9 * . . - 4- e~ Ct o ._ Ct o C~ - o - o ~0 e~ - C) C~ C. ~, - 5 o £` ,5 O ~ ~ er =, O ~ <¢ O O ' 4 ~ 5 0 C) C; t0 0 ._ _ ~ ~ ._ ·_ C-) C~ C~ _, ~ CO, O C~ - . _ .~ ._ C~ Ct ~ CJ C ~ - C~ C} ._ . _ _ :: ~ Ct U, ~ O _ U7 3) >= .= - C.) 5 zO~

OWNERSHIP AND ECONOMIC PERFORMANCE programs had higher debt-to-asset ratios and lower current ratios than those in less regu- lated states. Me higher debt-to-asset ratio may indicate both that hospitals in these states re- quired more borrowing and that these states assisted hospitals to borrow by issuing "com- fort orders" explaining that debt service was an allowable expense under the rate-setting program. Few other independent variables were strongly associated with differences in capital structure. Higher case-mix indices were as- sociated with slightly higher capital cost as a percentage of operating costs and higher fixed assets per bed, confinning the need for so- phisticated and costly equipment to render more intensive care. Higher numbers of sur- gical procedures per 100 admissions were re- lated to higher debt-to-asset ratios, lower current ratios, and higher capital cost as a per- centage of operating costs. However, we found no significant relationship between higher sur- geries and higher fixed assets per bed. Higher percentages of births among a hospital's case load were associated with lower capital cost as a percentage of operating costs, but had no strong effects on any of the other measures of capital structure. Higher wage indices, often indicating urban areas, were linked to higher debt-to-asset and lower current ratios. Capacity and facility utilization had statis- tically significant but small effects on capital structure ratios. Higher occupancy rates re- sulted in lower capital cost as a percentage of total operating costs, showing the effect of spreading the fixed capital costs over greater volumes. Higher occupancy also was present in younger physical plants and in hospitals with higher fixed assets per bed. This may indicate that new physical plants tend to be built in areas of need where the capacity will be uti- lized by growing populations. An alternative hypothesis, that patients are attracted to new buildings, is less in accord with our experi- ence. Higher proportions of total revenue from outpatient services, including ancillaries, were associated with lower capital cost as a per- centage of operating costs. Finally, hospitals with longer stays apparently also had some- what older physical plants. 283 Discussion Despite their better access to capital, IO multihospital systems appear to be more ju- dicious in its use, as shown by their lower fixed-asset values per bed. The significantly lower fixed-asset figures for freestar~ding IO hospitals are consistent with their lower case- mix indices and smaller size, and indicate hos- pitals which probably provide fewer sophisti- cated services. Consistent with national studies, IO hos- pitals of both classes had higher capital costs as a percentage of their total operating costs than other hospitals. This is due to their higher leverage, and also, in part, due to the younger age of their physical plants on accounting-based measures. However, since "accounting age" is affected by sales of hospitals (revaluation of the assets, which Medicare until recently allowed, "reset the accounting clock"), it is not clear whether other hospitals in fact will require replacement or major capital projects earlier than IO hospitals. Clearly, in the short run, IO hospitals would be less advantaged by a payment system that would include capital as a fixed-percentage add-on to the DRG rates. IO hospitals might be disadvantaged over the long term as well, unless their higher capital costs in fact show the financing of more "future costs," rather than being accounting artifacts of more recent purchase of plant assets of the same physical age as those of the NFPs. Activity and Productivity Table 9 presents the estimated coefficients for the activity and productivity regressions. The total asset turnover ratio measures the revenue dollars a hospital generates for each dollar invested in fixed assets, and the current asset turnover ratio measures revenue per dol- lar of current assets. Case flow is measured as admissions per bed, and is, thus, directly re- lated to occupancy and inversely related to length of stay. Chain-affiliated IO hospitals showed better use of assets on those measures based on dol- lars total and current asset turnover ratios than NFP system hospitals (ratios 0.39 and 2.32 higher), due to their higher charges and lower asset values. However, they showed

284 - s" ~ ~ v v v ~ - ~ ~ - e~ ~; c) o 3 v U) o ._ e~ ·~ ._ C) o s~ ._ ._ C) . . U) - U. P:; o ._ U. U. L. ~0 k o _ ._ V~ V, ~ ~_ C: ~ o C) ~ o ._ - C~ _ V) V, ¢ o Ct ~ * * * C ~CO Cl C ~CD ~00C ~CO 00 ~ C ~U~C ~_ ~O CD O C ~C~ o o o oo o oo oo o C] o o o o oo o oo o o 1 1 11 11 * * * * * * * ** * * * o _ ~ C ~ ~X ~o, 0 o o C ~ ~C ~oo ~Cp ~ CC ~_ C ~C~ ~o ~o ~_ o o ~0 U~ _ _ ~U: C ~C ~o ~o CD CD 1 1 1 1 1 1 ~ * * * ' ~0 u, ~'o ~cO a ~oo u) co ~_ ~co c ~ o 'o ' ~o ~cO ~o ~o a o co c~1 1 1 1 1 1 * * * * * * 00{D ~i~ 0 ~_ OCDX C~C~ C ~LO c ~ o'c ~o's ~ ~ ~.~ d4C~C~ OO OOO 11 ~o - o o o ~o lo o o o 1 1 1 ** c ** o * * ** -* * ~C ~0O ~CD - Ct'{DCD ~- ~_ 0 ~ cn0 0 - ~c~00 c~Q U:o U, ~_ _ ~o ~oo COo .. . .. . r~..... .. __ o o_ o ~ooooo _o o 1 1 1 C C .o r ~ ~ ~ _ ~ ~9 ~ ~ ~ e,- ~ v ~ 9 E ~j ~Z O C) ~

285 U. ~ * * 1 ** 1 * * * ~*** ~o * ~ =, ~0 cn ~ co o8 ~g~ g ~ ~o ~ o o o o ~o _ o _ o ~o o 1 1 1 0 1 ** * 1 * ~* _ ~o ~_ ~o o C ~C ~o C~ C~ o o ~o ~ _ C ~o C~ o o C ~o o ~ C ~CO o ~o C~ 1 - 1 i,~ * * * 1 * * * * * * * * * ~3 * CO * ~o C ~CD ~_ ~_ CO C ~C] ~oo ~- . o ~ U) C~ CO _ o U) o ~ _ o o ~C ~o C ~_ ~U. o 1 1 1 - c ~,,, 1 * ** ** ~* C ~433 _ o ~ ~ o 8 ~x 28 ~ 8 ~ ~ o o _ o _ ~o _ o ~_ ~o ~ 1 1 1 ,l, * * 1* * C ~CO o ~_ t ~C~ o- ~CO U ~* o 0 ~o 0 ~ _ ~o o o ~ o ~ C7 · · . . . . . . .. . . o o _ o o o o o __ C ~o o 1 1 1 1 ,,~, c) C1 ~ _ .o C ~ ~ ~_ - o ~>% .. ~ S . ~ ~ ° ~ ~d 5 D E 5, 7 ~ ~ e o cn ) ~4 , z ~ - ~ ~ ~ o ~5 c~ 9 o C~ - - o V - D D * * * .^ O V _ ._ D O * * O^ V o L' * 3 _ c~ c~ ._ ~7 co ~s 0 ~:s s~ cc c~ c~: ,bO c~ c' ._ . .u, ~ _ ~ e~ ~o 0 _ c~ { ;> ._ .. O Z ~ ._

286 weaker performance on case flow (2.4 fewer admissions per staffed bed per year), a pro- ductivity measure based on use of beds in ser- vice, due to their significantly lower occupancy. However, despite their lower occupancy, IO chain hospitals employed 0.28 fewer FlEs per adjusted average daily census than NFP sys- tem hospitals. NFP system members showed slightly better total asset turnover numbers than freestanding NFPs, but showed no other significant differences from NFPs on these measures of activity and productivity. Gov- ernment hospitals were also similar to free- standing NFPs on these measures. Unlike the chain-affiliated IOs, however, freestanding IO hospitals did not use fewer FlEs per average daily census than NFP hospitals. Hospitals chain-affiliated for 3 or more years showed higher turnover of current assets (ratio 0.65 higher), better case flows (2.4 more ad- missions per bed per year), and higher occu- pancy than hospitals that had joined multihospital systems more recently, but showed no better productivity in the use of personnel. Hospitals newly contract managed had fewer admissions per bed and lower oc- cupancy than other hospitals in the sample. Case flow was faster and occupancy higher in hospitals in rate-setting states (several states encouraged more admissions per bed through low-occupancy penalties in their reimburse- ment systems), but activity and productivity apparently were little affected otherwise by our competition and regulation variables. Hospitals with higher case-mix indices ap- parently had somewhat lower total asset turn- over, primarily due to their higher asset values. They also showed higher occupancy, graduate medical education (GME) costs, and ratios of personnel to daily census than did hospitals with less intense caseloads. Higher numbers of surgeries and births per 100 a~nissions were associated with better case flow and somewhat higher FTEs per adjusted average daily cen- sus. Higher proportions of charge-based pay- ers also were associated with better case flow, perhaps through shorter stays. Higher area wage indices, perhaps due to their relation to urban areas, were associated with slightly higher asset turnover ratios arid higher GME costs, but not with better per- sonnel productivity. Pacific Coast hospitals FOR-PROFIT ENTERPRISE IN HEALTH CARE showed lower current asset turnover, occu- pancy, and GME costs, but better personnel productivity than hospitals elsewhere. Larger hospitals had lower current asset turnover, higher occupancy, and slower case Dow (0.009 fewer admissions per bed per year for each additional bed) than the smaller hos- pitals in our sample, but were not different in personnel productivity. They also had higher GME costs. Occupancy, oddly, appears to be directly related to the number of FTEs per adjusted average daily census (0.025 increase per each 1 percent increase In occupancy). More expected results are that the proportion of a hospital's revenue from outpatient services also increases FTEs per adjusted daily census (0.086 additional FTE for each additional percent of outpatient revenue), as does GME. Longer lengths of stay are also associated with higher occupancy. Discussion . IO hospitals of both classes, due to their higher markups and lower asset values per bed, generate more revenue per dollar of total or current assets than do NFP hospitals. They also operate with fewer staffper occupied bed. While IO hospitals did not translate their higher productivity into lower costs per case in 1980, they achieved better productivity despite op- erating at much lower occupancy rates than NFP hospitals. Thus, IO hospitals, particularly those in systems, stand to benefit more Man NFP hospitals from improvements in facility utilization, which will allow them to spread Heir higher administrative costs and lower Heir overall costs per case. While their higher charges would disadvantage them in contract- ing with organizations like PPOs and HMOs which can ship blocks of patients, their cost functions might encourage the IO system fa- cilities to offer the deeper discounts necessary to secure greater patient volumes. CONCLUSIONS This study of a national sample of commu- nity hospitals has shown a number of differ- ences in economic performance across hospitals of different ownership and affiliation classes from which differences in strategies may be

OWNERSHIP AND ECONOMIC PERFORMANCE inferred. On 24 of our 28 measures of case mix, revenues and operating costs, markups and profitability, capital structure, and activity and productivity, statistic tests showed that hospital ownership and system affiliation were significant factors in explaining the differences across hospitals. However, on most measures except revenues per case, home office costs, and debt-to-asset ratios, other factors such as hospital capacity were also important explan- atory factors. Ownership and Chain Affiliation The clear pattern that emerges from this study is that in 1980 there were much greater similarities among hospitals of the same own- ership type (IO or NFP) than among hospitals of the same organizational type (multihospital system or freestanding). IO hospitals, whether freestanding or members of systems, were similar on the majority of measures in the study. The NFP members of multihospital systems in our study differed from freestanding NFPs only on the measures of home office costs, markup ratios, and asset turnover. Clearly, in 1980 the strategies and performance of IO and NFP multihospital system hospitals had not converged as some observers believe they have today. Functional Strategies The study suggests several conclusions about the functional strategies of IO and NFP hos- pitals. First, patient selection by the IO hos- pitals appears largely a function of siting and service-offering decisions. In 1980, as in 1984, less than 30 IO system hospitals, and none in our sample, were Medicare sole community providers, indicating that their service areas lacked competition. IO hospitals are primarily (74 percent) suburban or urban, and located in areas of relatively more rapid population growth and moderate bed-to-population ratios than characterize the home counties of NFP hospitals. They appear to have preferred to be Me second hospital in town, which allows them to differentiate their service offerings from those of other providers. This type of service deci- sion was shown in 1980 in their less frequent offering of obstetrical services, ones which ap 287 pear to be unprofitable among the hospitals in our sample. IO multihospital systems in 1980 also preferred to locate in areas where Blue Cross paid charges; this resulted in higher per- centages of their patients being covered by charge-based insurance than the patients of the other classes of hospitals in our sample. NFP multihospital systems, on the other hand, appear to have hospitals more evenly distributed across locations ranging from cen- tral cities to rural communities. Although sev- eral NFP system hospitals in our sample are sole community providers, on average, the NFP multihospital system facilities are located in areas with more competition, as indicated by bed-to-population ratios, than any other class of hospital. Second, in 1980 pricing was the key to the IO hospitals' higher profitability. Services in IO systems in 1980 cost the public more than in other classes of hospitals, even when ad- justed for differences in case mix, wage rates, and outpatient volumes, and controlling for a wide range of other factors affecting costs. The NFP system members in our sample, on the other hand, earned somewhat lower returns than other classes of hospitals by marking up their patient care services less in relation to their costs. Third, the financial strategies of IO hospi- tals of both types enabled them to earn higher returns for their investors and gave the IO hospitals an advantage in access to capital. These strategies combined higher patient service prices with financing a higher proportion of their assets through debt than the NFPs. Finally, the costs of medical education in 1980 were virtually all borne by NFP hospi- tals. The direct expenses associated with train- ing interns and residents were dramatically higher in hospitals in NFP multihospital sys- tems than in IO ones. While not quantified in this study, it is intuitive that the indirect costs of teaching were also higher in NFP multi- hospital systems. Multihospital System Economies This study does not show evidence of net economies achieved through multihospital system operation in either the IO or the NFP sector. The total patient care costs and total

288 FOR-PROFIT ENTERPRISE IN HEALTH CARE operating expenses per adjusted admission of tats in less challenging environments. This the muitihosoital system hospitals were not strategy is open to all systems, but not to free- standing hospitals. Second, the greater capital available in systems makes them more able to offer their own insurance products. Finally, the IO chains cost and profit performance has been achieved in hospitals with generally low occupancy. Winning or offering contracts to improve occupancy might result in greater economies of scale in the IO hospitals, allow- ing them to achieve their historical profit mar- gins despite the necessarily lower net prices of services. IO hospitals might use these fac- tors to erode the NFP hospitals initially greater attractiveness to payers, but they will need to run faster and farther than the NFPs to do so. Second, faced with potentially shrinking margins, IO systems may redeploy capital to develop other sources of revenues. Humana, for example, was initially a nursing home com- pany, but left that business to concentrate on hospitals. In recent years, it has sold more hospitals than it has acquired, and concen- trated on freestanding centers and insurance products. If changing market economics force hospitals to become price-takers and allow in- surers to be price-setters, IO systems may in- creasingly redefine themselves as insurers to capture the higher margins possible Tom other levels of the health care industry. NFP hos- pitals, whether freestanding or in systems, may be slower to change their historical missions. Third, their higher capital costs as a per- centage of operating costs would disadvantage IO hospitals in the short run if a method were enacted to pay for capital as a fixed percentage increment to operating costs. This disadvan- tage would persist in the long run as well to the extent that the IO hospitals higher current capital costs are an artifact of the revaluation of order assess on purchase, rasher then related to physically newer buildings. Fourth, the current magnitude of the in- direct teaching cost adjustment under Medi- care gives teaching hospitals an edge over nonteaching ones that may become even more important as the payment formula becomes based more on national averages. Along with their desire to integrate more vertically, this may be part of the reason that IO chain hos- pitals have sought to acquire several major teaching hospitals recently. However, the fu significantly different from those of freestand- ing NFPs. This may be due in part to the incentives of He Medicare reimbursement system in place in 1980, under which hospitals were paid on the basis of costs rather than revenues. However, it is also important to note that these operating cost and patient care cost results were obtained by system hospitals de- spite their home office costs, a category of costs not borne by freestanding facilities. The total general and administrative costs per admission (including home office costs) of the IO chains were only slightly higher than those of free- standing NFPs, and those of multihospital sys- tem NFPs were lower than those offreestanding hospitals. This suggests that the home office operations of both types of systems, in the main, substituted for costs borne by hospitals locally and may, in fact, have resulted in some offsetting direct cost savings as well. Potential for Future Success The environment that hospitals of all classes face in 1984 is considerably different from the world of 1980. HMOs, PPOs, employers, and insurers have begun to shop for hospital ser- vices. Medicare has set prospective prices, and other public payers have experimented with contracting with selected hospitals. New types of providers such as surgicenters have also emerged that chip away at the edges of hos- pitals traditional markets. This snapshot of the behavior and performance of hospitals under cost reimbursement, while not conclusive, may allow some inferences about the potential suc- cesses of the classes of hospitals in a more price-competitive environment. First, the considerably higher prices per ease among the IO system facilities, wider geo- graphic dispersion (rather than regional con- centration), and their less frequent offering of outpatient and obstetric services may put them at a disadvantage to NFP hospitals In winning contracts to provide services for organized lo- cal groups of patients. However, this expec- tation could be tempered by three factors. First, through multiunit operation, systems can sub- sidize lower prices and profits in areas where competition is keen, with profits from hospi

OWNERSHIP AND ECONOMIC PERFORMANCE ture of this adjustment factor is in doubt. Cur- rent proposals suggest halving the indirect al- lowance factor and freezing allowable salary and benefit costs per intern and resident at some historical level. The advantage of more favorable reimbursement to teaching hospitals may be quickly eroded. Finally, for multihospital systems of both classes, the challenge of the new environment will be to use their greater capital resources effectively and achieve the scale economies that theory suggests should be possible through system operation. Neither the IO nor the NFP multihospital systems demonstrated solid ev- idence of such economies in 1980. Basic issues such as how to motivate physicians and staff, reduce the costs of service, increase patient volumes in maturing markets, anticipate the behavior of third-party payers, and develop new revenue sources will be a large part of the new equation. The question of relative fu- ture success thus depends significantly on the managerial creativity and leadership each sec- tor of the. industry brings to solving problems. On these factors of production, no sector of the industry has a monopoly. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical assistance of George D. Pillari. This research was supported in part by As- sociated Hospital Systems. The views ex- pressed are solely those of the authors. 289 NOTES Using a matrix for each state, these data were used to construct the "percent charge-based" and "percent of admissions subject to rate-setting" variables de- scribed below. 2 Watt, T. M., R. A. Derzon, S. C. Renn, and C. J. Schramm, T. S. Hahn, and G. D. Pillari (1986) The comparative economic performance of investor-owned and not-for-profit hospitals, New England Journal of Medicine 314:89-96; Lewin and Associates (1981) Stud- ies in the Comparative Performance of Investor-Owned and Not-for-Profit Hospitals, Vol. 1: Industry Analysis (Washington, D.C.: Lewin and Associates); and Lewin, L., R. Derzon, and R. Margulies (1981) Investor-owneds and nonprofits diner in economic performance, Hos- pitals 55:52-58. 3Lewin and Associates (1976) Investor-Owned Hos- pit~s: An Examination of Performance (Chicago: Heals Services Foundation). 4This measure is also more restrictive than those based solely on distance between hospitals, e.g., no other hospitals in a 15-mile radius. Using the Medicare designation, none of the IO hospitals in our sample are sole community providers. However, using other def- initions, HCA, among other systems, may have a large number of sole community provider hospitals. sIn this Chapter, "significant" indicates p <.05. Here, significant differences are measured against not-for-profit freestanding hospitals, which constitute the majority of the hospitals in the universe Mom which our samples were drawn. 6This inference is supported by our earlier case study research. See Lewin and Associates, Inc. (1981) Studies in the Comparative Performance of Investor-Owned and Not-For-Proft Hospitals. Volume III: Two Case Studies of Competition Between Hospitals (Washing- ton, D.C.: Lewin and Associates).

Next: An Analysis Based on California Data, 1978-1982 »
For-Profit Enterprise in Health Care Get This Book
×
Buy Paperback | $140.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

"[This book is] the most authoritative assessment of the advantages and disadvantages of recent trends toward the commercialization of health care," says Robert Pear of The New York Times. This major study by the Institute of Medicine examines virtually all aspects of for-profit health care in the United States, including the quality and availability of health care, the cost of medical care, access to financial capital, implications for education and research, and the fiduciary role of the physician. In addition to the report, the book contains 15 papers by experts in the field of for-profit health care covering a broad range of topics—from trends in the growth of major investor-owned hospital companies to the ethical issues in for-profit health care. "The report makes a lasting contribution to the health policy literature." —Journal of Health Politics, Policy and Law.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!