Journal of Health Care Finance

Journal of Health Care Finance (USPS 048-090) Indexing: JHCF is indexed in Academic (ISSN 1078-6767) is published quarterly by Aspen Search/CD-ROM, the Business Periodicals Index, the Publishers, 76 Ninth Avenue, New York, NY 10011. Cumulative Index to Nursing & Allied Health Copyright © 2012 CCH Incorporated. All rights reserved. Literature (CINAHL), EMBASE, Health Source, Index www.aspenpublishers.com. Reproduction in whole Medicus, MEDLINE, MEDLARS, the UP-TO-DATE or in part without permission is strictly prohibited. Library/Health Services Management and Wilson Postmaster: Send address changes to Subscription Dept. Business Abstracts®, Wilson Business Abstracts Full IP. P.O. Box 3000, Denville, NJ 07834. Text® by The H.W. Wilson Company. Currently Subscription rate is $369 (plus postage and handling) available on CD-ROM and online via the WilsonWeb. per year in the United States and Canada (four issues), For more information, please visit http://www. payable in advance. The two-year subscription rate is hwwilson.com. $627, the three-year subscription rate is $886, and the “This publication is designed to provide accurate and cost of one issue is $111. Subscribers may specify any authoritative information in regard to the Subject Matter issue to begin the subscription. Subscribers in the covered. It is sold with the understanding that the United States and Canada: Address inquiries to publisher is not engaged in rendering legal, accounting, Fulﬁ llment, Aspen Publishers, 7201 McKinney Circle, or other professional service. If legal advice or other Frederick, MD 21704, or call 1-800-234-1660. To place expert assistance is required, the services of a an order, call 1-800-638-8437. Subscribers in all competent professional person should be sought.” countries other than the United States, Canada, and (From a Declaration of Principles jointly adopted by a Japan: Address inquiries to Aspen Publishers, c/o Swets Committee of the American Bar Association and a & Zeitlinger, P.O. Box 825, 2160 SZ Lisse, The Committee of Publishers and Associations.) Netherlands, telephone 31 252 435111, fax 31 252 Issue: Vol. 39, No. 1, 9900610022 415888. ISSN: 1078-6767 Permission requests: For information on how to Printed in the United States of America obtain permission to reproduce content, please go to the Aspen Publishers Web site at www.aspenpublishers. The paper used in this publication meets the com/permissions. requirements of the American National Standard for Purchasing reprints: For customized article reprints, Information Sciences—Permanence of Paper for Printed please contact Wright’s Media at 1-877-652-5295 or Library Materials, ANSI Z39.48-1992, effective with go to the Wright’s Media Web site at www.wrights Volume 13, Issue 3. media.com.

For more information resources, directories, articles, and a searchable version of Aspen Publishers’ full catalog, visit us at: www.aspenpublishers.com

1 2 3 4 5 Editorial Board

Editors Editors Emeritus

James J. Unland, MBA Judith J. Baker, PhD, CPA President Partner The Health Capital Group Resource Group, Ltd. Chicago, IL Dallas, TX

Paul Gibson, Publisher William O. Cleverley, PhD Joanne Mitchell-George, Senior Managing Editor Professor Elizabeth Venturo, Managing Editor Ohio State University Dom Cervi, Marketing Director Columbus, OH

Editorial Board Dana A. Forgione, PhD, CPA, CMA, CFE, Janey Elizabeth Simpkin, President, The Lowell Group, S. Briscoe Endowed Chair in the Business Inc., Chicago, IL of Health, and Professor of Accounting, Elaine Scheye, President, The Scheye Group, Ltd., College of Business, University of Texas at Chicago, IL San Antonio,TX Pamela C. Smith, PhD, Associate Professor, Ellen F. Hoye, MS, Principal, Hoye Consulting Ser- Department of Accounting, The University of vices, Elmhurst, IL Texas at San Antonio, San Antonio, TX Daniel R. Longo, ScD, Professor and Director Jonathan P. Tomes, JD, Partner, Tomes & Dvorak, of Research, ACORN Network Co-Director, Overland Park, KS Department of Family Medicine, Virginia Mustafa Z. Younis, Professor of Health Economics & Commonwealth University, Richmond, VA Finance, Jackson State University, School of Kevin T. Ponton, President, SprainBrook Group, Health Sciences, Department of Health Policy & Hawthorne, NY Management, Jackson, MS JHCF 39:1, Fall 2012 Contents

1 A Comparison of the Capital Structures of Nonprofi t and Proprietary Health Care Organizations John Trussel 12 A Comparative Analysis of the CVP Structure of Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals Li-Lin (Sunny) Liu, Dana A. Forgione, and Mustafa Z. Younis 39 The Economics of Health Care Quality and Medical Errors Charles Andel, Stephen L. Davidow, Mark Hollander, and David A. Moreno 51 Complicated Billing Requirements Challenge Physical Therapy Industry, Creating Ineffi ciencies and Confusion Annette R. Ciavarella 79 Americans with Disability Act: Financial Aspects of Reasonable Accommodations and Undue Hardship Sandra K. Collins and Eric P. Matthews 87 A Real Options Approach to Clinical Faculty Salary Structure Marc J. Kahn and Hugh W. Long From the Editor—About This Issue

Once again, this issue of the Journal of —James J. Unland, MBA Health Care Finance is illustrative of the The Health Capital Group breadth of topics we cover. 244 South Randall Road, Ste 123 We are always interested in new article ideas Elgin, Illinois 60123 that directly or indirectly relate to health care (800) 423-5157 ﬁ nance. To submit ideas or articles, please send healthﬁ [email protected] an email to: [email protected].

John Trussel

The relative amount of debt used by an organization is an important determination of the organization’s likelihood of fi nancial problems and its cost of capital. This study addresses whether or not there are any differences between proprietary and nonprofi t health care organizations in terms of capital structure. Controlling for profi tability, risk, growth, and size, analysis of covariance is used to determine whether or not proprietary and nonprofi t health care organizations use the same amount of leverage in their capital structures. The results indicate that there is no difference in the amount of leverage between the two institutional types. Although nonprofi t and proprietary organizations have unique fi nancing mechanisms, these differences do not impact the relative amount of debt and equity in their capital structures. Key words: institutional convergence, capital structure, leverage, nonprofi t fi nancial management, health care fi nancial management.

apital structure is the relative amount directly and empirically compared the capital of debt that an organization uses to structures of these two types of institutions. Cfi nance its investment projects and This study focuses on the health care industry, programs. The relative amount of debt used by in order to eliminate the effects of industry on an organization is an important determination capital structure. I use a methodology similar of the organization’s likelihood of fi nancial to Kester,4 who compared the capital structure problems and its cost of capital.1 Therefore, of Japanese and US manufacturing corpora- capital structure is a critical issue in the fi nan- tions. Using a sample of 163 US proprietary cial management of both proprietary and non- health care organizations, this study analyzes profi t organizations. Although nonprofi t and the capital structures and compares them proprietary organizations have unique legal to a sample of 163 US nonprofi t health care characteristics that potentially affect fi nanc- organizations. Specifi cally, I test the hypoth- ing, these differences may not impact the rela- esis of institutional convergence on capital tive amount of debt and equity in their capital structure—that there is no difference in the structures. This study addresses whether or capital structures of proprietary and nonprofi t not nonprofi t health care organizations have organizations. capital structures similar to proprietary health care organizations. The results will be of interest to creditors and other stakeholders when John Trussel, PhD, CPA, is an Associate Profes- sor of Accounting at the University of West Florida determining the organization’s cost of capital in Pensacola, Florida. His research interests include and its likelihood of fi nancial problems. accounting and fi nancial management of nonprofi t, Although several studies have addressed governmental, and proprietary organizations. He can capital structure in publicly traded corpo- be reached at [email protected]. 2 rations and a few have addressed the issue J Health Care Finance 2012; 39(1):1–11 in nonprofi t organizations,3 no studies have Copyright © 2012 CCH Incorporated 1 2 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

The results of this study indicate that DiMaggio and Anheier10 support this view when capital structure is proxied by leverage and also suggest that organizations of vari- and is measured as the book value of debt- ous forms (proprietary, nonprofi t, and pub- to-assets, there is no signifi cant difference lic) converge, according to industry, due between proprietary and nonprofi t organi- to these isomorphic and other competitive zations beyond that which the profi tability, pressures. As an example, Gray11 conjec- risk, growth, and size of the organizations tures that competition makes proprietary can explain. health care providers more socially respon- The following section provides a brief sible and nonprofi t health care providers overview of the unique legal characteris- more effi cient than they would otherwise be. tics that potentially affect fi nancing and the Empirical evidence supports Gray’s hypoth- capital structure models. The next section esis in the development of the hospital indus- discusses the hypotheses and methodology try.12 Since institutions seem to converge for testing. That section is followed with a according to industry, I focus on one broadly discussion of the empirical results, and the defi ned industry, health care, in testing the fi nal section summarizes this article. hypothesis of institutional convergence in the area of capital structure. Nonprofi t Organizations and Brody focuses on the principal-agent Capital Structure Models problem and concludes that nonprofi t and for-profi t fi rms “bear more resemblance to Whether or not the capital structures of each other than their organizational differ- nonprofi t and proprietary organizations ences suggest.”13 Concentrating on economic are similar is an empirical issue; however, forces and agency costs, she argues that the previous research addresses this issue con- two institutions are similar due to internal ceptually. Some researchers5 suggest that and external constraints, such as resource there is reason to believe that nonprofi t and dependency, institutional isomorphism, and proprietary organizations behave similarly organizational slack. from an economic standpoint, while oth- Hansmann14 suggests that the nondistri- ers6 argue that the legal differences make the bution constraint solves the principal-agent economics diverge. Research in the area of problem in nonprofi t organizations, since institutional convergence between nonprofi t this promise not to distribute profi ts acts to and proprietary organizations has become make the nonprofi t organization more trust- increasingly popular since the pioneering worthy. Monitoring costs are eliminated work of DiMaggio and Powell,7 who discuss as the nonprofi t delivers “trust goods.” In the isomorphism of organizations in general, contrast to Hansmann,15 Brody16 asserts and Brody,8 who focuses on the economic that the nondistribution constraint theory of convergence of nonprofi t and proprietary nonprofi t organizations cannot effectively organizations in particular. distinguish between a nonprofi t organiza- DiMaggio and Powell9 argue that organi- tion and a proprietary organization for two zational change occurs as a result of processes reasons. First, there is information asymme- that make organizations homogeneous with- try causing external agency costs. The prin- out necessarily making them more effi cient. cipals of a nonprofi t organization (donors, Capital Structures: Nonprofi t and Proprietary Health Care Organizations 3

customers, government agencies, etc.) may systems. Third, nonprofi t organizations can not be able to judge the quality of the ser- use the market value (net present value) rule vices provided by the agents (management), with discount rates determined by outside and the nonprofi t organization may not be capital markets when evaluating investment able to judge the quality of its services either. opportunities. Second, there is information asymmetry Fama and Jensen do not explicitly address caused by internal agency costs between the the issue of capital structure in nonprofi t management (principals) and the employ- organizations, however. To fi ll this void, ees (agents) of the nonprofi t organization. Wedig, Hassan, and Morrisey19 developed Brody also claims that the nondistribution a theory of capital structure for nonprofi t constraint does not explain how the public organizations. Their model explains how a should choose between competing nonprofi t nonprofi t organization, despite its exemp- organizations. tion from corporate tax, can indirectly ben- Bowman17 notes that there are four distinct efi t from the tax shields in the municipal legal characteristics relative to fi nancing bond market. Nonprofi t organizations have behavior that most nonprofi t organizations a fi nancial incentive to issue tax-exempt have when compared with proprietary bonds, since some of the personal tax abate- organizations. First, nonprofi t organizations ment is refl ected in the yields of municipal cannot issue stock nor distribute residual bonds. They argue that a nonprofi t organiza- claims. Second, they have the ability to raise tion can earn an indirect arbitrage by issuing funds through donations, and the donors low-yield, tax-exempt debt in lieu of internal can restrict their donated assets. Third, they fi nancing with existing cash fl ow. Opposing are not subject to involuntary bankruptcy. this incentive to issue debt, they claim, are the Fourth, many can sell bonds at tax-exempt factors, such as agency and bankruptcy costs, rates. He also notes that many hold signifi - that would lead to a reduction in the use of cant endowment assets, which are typically debt. Thus, similar to proprietary fi rms, a less risky than other assets. nonprofi t organization has an internal opti- Fama and Jensen18 provide the founda- mum mix of debt and equity, even though the tion for fi nancial management in nonprofi t choice of issuing stock doesn’t exist. From the organizations. They discuss the relationships nonprofi t organization’s point of view, equity among agency, contracting, and ownership is the retention of earnings for future program in (among others) nonprofi t organizations purposes. Empirical tests on a sample of non- and reach three general conclusions. First, profi t hospitals support their hypotheses. they conclude that the absence of residual The balance between opposing factors claims avoids certain agency problems and for issuing debt (i.e., the tax benefi ts versus explains the dominance of nonprofi t organi- agency, bankruptcy, and other such costs) zations in donor-fi nanced activities. How- espoused by Wedig et al.,20 is the so-called ever, residual risk still exists and is borne by static trade-off model of capital structure.21 consumers and factors used to produce out- This model stands in contrast to the pecking- puts. Second, like proprietary organizations, order model, which was advanced by Myers nonprofi t organizations require separation of and Majluf22 and Myers23 building upon ownership and control in designing decision Donaldson.24 Under this model, fi rms prefer 4 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

retained earnings (available liquid assets) within an industry are similar. To address this to debt and debt to equity when fi nancing issue empirically, I test the null hypothesis investments. Using a sample of 200 nonprofi t that there is no difference in the capital struc- hospitals, Bacon25 fi nds support for the peck- tures of nonprofi t and proprietary health care ing order model for nonprofi t institutions. organizations. The empirical model uses Bowman26 uses a sample of 1,393 non- cross-sectional data for 163 proprietary profi t organizations from four sectors, and 163 nonprofi t health care organizations including hospitals, and has mixed results from 1995 and is replicated using data from for tests of the pecking-order and static 2005. The year 1995 was chosen, since the trade-off models. However, neither he nor accounting rules change signifi cantly for the other researchers in nonprofi t capital nonprofi t organizations in the United States structure (Wedig et al.27 and Bacon28) com- after 1995. These data are drawn from two pare nonprofi t organizations with propri- primary sources. Data for proprietary health etary organizations in their empirical tests. care organizations are obtained from the This article extends the research on capital Standard and Poor’s Compustat database. structure to compare, directly and empiri- I included all 163 “health services” organi- cally, the capital structures of nonprofi t and zations (Standard Industrial Code (SIC) of proprietary organizations. 8000 to 8099) that had all of the necessary Kester29 is one of the few studies that data to apply the empirical model. I also directly compares the capital structures of selected a sample of nonprofi t health care two different populations, although both organizations from the Statistics on Income are proprietary. Using cross-sectional data database of the National Center on Charita- for 344 Japanese and 452 US manufactur- ble Statistics (NCCS). I included a random ing companies in 27 different industries, he sample of 163 health care organizations that tests whether or not the Japanese and the had all of the necessary data. The nonprofi t US manufacturing fi rms have similar capital health care organizations had codes desig- structures. Controlling for profi tability, risk, nated by the National Taxonomy of Exempt growth, size, and industry in his empirical Entities (NTEE) that were matched to the tests, he fi nds that there may or may not be a SIC health services classifi cations 8000 to difference in capital structure, depending on 8099. The NTEE is a comprehensive classi- how leverage is measured. However, when fi cation system developed by the NCCS for there was a difference, it was concentrated in tax-exempt organizations. Only health care certain industries. This article uses a method- organizations are included to control for the ology similar to Kester’s with a few refi ne- effects of a specifi c industry on capital struc- ments, as discussed below. ture (see Figure 1). Following Kester,30 each organization’s Data and Methodology leverage is used as a proxy for its capital structure (the dependent variable). To reduce Much of the literature discussed above, the amount of variance in the error terms, including the theory of institutional conver- several covariates are utilized. Again, fol- gence, suggests that the capital structures lowing Kester,31 these additional independ- of nonprofi t and proprietary organizations ent variables are profi tability, risk, growth, Capital Structures: Nonprofi t and Proprietary Health Care Organizations 5

Figure 1. Classiﬁ cations of Health Care Organizationsa

SIC Classifi cation SIC NTEE Offices and Clinics of Doctors of Medicine Surgical and 8011 E31 Emergency Centers, HMOs Offices and Clinics of Dentists 8021 E31 Offices and Clinics of Doctors of Osteopathy Offices of Doctors 8031 E31 of Osteopathy Offices and Clinics of Chiropractors 8041 E31 Offices and Clinics of Optometrists 8042 E31 Offices and Clinics of Podiatrists 8043 E31 Offices and Clinics of Health Practitioners, Mental Health 8049 E50 Practitioners Skilled Nursing Care Facilities Continuing Care Retirement 8051 E90, E91 Communities Intermediate Care Facilities Continuing Care Retirement 8052 E90 Communities Nursing and Personal Care Facilities, NEC Continuing 8059 E90 Care Retirement Communities General Medical and Surgical Hospitals 8062 E20, E21, E22 Psychiatric Hospitals 8063 E20 Specialty Hospitals 8069 E20, E24 Medical Laboratories Diagnostic Imaging Centers 8071 E31 Dental Laboratories 8072 E31 Home Health Care Services 8082 E92 Kidney Dialysis Centers 8092 E31 Specialty Outpatient Facilities, NEC Family Planning Centers 8093 E30, E32, E40, E42 Health and Allied Services, NEC Blood and Organ Banks 8099 E40, E60, E61, E65, E70, E86, E99 a Figure 1 represents the various sub-sectors within the SIC classifi cation of “health services” organizations (SIC 8000–8099). The SIC classifi cations are matched with the National Taxonomy of Exempt Organiza- tions (NTEE) codes for nonprofi t organizations (National Center of Charitable Statistics 2000). The samples for this study were drawn from these sub-sectors.

␤ and size. A dummy independent variable 8 GROWTH*TYPE + ␤ ⑀ will be used to represent the institutional 9 SIZE*TYPE + (1) type. The full regression model (with observation subscripts removed) is: Leverage (LEV) is measured as the book value of total liabilities divided by total assets. ␤ ␤ ␤ LEV = 0 + 1 ROA + 2 RISK + Proﬁ tability (ROA) is deﬁ ned as the excess of ␤ ␤ ␤ 3 GROWTH + 4 SIZE + 5 TYPE + revenues over expenses scaled by total assets, ␤ ␤ 6 ROA*TYPE + 7 RISK*TYPE + and is calculated as the average from the 6 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

previous fi ve years. The standard deviation of The second step addresses the equality the ROA over the previous fi ve years is used as of the intercept terms of the two regression a proxy of the volatility (RISK) of the return on lines. In testing for the equality of intercepts, assets. GROWTH was measured as the com- the hypotheses are: pound average annual rate of growth in revenues over the last fi ve years. SIZE, defi ned as ␤ H0: 5 = 0 the natural log of the previous year’s revenues, ␤ ≠ is included as a proxy for other relevant vari- HA: 5 0 ables, such as information asymmetries and age.32 Finally, TYPE is a dichotomous dummy If the parameter is insignifi cant, then the null variable defi ned as one, if the organization is cannot be rejected. If the slopes are parallel nonprofi t and zero, otherwise. The interaction and the intercepts are the same, then there β β is no evidence to reject the hypothesis that parameters, 6– 9, were included in the full regression model to express the potential rela- the regression lines are coincident. If this tionship between the covariates ROA, RISK, is the case, then there is no basis for rejecting GROWTH, and SIZE and the experimental the hypothesis that nonprofi t and proprietary (dummy) variable TYPE. Bowman33 suggests health care organizations have similar capi- that the relative amount of endowment assets, tal structures. The results are included in the defi ned as investments, be included as a pre- next section. dictor variable. Inclusion of investments as a percent of total assets does not change the Results results signifi cantly. He also suggests remov- ing investments from the ROA and other cal- Figure 2, panel A, includes the descriptive culations. This also does not change the tenor and univariate results of the testing using the of the results. 1995 data. The results using 2005 data are To test the hypothesis that there is no dif- similar and thus are not shown. Mean ROA ference in the capital structures of the two for proprietary organizations is signifi cantly institutions, I use analysis of covariance lower than nonprofi t organizations, and they and proceed in two steps. The fi rst step are also are signifi cantly more risky (RISK) addresses the equality of the slopes of the than nonprofi ts. The proprietary organiza- two regression lines (i.e., nonprofi t versus tions grew more than the nonprofi t organi- proprietary). I test for parallelism using zations, but the difference in GROWTH interaction terms for the dummy variable was not signifi cant. This result is interesting (TYPE) on each of the covariates. In testing since I assumed that the ability to issue stock for parallelism, the hypotheses are: would give a growth advantage to the proprietary organizations. However, the nonprofi t ␤ ␤ ␤ ␤ H0: 6 = 7 = 8 = 9 = 0 organizations were signifi cantly smaller in SIZE than the proprietary ones, which may

HA: at least one of the parameters is not zero. explain why GROWTH was insignifi cant. If the nonprofi t organizations were smaller over If the parameters of the interaction terms are the long run, then the SIZE difference may insignifi cant, then the null cannot be rejected. explain why GROWTH was insignifi cant. Capital Structures: Nonprofi t and Proprietary Health Care Organizations 7

Figure 2. Summary Statisticsa

ROA RISK GROWTH SIZE Panel A: Descriptive statistics and univariate tests Proprietary Mean Ϫ0.166 0.229 0.261 18.49 (Std. Dev.) (0.341) (1.328) (0.399) (2.241)

Nonprofi t Mean 0.101 0.055 0.206 16.95 (Std. Dev.) (0.023) (0.030) (0.867) (1.550)

t-statistic Ϫ9.966** 2.150* 0.739 7.947**

Panel B: Pearson correlation coefficients ROA RISK GROWTH

RISK Ϫ0.230**

GROWTH Ϫ0.025 0.007

SIZE 0.138* Ϫ0.110* 0.033

a The sample consists of 163 health care organizations that are proprietary and 163 that are nonprofi t. The independent variables are as follows: ROA is the ratio of total revenues less total expenses to total assets; RISK is the volatility of ROA; GROWTH is the rate of growth of revenues over the last fi ve years; SIZE is the natural log of revenues; and TYPE is 1 if the organization is nonprofi t and zero otherwise. * Signifi cance at the 0.05 level (two-tailed). ** Signifi cance at the 0.01 level (two-tailed).

The Pearson correlation coeffi cients are the interaction terms were insignifi cant, I included in Figure 2, panel B. The highest focused on the reduced regression model. correlation is between ROA and RISK Panel A of Figure 4 includes the results of (Ϫ0.230). The correlations appear within the reduced regression model, without the reasonable bounds and should not present interaction parameters. Panel B of Figure 4 a serious threat to the estimation of the includes the parameter estimates. Overall the regression model. model is signifi cant at the 0.05 level. ROA The results of the analysis of covariance and RISK are signifi cant at the 0.05 level. are included in Figures 3 and 4. Figure 3 GROWTH and SIZE are not signifi cant at includes the full regression model (1), with the 0.05 level. Although not a focus of this the interaction parameters. The interaction study, the negative sign on ROA supports terms were included to test for parallelism. the pecking-order model of capital structure. Since none of the interaction parameters is More importantly, TYPE is insignifi cant at signifi cant at the 0.05 level, the null hypoth- the 0.05 level, indicating that the null hypothesis of equal slopes is not rejected, and there esis of equal intercepts cannot be rejected. is no evidence to reject the hypothesis that Since the regression lines for proprietary the regression lines for proprietary and and nonprofi t health care organizations are nonprofi t organizations are parallel. Since parallel and have the same intercept, there is 8 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 3. Test for Parallelism: Full Modela

Variableb Df Mean Square F p-valuec Corrected Model 9 5.884 11.596 .001 CONSTANT 1 0.077 0.153 .695 ROA 1 0.747 1.484 .224 RISK 1 0.010 0.020 .888 GROWTH 1 0.008 0.017 .897 SIZE 1 0.004 0.009 .925 TYPE 1 1.029 2.046 .154 ROA*TYPEd 1 0.156 0.309 .578 RISK*TYPEd 1 0.047 0.095 .759 GROWTH*TYPEd 1 0.050 0.099 .753 SIZE*TYPEd 1 0.524 1.041 .308 Adjusted R2 = 0.229

a The full ANACOVA model is estimated using a sample of 163 proprietary health care organizations and 163 nonprofi t health care organizations. b The dependent variable is leverage, the ratio of debt-to-total assets. The independent variables are: ROA is the ratio of total revenues less total expenses to total assets; RISK is the volatility of ROA; GROWTH is the rate of growth of revenues over the last fi ve years; SIZE is the natural log of revenues; and TYPE is 1 if the organization is nonprofi t and zero otherwise. c All p-values on coefficients are based on one-tailed t-tests, with the exception of the model, the intercept, TYPE, and the interaction terms, which are based on two-tailed t-tests. d The interaction terms are included to test for parallelism. Since the interaction terms are insignifi cant, the null hypothesis of equal slopes is not rejected.

no evidence for rejecting the hypothesis that Summary and Conclusions they are coincident. Thus, there is no evidence that the capital structures (as measured This study addresses whether or not there by the debt-to-assets ratio) of proprietary and are any differences in the capital structures nonprofi t health care organizations are dif- of proprietary and nonprofi t health care ferent, supporting the theory of institutional organizations. The theory of institutional convergence on capital structure. Although convergence holds that entities with different nonprofi t and proprietary health care organi- institutional forms have similar characteris- zations have unique fi nancing mechanisms, tics. That is, even though nonprofi t organiza- these differences do not impact the relative tions have unique legal characteristics when amount of debt and equity in their capital compared to proprietary organizations, these structures. Using data from 2005 (results do not cause differences in capital structures. not shown), the results are similar and thus Although there have been several studies on robust as to the different time periods. the capital structures of both proprietary Capital Structures: Nonprofi t and Proprietary Health Care Organizations 9

Figure 4. Test for Equality of Intercepts: Reduced Modela

Panel A: Reduced ANACOVA Model

Variableb Df Mean Square F p-valuec Corrected Model 5 10.347 20.671 .001 CONSTANT 1 0.001 0.002 .963 ROA 1 16.127 32.221 .001 RISK 1 18.972 37.903 .001 GROWTH 1 0.193 0.386 .535 SIZE 1 0.961 1.919 .167 TYPEd 1 0.016 0.032 .857 Adjusted R2 = 0.233

Panel B: Parameter Estimates

Variable Coefficient Standard Error t p-valuec CONSTANT Ϫ0.010 0.388 Ϫ0.026 .979 ROA Ϫ1.002 0.177 Ϫ5.676 .001 RISK 0.264 0.043 6.157 .001 GROWTH Ϫ0.0363 0.058 Ϫ0.622 .535 SIZE 0.0315 0.023 1.385 .167 TYPE Ϫ1.830 0.102 Ϫ0.180 .857

a The reduced model is estimated using a sample of 163 proprietary health care organizations and 163 nonprofi t health care organizations. b The dependent variable is leverage, the ratio of debt-to-total assets. The independent variables are: ROA is the ratio of total revenues less total expenses to total assets; RISK is the volatility of ROA; GROWTH is the rate of growth of revenues over the last fi ve years; SIZE is the natural log of revenues; and TYPE is one if the organization is nonprofi t and zero otherwise. The interaction terms are eliminated, since they were insignifi cant (see Figure 3). c All p-values on coefficients are based on one-tailed t-tests, with the exception of the model, the intercept, and TYPE, which are based on two-tailed t-tests. d Since the experimental variable TYPE is insignifi cant, the null hypothesis of equality of intercepts is not rejected. and nonprofi t organizations, no studies have in the amount of leverage between the two directly and empirically compared the two. institutional types, supporting the institu- Controlling for profi tability, risk, growth, tional convergence on capital structure. and size, I use analysis of covariance to The limitations of this study are the limited determine whether or not the proprietary and time period, the single industry focus, and the nonprofi t organizations use the same amount use of only organizations in the United States. of leverage (measured as the ratio of debt-to- Due to changes in accounting rules, this study total assets) in their capital structures. The focused on data from the mid-1990s. How- results indicate that there is no difference ever, data from the mid-2000s also support 10 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

the hypothesis. Also, to eliminate the impact comparing their structures to those of propri- of industry on capital structure, only one, etary organizations. The relative amount of albeit broad, industry was used for testing. debt used by an organization is an important Further research is needed to expand the determination of the organization’s likeli- time periods of the study and to apply the hood of fi nancial problems and its cost of tests to other industries and other countries. capital.34 Therefore, capital structure is a This research contributes to the existing lit- critical issue in the fi nancial management of erature on capital structure of nonprofi ts by both proprietary and nonprofi t organizations.

REFERENCES

1. Jensen, M, Meckling, W, “Theory of the Firm: 4. Kester, C, “Capital and Ownership Structure: Managerial Behavior, Agency Costs and A Comparison of United States and Japanese Ownership Structure,” Journal of Financial Manufacturing Corporations,” Financial Man- Economics, 3: 305–360 (1976). agement, Spring: 5–16 (1986). 2. E.g., Toy, N, Stonehill, A, Remmers, L, Wright, 5. E.g., Brody, E, “Agents Without Principles: The R, Beekhuisen, T, “A Comparative Interna- Economics Convergence of the Nonprofi t and tional Study of Growth, Profi tability, and Risk For-Profi t Organizational Forms,” New York as Determinants of Corporate Debt Ratios in Law School Law Review, 40: 457–535 (1996). the Manufacturing Sector,” Journal of Finan- 6. E.g., Bowman, supra, n.3. cial and Quantitative Analysis, Nov: 875–886 7. DiMaggio, P, Powell, W, “The Iron Cage (1974); Baskin, J, “An Empirical Investigation Revisited: Institutional Isomorphism and Col- of the Pecking Order Hypothesis,” Financial lective Rationality in Organizational Fields,” Management, Spring: 26–35 (1989); Kes- American Sociological Review, 82: 147–160 ter, C, “Capital and Ownership Structure: A (1983). Comparison of United States and Japanese 8. Brody, supra, n.5. Manufacturing Corporations,” Financial Man- 9. DiMaggio and Powell, supra, n.7. agement, Spring: 5–16 (1986). 10. DiMaggio, PJ, Anheier, HK, “The Sociology 3. E.g., Wedig, GJ, Sloan, F., Hassan, M, Mor- of Nonprofi t Organizations and Sectors,” risey, M, “Capital Structure, Ownership and Annual Review of Sociology, 16: 137–159 Capital Payment Policy: The Case of Hospi- (1990). tals,” Journal of Finance, 43: 21–40 (1988); 11. Gray, BH, Profi t, Corporate Change and Wedig, GJ, “Risk, Leverage, Donations, and Accountability in American Health, 20th Cen- Dividends-in-Kind: A Theory of Nonprofi t tury Fund (1990). Financial Behavior,” International Review of 12. Starr, P, “Medical Care and the Boundaries of Economics and Finance, 3: 257–278 (1994); Capitalist Organization,” unpublished manu- Wedig, GJ, Hassan, M, Morrisey, M, “Tax- script, Program on Non-Profi t Organizations, Exempt Debt and the Capital Structure of New Haven: Yale University (1980). Nonprofi t Organizations: An Application to 13. Brody, supra, n.5, p. 458. Hospitals,” Journal of Finance, 51: 1247– 14. Hansmann, H, “The Role of the Nonprofi t 1283 (1996); Bowman, W, “The Uniqueness Enterprise,” Yale Law Journal, 89: 835–901 of Nonprofi t Finance and the Decision to Bor- (1980). row,” Nonprofi t Management and Leadership, 15. Id. 12(3): 293–311 (2002). 16. Brody, supra, n.5. Capital Structures: Nonprofi t and Proprietary Health Care Organizations 11

17. Bowman, supra, n.3. Information Investors Do Not Have,” Journal 18. Fama, E, Jensen, M, “Separation of Owner- of Financial Economics, 13: 187–221 (1984). ship and Control,” Journal of Law and Eco- 23. Myers, S, “The Capital Structure Puzzle,” Jour- nomics, 26: 301–325 (1983); Fama, E, Jensen, nal of Finance, 39: 575–592 (1984). M, “Agency Problems and Residual Claims,” 24. Donaldson, G, Corporate Debt Capacity: Journal of Law and Economics, 26: 327–349 A Study of Corporate Debt Policy and the (1983b); Fama, E, Jensen, M, “Organizational Determination of Corporate Debt Capacity, Forms and Investment Decisions,” Journal of Boston: Harvard Business School (1961). Financial Economics, 14: 101–119 (1985). 25. Bacon, P, “Do Capital Structure Theories Apply 19. Wedig, GJ, Hassan, M, Morrisey, M, “Tax- to Nonprofi t Hospitals?” Journal of the Mid- Exempt Debt and the Capital Structure of west Finance Association, 21: 86–90 (1992). Nonprofi t Organizations: An Application to 26. Bowman, supra, n.3. Hospitals,” Journal of Finance, 51: 1247– 27. Wedig, et al. (1988), supra, n.3. 1283 (1996). 28. Bacon, supra, n.25. 20. Id. 29. Kester, supra, n.4. 21. Taggart, R, “A Model of Corporate Financing 30. Id. Decisions,” Journal of Finance, 32: 1467– 31. Id. 1484 (1977). 32. Myers and Majluf, supra, n.22. 22. Myers, S, Majluf, N, “Corporate Financing 33. Bowman, supra, n.3. and Investing Decisions When Firms Have 34. Jensen and Meckling, supra, n.1. A Comparative Analysis of the CVP Structure of Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals

Li-Lin (Sunny) Liu, Dana A. Forgione, and Mustafa Z. Younis

Due to the market turbulence facing the hospital industry, the fi nancial viability of teaching hospitals has been severely threatened. Their missions of education, research, and patient care even strengthen this crisis. Therefore, the objective of this study is to conduct a comparative analysis of the cost, volume, and profi t (CVP) structure between large nonprofi t urban teaching hospitals and small for-profi t rural/ suburban non-teaching hospitals. The following two hypotheses were developed: (1) large nonprofi t urban teaching hospitals tend to have higher fi xed cost, lower variable cost, lower total revenue adjusted by case mix index (CMI), and lower return on total assets (ROA); and (2) small for-profi t rural/suburban non-teaching hospitals tend to have lower fi xed cost, higher variable cost, higher total revenue adjusted by CMI, and higher ROA. Using 117 teaching hospitals and 102 non-teaching hospitals selected from the Medicare Cost Report database in 2005, the results from multiple regression indicated that large nonprofi t teaching hospitals located in urban areas are more likely to have higher fi xed cost and lower variable cost. While such cost structure doesn’t necessarily affect their total revenue adjusted by CMI, it does lead to a lower return on hospitals’ total assets. The results support our hypotheses in terms of fi xed cost percentage, variable cost percentage, and ROA, but not total revenue adjusted by CMI. The results suggest that cost structure is signifi cantly associated with hospitals’ performance. Also, as teaching hospitals’ portfolios of services and programs increase (e.g., provision of uncompensated care to Medicare and Medicaid patients and doing research), it becomes strategically necessary and critical to manage the allocation of resources or investments into the fi xed capital that supports the business. Key words: CVP structure, teaching hospitals, non-teaching hospitals.

I. Introduction viability of all hospitals, the teaching hospitals are especially affected due to their lack of Teaching hospitals are facing a highly exposure to a competitive industry and their uncertain and unpredictable future. The once unusually large and complex organizational placid and stable industry that allowed teach- and capital infrastructures. Additionally, ing hospitals to practice academic medicine the existence of several business missions in a fi nancially secure environment has been replaced by an extremely turbulent and Li-Lin (Sunny) Liu, PhD, CPA, is Assistant Profes- 1 competitive marketplace. A comprehensive sor in the Department of Accounting and Finance, review of prior studies regarding Academic College of Business Administration and Public Pol- Health Centers (AHCs), Academic Medi- icy, California State University, Dominguez Hills. cal Centers (AMCs), and teaching hospitals Dana A. Forgione, PhD, CPA, CMA, CFE, is the has indicated that changes in public policy, Janey S. Briscoe Endowed Chair in the Business of increases in uncompensated care, growth in Health and Professor of Accounting in the College of Business at the University of Texas at San Antonio, managed care, increases in input prices, par- San Antonio, Texas. adox of higher intensity, and drop of oper- Mustafa Z. Younis, is Professor of Health Econom- ating margins are all key challenges facing ics and Finance, at Jackson Sate University, School these institutions and derive a turbulent and of Health Sciences, Jackson, Mississippi. He can be competitive market. reached at [email protected]. 2 As Langabeer indicated, although this J Health Care Finance 2012; 39(1):12–38 market turbulence threatens the fi nancial Copyright © 2012 CCH Incorporated 12 A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 13

(i.e., education, research, and patient care) literature regarding challenges facing AHCs, rather than the singular profi t-maximizing AMCs, and teaching hospitals, and the possi- mission of the industrial organization, poses ble determinants of performance differences even more critical concerns for the fi nan- between teaching and non-teaching hospi- cial viability of the teaching hospitals in tals. Section III provides research design, extremely turbulent markets.3 including propositions, sample selection, While the private sector has paid signifi - data source, and research method, and the cant attention to the links among the envi- fi nal section concludes with a summary of ronment, strategy, and fi nancial performance expected conclusion, limitations, and possi- for the industrial organization,4 very little ble future research. research has focused on the academic teaching hospitals. Consequently, the objective of II. Literature Review this study is to conduct a comparative analysis of the cost, volume, and profi t (CVP) struc- Challenges Facing AHCs, AMCs, and ture between large non-profi t urban teaching Teaching Hospitals hospitals and small for-profi t rural/suburban The Association of Academic Health non-teaching hospitals. Based on an exten- Centers5 defi nes an AHC as an allopathic or sive review of prior literature concerning osteopathic medical school with at least one teaching hospitals’ performance, we hence or more other health professional schools develop the following two propositions: and an associated teaching clinical enterprise.6 In other words, an AHC refers to a 1. Large nonprofi t urban teaching hospitals medical teaching institution affi liated with tend to have lower CMI adjusted vol- a primary hospital. These institutions have ume, higher fi xed cost, lower variable three goals: cost, and lower return on assets (ROA, profi t margin x asset turnover); and 2. Small for-profi t rural/suburban non- 1. To provide medical and health educa- teaching hospitals tend to have higher tion and training; CMI adjusted volume, lower fi xed 2. To conduct research and technological cost, higher variable cost, and higher innovations; and ROA. 3. To provide patient care, especially to populations who are unable to pay. By obtaining data primarily from the Medicare Cost Report, we plan to utilize The environment facing AHCs has the CVP technique to conduct the analy- been described as turbulent,7 endangered,8 sis. Hopefully, the fi ndings can be used to competitive,9 and opportunistic.10 Such a enhance the decision effectiveness of hospi- dynamic environment exists due to the fol- tal administrators in selecting their product lowing paradoxical view: on one hand, mix and designing their CVP structure as AHCs share distinctive competitive advan- well as fi nancial strategies appropriately. tages as clinical and research hubs in the The remainder of this article is organized health care sector although they confront as follows: Section II describes the related increased scrutiny from federal regulators 14 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

to ensure that their clinical services and 5. Paradox of Higher Intensity; and resident practice patterns are in compli- 6. Drop in Operating Margins. ance with federal statutes;11 and on the other hand, reimbursement systems that reward Changes in Public Payment Policy effi ciency have replaced traditional fee- The enactment of the Balanced Budget Act for-service plans and increased penetration (BBA) of 1997 has reduced both Medicare by managed care organizations have also and Medicaid payments to teaching hospitals reversed the fi nancial incentives that were and affected hospital profi tability14 and the previously available under the traditional average length of hospitalization.15 Medicare fee-for-service system.12 has been a key source of funding for graduate AMCs are generally defi ned to include the medical education (GME) which is the train- medical center, the teaching hospital, and the ing that interns or residents receive follow- faculty practice plan. Fox et al.13 examined ing medical school.16 From 1990 to 1997, the the factors infl uencing AMCs’ fi nancial per- average US hospital payment-to-cost ratio for formance. Their analysis showed that AMCs fee-for-service Medicare patients increased tend to be more expensive than their commu- from 89.2 percent to 103.6 percent. However, nity counterparts, that there are differences following the enactment of the BBA of 1997, in the mission and culture of AMCs and Medicare Prospective Payment System (PPS) managed care organizations, and that AMCs payments were reduced and the Medicare and managed care are incompatible on sev- hospital payment-to-cost ratio fell to 101.1 eral levels, including distribution of patient percent in 1999.17 In addition to the reduction revenues, attitudes toward patients, and the of Medicare hospital payments, the provi- role of primary care physicians. In other sions of the BBA of 1997 also reduced Med- words, AMCs need to act like multispecialty icaid hospital payments. Most of the change group practices, to adopt new methods for in Medicaid hospital payments has been controlling service use, and to encourage attributed to growth in the Medicaid Dispro- recognition of the costs of medical educa- portionate Share Hospital (DSH) program.18 tion in order to survive in this increasingly This program provides additional funding to turbulent environment. hospitals serving a relatively large number Due to the emerging dynamic environ- of indigent patients. From 1990 to 1998, the ment, all AHCs, AMCs, and teaching hos- Medicaid hospital payment-to-cost ratio for pitals need to identify the key challenges fee-for-service patients increased from 79.7 facing them so as to position themselves percent to a high of 97.9 percent. However, more favorably for future survival or suc- following the enactment of the BBA of 1997, cess. Based on an extensive review of prior the Medicaid hospital payment-to-cost ratio literature, we thus summarize the possible fell to 96.7 percent in 1999.19 challenges facing them as follows: What are the impacts on AHCs of these reductions in funding? Guo20 reported that 1. Changes in Public Payment Policy; the reductions in Medicare and Medicaid 2. Increases in Uncompensated Care; funding decreased AHCs’ revenue and 3. Growth in Managed Care; increased uncompensated care. Alexander 4. Increases in Input Prices; et al.21 indicated that reductions in these A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 15

funding sources adversely affect AHCs’ reduced fees, which is called “uncompen- mission of providing care to the poor and sated care.” The increases in uncompensated uninsured people. The costly care for care are associated with the growth in the these patients has contributed to the ris- number of Americans without health insur- ing uncompensated care of AHCs. Just as ance. The number of uninsured increased Reuter22 reported, AHCs have the highest from 35.6 million in 1990 to 42.6 million levels of uncompensated care, accounting in 1999.25 Due to the competitive forces in for 37 percent of all the uncompensated the health care market, it’s getting more dif- care in 1994. Reuter23 indicated that such fi cult for hospitals to shift the costs of caring funding reductions also affect AHCs’ train- for poor and uninsured patients onto paying ing and education mission. As he stated in patients. As Bellandi26 reported, the uncom- his study, programs in GME have more than pensated care represented about 6 percent doubled from 1970 to 1994. Since Medi- of hospital expenses in 1998. When uncom- care is the largest payer of GME, AHCs are pensated care reaches a substantial portion especially vulnerable to government legis- of a hospital’s business, it could become a lations that call for reductions in funding to fi nancial burden.27 Consequently, hospitals Medicare and Medicaid. that serve large numbers of indigent patients In addition, Linna et al.24 utilized the sto- need to be subsidized either with public chastic frontier cost function to estimate the funds or private charities in order to sur- teaching and research costs of Finnish hospi- vive in competitive markets.28 Federal and tals and demonstrated that: state governments currently are considering revisions in payment policies to support 1. The effi ciency adjustment had signifi - hospitals that disproportionately serve poor, cant impact on the marginal and aver- uninsured patients.29 age cost estimates of the teaching and The degrees to access the above-described research output; and subsidies could be various depending upon 2. The university teaching hospitals were the types of hospitals. For example, public underfunded with respect to both hospitals and teaching hospitals may be research and teaching output. particularly affected by changes in these payment policies because many of them Their results further suggested that the are located in urban centers and serve large average rate of teaching and research reim- numbers of low-income patients.30 Espe- bursement should be approximately 14.6 cially teaching hospitals, they must bal- percent of the total operating costs in uni- ance their patient care activities with their versity teaching hospitals. Their fi ndings research and education missions in order may provide some important implications to to position themselves favorably in the policymakers when there is a need to initiate market for survival. Moreover, the care of some policy changes in the future. uninsured patients in teaching hospitals may even be affected by growth of managed Increases in Uncompensated Care care during the 1990s. Weissman et al.31 Many hospitals provide health services examined the relation of teaching status to the uninsured for free or substantially and managed care to changes in market 16 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

share and market concentration and found negative impacts of the proliferation of the following: managed care. For example, Levey et al.33 indicated that managed care, while affect- • The market share of uninsured patients ing signifi cant cost savings, has at no small during the 1990s grew at AMC hospitals cost to America’s major teaching hospi- relative to other hospitals in their ser- tals and their social missions of teaching, vice areas; and these increases occurred research, and patient care. Also, as more primarily in areas with high levels of and more consumers enroll in managed managed care; and care plans, they cannot live by primary • In urban areas, the care of uninsured care alone; in other words, they will, at patients was twice as concentrated as one time or another, need the kinds of spe- that of all patients, and the ratio of con- cialized care that only major teaching hos- centration (uninsured patients vs. all pitals and the specialist physicians trained 34 patients) was greater in areas with high in such institutions can provide. Guo managed care levels than in areas with reported that the growth of managed care low managed care levels. has affected AHCs in several ways. First, managed care has initiated cost contain- ment strategies by seeking to contract with Growth in Managed Care health care organizations that offer lower Managed care here is defi ned as a system fees. As they are more likely to contract that either pays hospitals on a capitated basis with non-teaching hospitals that charge (a per member per month basis) or pays the less for patient care than AHCs, AHCs are hospitals based on a discounted price. This currently facing increasing price competi- managed care system includes two types tion. More specifi cally, while AHCs have of agreements: one between hospitals and always charged 30 to 40 percent more than health maintenance organizations (HMOs) non-teaching hospitals to offset their mis- and the other between hospitals and pre- sion of caring for the poor and uninsured,35 ferred provider organizations (PPOs). As they cannot afford to do so any more under Fox32 documented, about 30 percent of the managed care. Second, more for-profi t population is enrolled in HMOs and 34 per- health care organizations entering the mar- cent in PPOs, with only 14 percent remain- ket and integrating also threaten AHCs’ ing in traditional fee-for-service plans. survival. Although managed care may be credited In other words, the growth of managed with reducing hospital and health care cost care has brought drastic competition, thus infl ation, such fi nancial constraints are also leading to higher AHC costs and lower rev- blamed for jeopardizing other important enues. Higher costs result from the fact that hospital objectives, especially teaching hos- AHCs must serve patients with sicker and pitals’ primary missions in providing direct more complex needs which non-teaching patient care, training the US medical work- hospitals are not able to accept. Lower rev- force, and engaging in medical research. enues result from AHCs’ inability to compete Several studies dealing with AHCs or with non-teaching hospitals to secure man- teaching hospitals have pointed out the aged care contracts. A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 17

Increases in Input Prices Despite the widely held perception that Shortages of hospital workers, such as AMCs’ important mission and perceived nurses, radiology technicians, and pharma- quality advantages have been used to justify cists, have led to upward pressures on hospi- their higher costs and ensure their inclu- tal wages.36 This pressure can be even more sion in preferred networks of providers, the 39 severe for a labor-intensive industry such as Fisher et al. study suggests that: hospitals. As Rodgers et al.37 indicated, an estimated 38.8 percent of the increase in 1. The challenge facing AMCs is to learn hospital care spending between 1997 and how to improve both the quality and 2001 has been attributed to labor costs (e.g., effi ciency of care; and wages and benefi ts). 2. Academic institutions that demonstrate both high quality and lower long-term Paradox of Higher Intensity costs will be more likely to compete successfully in a health care market- 38 Fisher et al., focusing on patients whose place that is increasingly concerned initial hospitalization was at one of the 299 about longitudinal effi ciency. hospitals that are members of the Council of Teaching Hospitals (COTH), examined Drop in Operating Margins the content, quality, and outcomes of care According to a study commissioned by across AMCs that differ by up to 60 percent the New York City–based Commonwealth in the overall intensity of medical services Fund for the fi nancial performance of AHC delivered to patients with serious chronic ill- hospitals, 1994–2000, the main challenges nesses. Their fi ndings indicated that: facing AHCs are the declining operating margins. This report documents a drop of the 1. Although major US AMCs differ dra- operating margins of 1.4 percent and 2.6 per- matically in the overall intensity of ser- cent in 2000 for AHCs and major teaching vices they provide to similar patients, hospitals, respectively. Additionally, among the increased intensity doesn’t appear the several factors contributing to the declin- to be associated with higher quality of ing fi nancial performance of AHC hospitals, care or to result in better survival; and the primary two are: 2. Characteristics of the institutions themselves (or the physicians practic- 1. The decrease in private payer payment- ing in them) provide the major alter- to-cost ratio; and native explanation for the differences 2. The amount of resources that AHC in intensity. hospitals spent in caring for the poor. With respect to the second fi nding, one According to this report, since private pay- factor could be the size of the teaching pro- ers have been less willing than in the past grams, which is consistent with the fi ndings to support the high costs of AHC hospitals, of prior studies which have shown that as the the private payer payments as a percentage size of the teaching programs increases, so of costs have declined from 128.4 percent does the cost of care. in 1996 to 112.5 percent in 2000.40 Also, 18 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

although public AHC hospitals depend more percentage points in 1998 to 2.3 percent, on Medicaid as a payment source, which is about half the margin of non-teaching hospi- different from private AHC hospitals which tals. Dick Knapp, executive vice president of depend more on Medicare and private pay- the AAMC, evaluated this situation and com- ers, high levels of uncompensated care have mented that the negative margins reported by reduced their fi nancial strength.41 a number of major teaching hospitals may be Moreover, the Association of American partly attributed to the increases in the cost Medical Colleges (AAMC) predicted on of new drugs, medical devices, information its Web site in 2000 that nearly 50 percent technology, and labor, as well as the declining of all AMCs that are major teaching hospi- revenues from managed care, Medicare, and tals would have negative margins by 2002, Medicaid, due in part to the BBA of 1997.44 compared to only about 25 percent of non- In order to gain a better understanding teaching and other teaching institutions. of the factors driving such margin differ- According to the AAMC data, negative ences and to achieve our research objective, margins for major teaching hospitals would we conducted a comprehensive review of double between 1996 and 2002 due to the prior studies regarding teaching hospitals’ provision of the BBA. Besides, while no performance analysis and categorized our absolute standard exists, total margins below observations of the possible factors driving 4 percent are thought of as insuffi cient to differences into the following eight groups: sustain teaching hospitals. In sum, the enactment of the BBA of 1997 has added consider- 1. Cost differences between teaching and able fi nancial pressure to an already diffi cult non-teaching hospitals; fi nancial environment. As such pressure may 2. Cost differences between urban and threaten the teaching, research, and indigent rural hospitals; care missions of teaching hospitals, appropri- 3. Productivity (effi ciency) differences ate policy modifi cations may be necessary to between teaching and non-teaching ensure teaching hospitals fulfi ll their special hospitals; missions with suffi cient fi nancial resources.42 4. Competitive effects (effects of growth of managed care) on teaching hospitals; Possible Determinants of Performance 5. Impacts of organizational features on Differences Between Teaching and teaching hospitals; Non-Teaching Hospitals 6. Form of ownership; Due to both public policy changes and 7. Signifi cance of size on teaching hospi- market reforms, the US health care industry tals’ performance; and is facing a number of critical changes. While 8. Margin differences between teaching all hospitals are seeing less income after and non-teaching hospitals. expenses given these changes, major teaching hospitals are experiencing a particularly tight fi nancial squeeze. Cost Differences Between Teaching As the Medicare Payment Advisory Com- and Non-Teaching Hospitals mission reported in 2000,43 the total margin Teaching hospital costs are substantially for major teaching hospitals declined by 2.8 higher than non-teaching hospital costs.45 A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 19

For example, Foster46 examined the estima- hospitals and thus incur higher costs. Rosko tion of the costs of the products of teaching proposed that the possible explanations for hospitals through a review of the literature this phenomenon may include: and found that the following factors have the potential to explain why the costs of the 1. Large size of teaching hospitals; products of teaching hospitals are greater 2. Educational value of attracting patients than those of non-teaching hospitals: with a wide variety of diagnoses;51 and 3. Generous indirect medical education 1. Complexity of case mix; payments from Medicare as a subsidy. 2. Severity of illness; 3. Research activity; 4. Use of innovative modes of treatment; Cost Differences Between Urban and Rural Hospitals and 5. Teaching of students. An analysis of the 1981 Medicare cost report data showed that average costs per In addition, Foster’s study suggested the case in rural hospitals were about 40 per- importance of designing a framework that cent lower than that in urban hospitals. could determine reliable, standard costs for Even after accounting for differences in diagnostic case types. Rosko et al.47 and case mix, labor costs, and indirect teaching Thorpe,48 after controlling for a number of costs, a difference of more than 20 percent characteristics such as location, patient mix, still remained. Although the urban-rural and size, reported that major teaching hospi- cost differential has often been attributed tals may be 30 to 42 percent more expensive to systematic differences in patient sever- than non-teaching hospitals. ity of illness across hospital groups that Koenig et al.49 estimated the mission- are not accounted for by the current DRG related costs of teaching hospitals and system, the results of O’Dougherty et al.52 found that AHCs and other teaching hos- suggested that most of the cost difference pitals face higher patient care costs than between urban and rural hospitals is not non-teaching hospitals, because of their attributable to within-DRG differences in missions of GME, biomedical research, average patient severity. and the maintenance of standby capacity Moreover, Thorpe53 investigated the cost for medically complex patients. More spe- differences between inner city hospitals ciﬁ cally, it’s estimated that total mission- and suburban and rural hospitals and found related costs were $27 billion in 2002 for that the key sources of cost variation all teaching hospitals, with GME (includ- include: ing indirect and direct GME) and standby capacity accounting for roughly 60 percent 1. The higher costs of treating low-income and 35 percent of these costs, respectively. patients; Moreover, Rosko50 indicated that teach- 2. The difference in teaching commitment; ing hospitals, especially those with substan- 3. Market competition; tial GME programs, provide larger volumes 4. Higher numbers of emergency room of uncompensated care than non-teaching patients; 20 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

5. The ratio of forecasted to actual admis- due to the originally relatively high pay- sions; and ments of both. 6. Wages. Productivity (Effi ciency) Differences More specifi cally, the most expensive Between Teaching and Non-Teaching Hospitals urban hospitals have costs over 5.3 percent higher than average and nearly one half of If the productivivty of medical labor is lower this higher cost can be attributed to large in teaching hospitals, it may be concluded that teaching programs. In addition, case mix teaching hospitals are less effi cient and thus accounts for about 20 percent of the higher- more costly than their non-teaching coun- than-average costs; higher salaries and terparts. The evidence can be obtained from cost-increasing competition for physicians the following examples. First, Lehrer et al.55 explain 25 percent of the higher costs; and found that attending physicians’ productivity over 7 percent of these costs can be traced is lower in teaching hospitals than that in non- to emergency admissions, usually involving teaching hospitals. Second, Kralewski et al.56 poor patients. corroborated Lehrer et al.57 fi ndings by indi- In October 1983, Medicare changed the cating that productivity of clinical practition- payment method for inpatient hospital ser- ers is lower in teaching hospitals than that in vices from retrospective, cost-based reim- non-teaching clinics. bursement to a PPS, under which Medicare pays a fi xed basic rate for each inpatient Competitive Effects (Effects of Growth of stay. Initially, separate standard rates were Managed Care) on Teaching Hospitals established for urban and rural hospitals due The penetration of managed care could to the existence of the above-described cost drive increasing price competition for teach- differential; however, the Omnibus Budget ing hospitals. How does this market competi- Reconciliation Act of 1990 (OBRA 1990) tion affect teaching hospitals’ performance? phased out the separate standard rates for Will such market pressure hinder hospitals’ urban and rural hospitals between fi scal year performance or will it result in more effi - (FY) 1991 and FY 1995. cient hospitals providing the social good of O’Dougherty et al.54 assessed the need medical education? for payment adjustments in a PPS without According to a study by Pardes,58 teach- separate urban and rural rates and found ing hospitals should become as effi cient as that under a single standardized payment possible via decreasing costs, reducing the and current-law adjustments, there was a number of medical residents, and providing need to redistribute payments from rural to more primary care to respond to the growth urban hospitals. They further suggested that of managed care. Additionally, Grosskopf although refi nements for case mix, outliers, et al.59 assessed the performance of US teach- and the wage index can make a signifi cant ing hospitals operating in 1995 and found contribution to avoiding payment dispari- that competition (as measured by the number ties in a single-rate system, changes in the of managed care contracts per hospital and adjustments for teaching and disproportion- the number of patients covered by these con- ate-share (DSH) hospitals are also needed tracts per hospital) has positive effects on the A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 21

teaching hospitals; in other words, as compe- hospital is not required to pay taxes and is tition increases, so does the teaching hospi- not allowed to distribute earnings. Such non- tal’s relative effi ciency. Moreover, increased distribution constraint hence creates a dif- competition leads to higher effi ciency with- ferent competitive environment compared out compromising teaching intensity. to for-profi t hospitals. As Property Rights Rosko60 conducted a panel analysis of cost Theory suggests, the variable of ownership ineffi ciency on performance of US teaching form (FP) should be negatively associated hospitals and found that: with costs and ineffi ciency. Further, since teaching hospitals have a much different 1. Decreases in ineffi ciency are associ- cost structure, regulatory environment, and ated with the HMO penetration rate mission that is linked closely to research and and time; and teaching rather than just providing care,62 2. Increases in ineffi ciency are associated the separation of non-profi t versus for-profi t with for-profi t ownership status and in addition to teaching versus non-teaching Medicare share of admissions. hospitals is thus important and necessary for our analysis. Impacts of Organizational Features on Teaching Hospitals Signifi cance of Size on Teaching Teaching hospitals vary among them- Hospitals’ Performance selves in terms of standards and the severity Despite the general belief that the of patients’ case mix. Do such differences increased size of health care organizations in organizational features matter in account- may reduce costs through economies of ing for teaching hospitals’ performance? scale and scope, a review of prior literature Grosskopf et al.61 used two proxy variables for shows mixed results regarding the existence standards and case mix to examine this issue: of substantial scale economies in hospitals.63 For example, Carr et al.64 analyzed data on 1. Whether the hospital is a member of the 3,147 US hospitals using the average daily COTH; and patient census and a quadratic cost function 2. Whether the hospital has a medical and concluded that scale economies existed school affi liation (MDSCH). and the long-run average cost function reached a minimum at approximately 190 Their fi ndings indicated that COTH members patients. Lave et al.65 used an elaborate two- and hospitals with medical school affi liation stage model to estimate a Cobb-Douglas are relatively less effi cient than the hospitals hospital cost function and found that size without such memberships, and they maintain (defi ned by number of operating beds) was higher levels of teaching dedication (the num- negative in sign (economies of scale) but ber of residents per physician) and intensity insignifi cant. In other words, if economies (the number of residents per bed). of scale exist in the hospital industry, they are not very strong. Bays66 used a quadratic Form of Ownership cost function specifi cation for the number of The difference between for-profi t and beds, case mix, and case fl ow rate for teach- non-profi t hospitals lies in that—a non-profi t ing and non-teaching hospitals in California. 22 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

His fi ndings showed that case mix and case have lower costs due to affi liation. Further, fl ow rate are signifi cant and demonstrate Connor et al.71 reported that hospital merg- economies of scale but that the size of the ers have had a greater effect on bargaining hospital (beds) is not signifi cant. However, power than on costs. the coeffi cients of the linear and quadratic terms on size are of the right sign to indicate Margin Differences Between Teaching economies of scale. and Non-Teaching Hospitals Yafchak67 utilized the Cobb-Douglas Rosko72 employed a panel design to ana- production function to empirically esti- lyze changes in performance variables related mate whether or not larger hospitals have to profi tability, volume, and effi ciency in a lower long-run average costs per bed than national sample of major teaching hospitals smaller hospitals. His results indicated that from 1990 to 1999. Rosko’s fi ndings demon- although economies of scale have evolved strated that: recently for non-teaching and teaching hospitals, the primary market forces that may 1. Average operating margin was negative be creating economies of scale in the hospi- throughout the 1990s, which was in tal industry are decreasing revenues due to contrast with the positive mean operat- lower reimbursement and lower occupancy ing margin for non-teaching hospitals rates. As reimbursement and occupancy in six years of the last decade;73 and continue to trend down in the future, Yaf- 2. Major teaching hospitals responded chak further suggested that hospitals may to fi nancial pressures by downsizing seek additional economies to survive in inpatient capacity, expanding outpatient an increasingly competitive and shrinking activity, reducing length of stay, and industry. Moreover, Rosko,68 in his panel increasing labor productivity. analysis of major teaching hospitals’ performance, found that major teaching hospi- As Rosko suggested, the difference in tals increased their participation in systems, operating margins between major teaching hoping to reap the benefi ts of fi rm-level hospitals and non-teaching hospitals prob- economies of scale and scope and increased ably refl ects the higher cost structures and bargaining power. Compared to that in pri- greater uncompensated care burdens of vate major teaching hospitals where the rate major teaching hospitals. increased from 46 percent to 73 percent, How do teaching hospitals deal with the this trend was especially apparent in public pressure caused by the higher cost structures major teaching hospitals that had a 9 per- and greater uncompensated care burden cent system participation rate in 1990 and a described above? Langabeer74 examined the 40.3 percent rate in 1999. fi nancial and operating data for 100 major US Contrary to prior studies, Friedman et al.69 teaching hospitals to determine relationships found that hospitals’ average costs increase among competitive strategy, market envi- slightly with an increase in hospital size. ronment, and fi nancial return on invested Dranove70 studied whether or not system capital. Langabeer’s fi ndings indicated that affi liation results in economies of scale the single most signifi cant competitive strat- and found that systems do not appear to egy for improving fi nancial performance A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 23

was pricing strategy. As Langabeer stated, numbers of emergency room patients, organizations usually have two options to the ratio of forecasted to actual admis- increase their margins due to the only two sions, and wages; components of pricing strategy at a macro 2. Productivity differences; level: price and cost. The fi rst option is to 3. Competitive effects of growth of man- charge the same price as their competitors as aged care; long as they have a substantially lower cost 4. Organizational characteristics such as advantage. The second option is by charg- whether the hospital is a member of the ing higher prices than competitors, which COTH and whether the hospital has a only the high-performing hospitals are able medical school affi liation (MDSCH); to do. Due to the evolution of managed and care reform efforts by commercial insurers 5. Ownership form. and HMOs, teaching hospitals will not be allowed to exert much control over prices in To elaborate further, large non-profi t the future. Since they are facing increasingly urban teaching hospitals, due to their higher stricter price competition, their ability to complexity of case mix, greater severity of charge higher prices, even for higher quality patients’ illness, greater use of innovative and better positioned services, could become technology for treatment, missions of edu- much more diffi cult. Therefore, Langabeer cation, research, and patient care, provision further suggested that strategic management of larger volumes of uncompensated care, of teaching hospitals requires improvement larger size, higher level of teaching commit- of pricing strategies by focusing on both ments, higher labor cost, lower effi ciency, sides of the price-cost equation; that is, to increasing price competition resulting from reduce cost or improve internal effi ciency growth of managed care, organizational while maintaining or increasing prices for characteristics (teaching status/affi liation), services where possible. and non-profi t ownership form, tend to incur higher fi xed cost and lower variable Summary cost, have lower CMI-adjusted volume, To summarize the above discussion, the and thus create lower ROA, which can be possible factors driving performance dif- decomposed into two elements: profi t mar- ferences between large non-profi t urban gin and asset turnover. teaching hospitals and small for-profi t rural/ On the contrary, small for-profi t rural/ suburban non-teaching hospitals may suburban non-teaching hospitals, due to include: their lower complexity of case mix, less severity of patients’ illness, less use of 1. Cost differences resulting from com- innovative technology for treatment, provi- plexity of case mix, severity of illness, sion of smaller volumes of uncompensated use of innovative modes of treat- care, smaller size, lower level of teaching ment, missions of GME, biomedical commitments, lower wages, higher effi - research and patient care, provision ciency, lower penetration of managed care of uncompensated care, size, market in their located areas, organizational char- competition, teaching commitments, acteristics (non-teaching status/affi liation), 24 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

and for-profi t ownership form, are prone to III. Research Design incur lower fi xed cost and higher variable cost, have higher CMI-adjusted volume, Hypotheses and thus create higher ROA. Based on the foregoing discussion, we The above elaboration is consistent with developed two hypotheses, as follows: Langabeer’s fi ndings75 in product market and capital investment strategy. As Lang- H 1: Large nonprofi t urban teaching hospi- abeer76 stated, product market strategy in tals tend to have lower CMI adjusted volume, hospitals equates to a selection of the opti- higher fi xed cost, lower variable cost, and mal patient mix. Langabeer’s analysis con- lower ROA (profi t margin x asset turnover). fi rmed that the product market strategy is H 2: Small for-profi t rural/suburban non- one of the most fundamental elements that teaching hospitals tend to have higher CMI hospital administrators should consider adjusted volume, lower fi xed cost, higher when formulating a strategic plan. Selecting variable cost, and higher ROA (profi t margin the appropriate product market strategy will x asset turnover). not only enhance pricing strategy but also boost return on invested capital. More spe- We use the following four cross-sectional cifi cally, as hospitals select more complex regression models to test the above two and specialized procedures, as indicated by hypotheses: the adjusted CMI index, the expected fi nan- FC/TC = a0 + a1*Medicare % + cial performance decreases; the other way a2*Medicaid % + a3*DSH + a4*CMI + round is also true. a5*RESEAR + a6*GME + a7*OCC + Additionally, as teaching hospitals’ port- a8*OUTPAT + a9*RESBED + a10*SYS + folios of services and programs increase, it a11*TH + error (1) becomes strategically necessary to manage the allocation of resources or investments VC/TC = a0 + a1*Medicare % + into the fi xed capital (i.e., property, plant, a2*Medicaid % + a3*DSH + a4*CMI + and equipment) that supports the business. a5*RESEAR + a6*GME + a7*OCC + Langabeer’s fi ndings confi rmed that capi- a8*OUTPAT + a9*RESBED + a10*SYS + tal intensity (the overall level of resources a11*TH + error (2) invested into the assets of an organization and the average age of the capital infra- VOL = a0 + a1*Medicare % + structure) is signifi cantly negatively related a2*Medicaid % + a3*DSH + a4*CMI + to performance; that is, the more diversi- a5*RESEAR + a6*GME + a7*OCC + fi ed product mix teaching hospitals select, a8*OUTPAT + a9*RESBED + a10*SYS + the more capital they will invest into fi xed a11*TH + error (3) assets, and thus the lower expected return they will get eventually. ROA = a0 + a1*Medicare % +

Based on the above elaboration, two a2*Medicaid % + a3*DSH + a4*CMI + hypotheses were thus developed in Section a5*RESEAR + a6*GME + a7*OCC +

III and were tested using the CVP and multi- a8*OUTPAT + a9*RESBED + a10*SYS + ple regression analysis. a11*TH + error (4) A Comparative Analysis: Nonproﬁ t Teaching and For-Proﬁ t Non-Teaching Hospitals 25

Where assets. Based on the discussion in the literature review section (Section II), all inde- FC/TC = Fixed cost percentage, pendent variables except OCC and OUTPAT fi xed cost / total operating were expected to have a positive coeffi cient cost; in models (1) and a negative coeffi cient in VC/TC = Variable cost per- models (2), (3), and (4). centage, variable cost/ The test variable, TH, is defi ned as 1 if total operating cost; a hospital is a teaching hospital, has a non- Volume = Total revenue, adjusted profi t ownership, is located in an urban area, by case mix index (CMI); and has more than 400 beds, and 0 if not. If a ROA = Return on assets, net large nonprofi t urban teaching hospital tends income / total assets; Medicare % =100*Medicare dischar- to have higher fi xed cost, lower variable ges/total discharges; cost, lower CMI adjusted volume, and lower Medicaid % =100*Medicaid dischar- ROA, then the coeffi cient on TH should be ges/total discharges; positive in model (1) and negative in models DSH = 1 if a disproportionate (2), (3), and (4). share hospital, 0 otherwise; CMI = Medicare CMI; RESEAR = Research expenditure; Data Source GME = Graduate medical educa- The study data were obtained from the tion payments; Medicare Cost Report. The report is not only OCC = Occupancy rate, 100 one of the most comprehensive data sets *(patient days/365* licensed beds); available for every hospital in the United OUTPAT = % of outpatient revenue States that services Medicare patients and to total revenue; receives federal reimbursement, it also RESBED = Number of residents captures a variety of hospital income state- per bed; ments, balance sheets, and operational SYS = 1 if a system affi liated statistics for the entire hospital and is not hospital, 0 otherwise; limited to Medicare. Because of this, it can TH = 1 if a large nonprofi t be deemed as one of the best sources avail- urban teaching hospital, 0 able for national fi nancial data on hospitals if a small for-profi t rural/ and thus was was chosen as our primary suburban non-teaching data source. hospital. The dependent variables are FC/TC, VC/ Sample Selection TC, VOL, and ROA for each of the above four models respectively. FC/TC measures The unit of analysis is large teaching hos- the percentage of fi xed cost to total operat- pitals, with a non-profi t status and situated in ing cost; VC/TC measures the percentage an urban area. To achieve the research objec- of variable cost to total operating cost; VOL tive, a control sample of small non-teaching measures total revenue adjusted by CMI; and hospitals, with a for-profi t status and situated ROA measures the hospital’s return on total in a rural or suburban area, was also chosen. 26 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Hospitals with 400 beds or above are con- CVP analysis is often criticized for having sidered as large; otherwise, as small. Both some shortcomings, it is pliable enough sample sets were selected from the Medicare to overcome all of them, if necessary, and Cost Report for the single year 2005 because desirable, since most of these shortcomings of the restricted nature of the CVP analysis. are related to its basic underlying assump- We begin with 2,669 hospitals that have tions. Moreover, its restricted scope can a fi scal year end dated after July 1, 2005 be broadened with an extended version of from the Medicare Cost Report database. the basic model designed to mitigate cer- For those hospitals which fi led two medical tain shortcomings.77 As a consequence, it is reports in year 2005, we keep only the latest believed to be a very useful initial analysis fi ling in the sample. Therefore, 370 hospi- of strategic decisions.78 tals were deleted from this procedure. We In addition to its merit of simplicity, then removed 315 hospitals which did not through the provision of a sweeping fi nan- have data for teaching status in the Medi- cial overview of the planning process,79 it care Cost Report. The remaining 1,984 allows managers to examine the possible hospitals were partitioned into two sam- effects of a wide array of strategic decisions, ples: 549 teaching and 1,435 non-teaching including pricing policies, product mixes, hospitals. The teaching sample includes market expansions or contractions, outsourc- 470 hospitals with a non-profi t ownership ing contracts, idle plant usage, discretionary form. Among this group, 440 hospitals were expense planning, and a variety of other located in an urban areas and 122 of them important considerations in the planning had more than 400 beds. From the group of process. The key in CVP analysis would be 122 hospitals, fi ve were deleted due to their target income, which is defi ned as the mini- missing value for certain variables. The non- mum net income acceptable for a particular teaching sample includes 392 hospitals with decision. This minimum income level is for- a for-profi t ownership form. Among this mulated as the product of cost of capital and group, 102 hospitals were located in either total assets required for a decision. By using rural or suburban areas and all of them had CVP analysis, a hospital would be able to fewer than 400 beds. Thus, the fi nal sam- determine if a particular pricing policy and ple includes 117 teaching hospitals and 102 cost structure strategy might yield at least non-teaching hospitals. a minimum target income level that would meet the hospital’s cost of capital.

Research Method IV. Results Due to our objective to compare the CVP structure of large non-profi t urban Figure 1 provides descriptive evidence teaching hospitals and small for-profi t rural/ about the sample hospitals. The mean suburban non-teaching hospitals, the CVP (median) values of Medicare and Medic- analysis was conducted to test our hypoth- aid % are 42.90 (38.39) and 16.40 (14.17), eses. CVP analysis can be seen as one of respectively. On average, 84 percent of the the most powerful and simplest analytical sample hospitals are qualifi ed to receive tools in management accounting. Although the disproportionate share payments. The A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 27

Figure 1. Descriptive Statistics (n = 219)

25th 75th Variable Mean SD Percentile Median Percentile Medicare % 42.90 42.32 29.06 38.39 47.52 Medicaid % 16.40 13.23 7.32 14.17 21.37 DSH 0.84 0.37 1.00 1.00 1.00 CMI 1.44 0.36 1.11 1.44 1.71 RESEAR 2621433.96 13633160.75 0.00 0.00 102755 .00 GME 2374800.02 3949901.50 0.00 266006.00 3660212.00 OCC 62.93 48.22 40.45 65.97 77.63 OUTPAT 0.39 0.15 0.29 0.37 0.46 RESBED 0.16 0.25 0.00 0.01 0.23 SYS 0.09 0.29 0.00 0.00 0.00 Note: The sample includes 117 teaching hospitals and a control sample of 102 non-teaching hospitals from the Medicare Cost Report database with a fi scal year end dated after July 1, 2005. For those hospitals which fi led two medical reports in 2005, we keep only the latest fi ling in the sample. The variables are defi ned as follows:

Medicare % = 100 * Medicare discharges / total discharges; Medicaid % = 100 * Medicaid discharges / total discharges; DSH = 1 if a disproportionate share hospital, 0 otherwise; CMI = Medicare case mix index; RESEAR = Research expenditure; GME = Graduate medical education payments; OCC = Occupancy rate, 100 * (patient days / 365 * licensed beds); OUTPAT = % of outpatient revenue to total revenue; RESBED = Number of residents per bed; SYS = 1 if a system affiliated hospital, 0 otherwise. sample hospitals have a mean (median) per bed is 16 (1) percent. Only 9 percent CMI of 1.44 (1.44) percent. The average of the sample hospitals are affi liated with a research expenditures and GME payments medical system. are 2621433.96 and 2374800.02. The sam- Figure 2 provides results from compari- ple fi rms have a mean (median) occupancy sons of the teaching and non-teaching groups rate of 62.93 (65.97) percent. On average, for the variables included in our regression 39 percent of total revenue is outpatient rev- model. The data show that there are signifi - enue. The mean (median) ratio of residents cant differences between the two groups in 28 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 2. Mean (Median) Values for Teaching and Non-Teaching Groups

Teaching Group Non-Teaching Group p from T-test Variable (n = 117) (n = 102) (Mann-Whitney U test)

Medicare % 32.17 55.21 0.00 (32.36) (46.93) (0.00) Medicaid % 14.44 18.64 0.02 (12.67) (16.55) (0.02) DSH 0.88 0.79 0.08 (1.00) (1.00) (0.08) CMI 1.72 1.12 0.00 (1.69) (1.11) (0.00) RESEAR 4906786.64 0.00 0.01 (23597.00) (0.00) (0.00) GME 4427106.37 20683.92 0.00 (3360043.00) (0.00) (0.00) OCC 74.23 49.97 0.00 (76.30) (40.68) (0.00) OUTPAT 0.32 0.46 0.00 (0.31) (0.46) (0.00) RESBED 0.29 0.01 0.00 (0.18) (0.00) (0.00) SYS 0.09 0.10 0.75 (0.00) (0.00) (0.75)

Note: The variables are defi ned as in Figure 1. terms of all the control variables considered Panel B shows that the mean (median) pro- in this study except SYS. portion of variable cost to total operating Figure 3 provides results from comparisons cost is 0.60 (0.61) for the teaching group; of the teaching and non-teaching groups the corresponding value is 0.63 (0.64) for the for the dependent variables included in our non-teaching group. Panel C shows that the regression model. Panel A shows that the mean (median) value of the CMI adjusted mean (median) proportion of fi xed cost to revenue is 869508287.76 (796109726.64) total operating cost is 0.40 (0.39) for the for the teaching group; the corresponding teaching group; the corresponding value value is 46868431.22 (29592191.45) for is 0.37 (0.36) for the non-teaching group. the non-teaching group. Panel D shows that A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 29

Figure 3. Fixed Cost Percentage, Variable Cost Percentage, CMI Adjusted Volume, and Return on Total Assets in Teaching and Non-Teaching Samples

Panel A: Fixed Cost Percentage Teaching Non-Teaching Sample Sample Variable ( n = 117) ( n = 102) Mean 0.40 0.37 S.D. 0.06 0.06 25th percentile 0.35 0.33 Median 0.39 0.36 75th percentile 0.43 0.40

Panel B: Variable Cost Percentage Teaching Non-Teaching Sample Sample Variable ( n = 117) ( n = 102) Mean 0.60 0.63 S.D. 0.06 0.06 25th percentile 0.57 0.60 Median 0.61 0.64 75th percentile 0.65 0.67

Panel C: CMI Adjusted Volume Teaching Non-Teaching Sample Sample Variable ( n = 117) ( n = 102) Mean 869508287.76 46868431.22 S.D. 481738752.29 57739581.32 25th percentile 589811528.99 15046089.88 Median 796109726.64 29592191.45 75th percentile 1033368941.96 48650078.94

Panel D: Return on Total Assets Teaching Non-Teaching Sample Sample Variable ( n = 117) ( n = 102) Mean Ϫ0.11 0.19 S.D. 0.48 0.17 25th percentile Ϫ0.07 0.08 Median 0.00 0.13 75th percentile 0.02 0.25

Note: This fi gure presents descriptive statistics about the fi xed cost percentage, variable cost percentage, CMI adjusted volume, and return on total assets for the teaching and non-teaching samples. 30 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

the mean (median) return on total assets is and GME payments are more likely to have Ϫ.11 (.00) for the teaching group; the corre- higher a variable cost percentage, which sponding value is .19 (.13) for the non-teach- is against our expectation. The coeffi cient ing group. Both a T-test of the mean and a for TH is negative and signifi cant, suggest- Wilcoxon test of the median confi rm that ing that large nonprofi t teaching hospitals the two groups are signifi cantly different located in urban areas are more likely to have (p < .01) in terms of fi xed cost percentage, a lower proportion of variable cost to total variable cost percentage, CMI adjusted vol- operating cost. Or put another way, small ume, and return on assets. for-profi t non-teaching hospitals situated in rural or suburban areas are more likely to Regression Results have a higher proportion of variable cost to Figure 4 presents the results from the total operating cost. regression. The dependent variable in Panel The dependent variable in Panel C is the A is the proportion of fi xed cost to total CMI-adjusted revenue. The overall regres- operating cost. The overall regression is sig- sion is signifi cant (p < .001). Consistent with nifi cant (p < .001). Consistent with expecta- expectations, the coeffi cients for OCC and tions, the coeffi cient for RESEAR is positive OUTPAT are positive and signifi cant, indi- and signifi cant, indicating that hospitals cating that hospitals with a higher occupancy with higher research expenditures are more rate and higher proportion of outpatient rev- likely to have a higher proportion of fi xed enue to total revenue are more likely to have cost to total operating cost. The coeffi cients a higher total revenue adjusted by CMI. The for CMI and GME are negative and signifi - coeffi cients for CMI, RESEAR, and GME cant, indicating that hospitals with higher are positive and signifi cant, indicating that CMI and GME payments are less likely to hospitals with higher CMI, higher research have a higher fi xed cost percentage, which expenditures, and higher GME payments are is against our expectation. The coeffi cient more likely to have higher CMI-adjusted rev- for TH is positive and signifi cant, suggest- enue, which is against our expectation. The ing that large nonprofi t teaching hospitals coeffi cient for TH is positive and signifi cant, located in urban areas are more likely to suggesting that large nonprofi t teaching hos- have a higher proportion of fi xed cost to total pitals located in urban areas are more likely operating cost. to have higher total revenue adjusted by The dependent variable in Panel B is the CMI, which is also against our expectation. proportion of variable cost to total operat- The dependent variable in Panel D is the ing cost. The overall regression is signifi cant return on total assets. The overall regression is (p < .001). Consistent with expectations, signifi cant (p < .001). Consistent with expec- the coeffi cient for RESEAR is negative and tations, the coeffi cients for Medicaid % and signifi cant, indicating that hospitals with RESEAR are negative and signifi cant, indicat- higher research expenditures are more likely ing that hospitals with higher Medicaid per- to have a lower proportion of variable cost centage and research expenditures are more to total operating cost. The coeffi cients for likely to have a lower return on assets. Also CMI and GME are positive and signifi cant, consistent with expectations, the coeffi cient indicating that hospitals with higher CMI for OCC is positive and signifi cant, indicating A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 31

Figure 4. Regression Results

Panel A:

FC/TC = a0 + a1*Medicare % + a2*Medicaid % + a3*DSH + a4*CMI + a5*RESEAR + a6*GME + a7*OCC +

a8*OUTPAT + a9*RESBED + a10*SYS + a11*TH + error Variable Expected Sign Coefficient T-Statistic p-Value Constant .467 14.175 <.001 Medicare % + Ϫ.030 Ϫ.198 .421 Medicaid % + Ϫ.082 Ϫ.866 .193 DSH + Ϫ.084 Ϫ1. 17 0 .121 CMI + Ϫ.394 Ϫ3.252 <.001 RESEAR + .190 2.572 .005 GME + Ϫ.187 Ϫ1.502 .067 OCC Ϫ .003 .017 .493 OUTPAT ϪϪ.026 Ϫ.344 .365 RESBED + .108 .848 .198 SYS + .064 .986 .162 TH + .551 4.213 <.001

F = 3.27, p < .001; adjusted R-square = .10

Panel B:

VC/TC = a0 + a1* Medicare % + a2*Medicaid % + a3*DSH + a4*CMI + a5*RESEAR + a6*GME + a7*OCC +

a8*OUTPAT + a9*RESBED + a10*SYS + a11*TH + error Variable Expected Sign Coefficient T-Statistic p-Value Constant .533 16.175 <.001 Medicare % Ϫ .030 .198 .421 Medicaid % Ϫ .082 .866 .193 DSH Ϫ .084 1.170 .121 CMI Ϫ .394 3.252 <.001 RESEAR ϪϪ.190 Ϫ2.572 .005 GME Ϫ .187 1.502 .067 OCC + Ϫ.003 Ϫ.017 .493 OUTPAT + .026 .344 .365 RESBED ϪϪ.108 Ϫ.848 .198 SYS ϪϪ.064 Ϫ.986 .162 TH ϪϪ.551 Ϫ4.213 <.001

F = 3.27, p < .001; adjusted R-square = .10 Continues on next page... 32 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 4. Regression Results (continued)

Panel C:

VOL = a 0 + a1*Medicare % + a2*Medicaid % + a3*DSH + a4*CMI + a5*RESEAR + a6*GME + a7*OCC +

a8*OUTPAT + a9*RESBED + a10*SYS + a11*TH + error

Variable Expected Sign Coefficient T-Statistic p-Value Constant Ϫ654901828.776 Ϫ5.038 <.001 Medicare % ϪϪ.073 Ϫ1.066 .144 Medicaid % ϪϪ.029 Ϫ.670 .252 DSH ϪϪ.039 Ϫ1.209 .114 CMI Ϫ .379 6.861 <.001 RESEAR Ϫ .226 6.695 <.001 GME Ϫ .271 4.772 <.001 OCC ϩ .125 1.642 .051 OUTPAT ϩ .063 1.847 .033 RESBED Ϫ .014 .249 .402 SYS Ϫ .025 .837 .202 TH Ϫ .225 3.773 <.001

F = 87.66, p < .001; adjusted R-square = .81

Panel D:

ROA = a0 + a1*Medicare % + a2*Medicaid % + a3*DSH + a4*CMI + a5*RESEAR + a6*GME + a7*OCC +

a8*OUTPAT + a9*RESBED + a10*SYS + a11*TH + error

Variable Expected Sign Coefficient T-Statistic p-Value Constant Ϫ.049 Ϫ.425 .671 Medicare % Ϫ .028 .775 .219 Medicaid % ϪϪ.049 Ϫ1.204 .100 DSH Ϫ .025 .717 .237 CMI Ϫ .014 .282 .389 RESEAR ϪϪ.778 Ϫ23.720 <.001 GME ϪϪ.038 Ϫ.967 .167 OCC + .233 6.403 <.001 OUTPAT + .019 .552 .290 A Comparative Analysis: Nonproﬁ t Teaching and For-Proﬁ t Non-Teaching Hospitals 33

Figure 4. Regression Results (continued)

Variable Expected Sign Coefficient T-Statistic p-Value RESBED Ϫ .018 .460 .323 SYS Ϫ .033 1.072 .142 TH ϪϪ.081 Ϫ1.603 .046

F = 76.24, p < .001; adjusted R-square = .79

Note: This fi gure (including four panels) presents the results from a regression model where the dependent variable is fi xed cost percentage, variable cost percentage, CMI adjusted volume, and return on total assets for each panel, respectively. The sample includes 117 teaching hospitals and a control sample of 102 non- teaching hospitals from the Medicare Cost Report database, which have a fi scal year end dated after July 1, 2005. The variables are defi ned as in Figure 1. p-values are two-tailed. that hospitals with a higher occupancy rate are fi xed cost and lower variable cost. While such more likely to have a higher return on assets. cost structure doesn’t necessarily lead to lower The coeffi cient for TH is negative and sig- revenue adjusted by CMI, it indeed results in nifi cant, suggesting that large nonprofi t teach- lower profi t or asset turnover in terms of ROA. ing hospitals located in urban areas are more On the contrary, small for-profi t non-teaching likely to a have lower return on assets. hospitals located in rural or suburban areas Overall, the results provide empirical sup- tend to have a lower fi xed cost, higher vari- port to our hypothesis that large nonprofi t able cost, and higher profi t or asset turnover in urban teaching hospitals, due to their mis- terms of ROA due to their different features. sions of research and patient care, provision Langabeer80 stated that product market of larger volumes of uncompensated care, strategy is one of the most fundamental ele- larger size, higher level of teaching commitments that hospital administrators should ments, higher labor cost, lower effi ciency, consider when formulating a strategic plan. increasing price competition resulting from In other words, as hospitals select more com- growth of managed care, teaching status, plex and specialized procedures, the expected and non-profi t ownership form, tend to fi nancial performance decreases; the other incur a higher fi xed cost and lower variable way round is also true. While the results in cost. Such cost structure doesn’t necessarily this study seem not fully consistent with affect their total revenue adjusted by CMI; those of Langabeer81 in the effects of adjusted however, it does lead to a lower return on CMI on hospitals’ performance, the fi nd- hospitals’ total assets. ings do indicate that large nonprofi t urban teaching hospitals, because of their larger V. Conclusion size, nonprofi t ownership form, location and increasing price competition, teaching status, The empirical results confi rm our hypoth- mission of research and patient care, higher esis that large nonprofi t teaching hospitals labor cost and lower effi ciency, provision located in urban areas tend to have a higher of uncompensated care to larger volume of 34 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Medicaid patients, lower occupancy rate, CVP structure between teaching and non- and lower proportion of outpatient revenue teaching hospitals, without looking at other to total revenue, are more likely to incur a performance measures such as cash fl ow higher fi xed cost and thus lower ROA. ratio, bond rating, investment of technol- The fi ndings suggest that cost structure is ogy, and quality of care. Second, this study signifi cantly associated with hospitals’ per- focuses on the analysis of teaching hospitals, formance. That is, the higher proportion the without looking at AHCs and AMCs. Third, fi xed cost to total operating cost in a hospi- the use of the CVP technique to conduct tal’s cost structure, the lower return in terms analysis is subject to the following addi- of ROA will be expected. The fi ndings also tional limitations: suggest that as teaching hospitals’ portfolios of services and programs increase (e.g., pro- 1. The CVP analysis assumes that vision of uncompensated care to Medicare changes in volume have no effect on and Medicaid patients and doing research), elasticity of demand or on the effi - it becomes strategically necessary and criti- ciency of production factors; that is, it cal to manage the allocation of resources or ignores the curvilinear nature of total investments into the fi xed capital that sup- revenue and total cost schedules;82 ports the business. In other words, teaching 2. Since CVP analysis is typically hospitals aiming at improving performance restricted to one time period in each should attempt to do the following: case, this study only spans the single year 2000; 1. Adjust their cost structure; 3. The CVP analysis has a somewhat nar- 2. Evaluate the cost-benefi t of fulfi lling row scope on only sales revenue and the research mission; operating expenses, which could leave 3. Provide less volume of uncompen- some critical aspects of strategic deci- sated care or the same volume of sions overlooked; uncompensated care at higher reim- 4. The CVP analysis doesn’t measure the bursement; and impact of the decision on the hospital’s 83 4. Increase occupancy rate and the pro- wealth; portion of outpatient revenue to total 5. The CVP analysis does not incorporate revenue to achieve their objectives. the effect of asset structure changes required by the decision;84 and 6. The CVP analysis does not acknowl- Hopefully, the fi ndings can be used to edge the risk created by the decision.85 enhance the decision effectiveness of hospital administrators in selecting their product Fourth, this study does not incorporate the mix and designing their CVP structure as shift of services from inpatient to outpa- well as fi nancial strategies, appropriately. tient. Finally, since corporate overhead costs of system-affi liated hospitals are not Limitations included in the total cost reported in the This study doesn’t come without limita- Medicare Cost Report data,86 the focus of this tions. First, this study only compares the study is purely on the individual hospital and A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 35

not system-affi liated hospitals. As a result, impact of managerial compensation schemes whether system-affi liated hospitals perform on target profi t levels.87 Fourth, due to the one- better than their counterparts will not be year nature of CVP analysis, future research evaluated in this article. can employ research techniques other than CVP to explore the long-term performance Possible Future Research differences. Fifth, future research can extend There are some fruitful areas for future the basic model of CVP by entering some research to pursue. First, future research additional variables such as the dichotomy of can examine the performance differences variable and fi xed assets, cost of capital, the between teaching and non-teaching hospitals degree of operating leverage or an accounting in terms of other measures such as cash fl ow beta risk measure to incorporate the wealth ratio, bond rating, investment in technology, effects and the risk level imposed by a deci- and quality of care. Second, future research sion.88 Finally, with available data, future can extend the analysis to include both AHCs research can conduct a comparative analy- and AMCs. Third, future research can widen sis of the CVP structure between system- the scope of the CVP analysis to include the affi liated and non-system-affi liated hospitals.

REFERENCES

1. Ginn, GO “Strategic Change in Hospitals: 6. Guo, KL, “Market-Focused Management: An Examination of the Response of the Acute A Model for US Academic Health Centers,” Care Hospital to the Turbulent Environment Journal of Health Organization and Manage- of the 1980s,” Health Services Research, ment, 17(2): 88–101 (2003). 25(4): 565–591 (1990); Shortell, SM, Morri- 7. Topping, S, Hyde, J, Barker, J, and Woodrell, E, son, EM, and Friedman, B, “Strategic Choices “Academic Health Centers in Turbulent Times: for America’s Hospitals: Managing Change in Strategies for Survival,” Healthcare Manage- Turbulent Times,” San Francisco: Jossey-Bass ment Review, 24: 7–18 (1999). Publishers (1990). 8. Blumenthal, D, Meyer, GS, “Academic Health 2. Langabeer, JR, II, “Competitive Strategy in Centers in a Changing Environment,” Health Turbulent Healthcare Markets: An Analysis Affairs, 15: 200–214 (1996). of Financially Effective Teaching Hospitals,” 9. Reuter, J, Gaskin, D, “Academic Health Cent- Journal of Healthcare Management, 43(6): ers in Competitive Markets,” Health Affairs, 512–525 (Nov./Dec. 1998). 16: 242–252 (1997). 3. Langabeer, JR, II, “Competitive Strategy in 10. Shine, KI, “The Future of Academic Health Turbulent Markets: A Multivariate Analysis of Centers,” The Physiologist, 38: 51–55 (1995). Strategic Management in Teaching Hospitals,” 11. Reuter and Gaskin, supra, n.9. PhD diss., University of Houston (1996). 12. Walsh, AM, Szabat, K, and Grass, LB, Sus- 4. Buzzell, RD, Gale, BT, “The PIMS Principles: taining the Edge: Factors Inﬂ uencing Strategy Linking Strategy to Performance,” New York: Selection in Academic Health Centers,” Jour- The Free Press (1987). nal of Healthcare Management, 47(6): 360– 5. Association of Academic Medical Centers, 374 (Nov./Dec. 2002). www.aamc.org/about/progemph/camcam/ 13. Fox, PD, Wasserman, J, “Academic Medi- factshts/vol4_no3, htm. cal Centers and Managed Care: Uneasy 36 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Partners,” Health Affairs, 12(1): 85–93 (Spring 27. Weissman, JS, “Uncompensated Care: Will It 1993). Be There When We Need It?” Journal of the 14. Younis, M, “The Impact of the Balance Budget American Medical Association, 276(10): 823– Act on Access to Rural Hospital Services for 828 (1996). Medicaid Patients,” Journal of Health Care 28. Weissman, JS, Gaskin, DJ, Reuter, J, “Hospi- Finance, 33(1): 72–82 (Fall 2006). tals’ Care of Uninsured Patients during the 15. Younis, M, Forgione, D, “The Relationship 1990s: The Relation of Teaching Status and Between the Balanced Budget Act and Length Managed Care to Changes in Market Share of Stay for Medicare Patients in US Hospitals,” and Market Concentration,” Inquiry—Excellus European Journal of Health Economics, Health Plan, 40(1): 84–93 (Spring 2003). 10(1): 57–63 (2009). 29. Coughlin, TA, Ku, L, Kim, J, Reforming the Med- 16. Rosko, MD, “Performance of US Teaching Hospi- icaid Disproportionate Share Hospital Program tals: A Panel Analysis of Cost Ineffi ciency,” Health in the 1990s, Washington, DC: Urban Institute Care Management Science, 7: 7–16 (2004). (2000); Medicare Payment Advisory Com- 17. Medicare Payment Advisory Commission mission (MedPAC), Report to the Congress: (MedPAC), Report to the Congress: Selected Medicare Payment Policy, Washington, DC: Medicare Payment Issues, Washington, DC: MedPAC (Mar. 2001). MedPAC (2001). 30. Moy, E, Valente, E Jr, Levin, RJ, Griner, PF, 18. Liska, D, Bruen, B, Salganicoff, A, Long, P, “Academic Medical Centers and the Care of Kessler, B, “Medicaid Expenditures and Ben- Underserved Populations,” Academic Medi- efi ciaries: National and State Profi les and cine, 71(12): 1370–1377 (1996). Trends, 3rd Ed” Report of the Kaiser Commis- 31. Weissman et al., supra, n.28. sion on the Future of Medicaid (1997). 32. Fox, PD, “An Overview of Managed Care,” 19. MedPAC, supra, n.17. In Kongstvedt, PR (Ed.), The Managed Health 20. Guo, K, “Roles of Managers in Academic Care Handbook, 4th Ed., Aspen: Georgia, Health Centers: Strategies for the Managed MD, 3–16 (2001). Care Environment,” Health Care Manager, 33. Levey, S, Anderson, L, Howell, RE, “Pain- 20(3): 43–58 (2002). ful Medicine: Managed Care and the Fate of 21. Alexander, B, Davis, I, Kohler, PO, “Changing America’s Major Teaching Hospitals,” Journal Structure to Improve Function: One Academic of Healthcare Management, 44(4): 231–249 Health Center’s Experience,” Academic Medi- (July/Aug. 1999). cine, 72(4): 259–68 (1997). 34. Guo, supra, n.6. 22. Reuter, JA, “The Financing of Academic Health 35. Gottlieb, S, “Academic Medical Centers: Lean Centers: A Chart Book.” Institute for Health Times Ahead,” USA Today, 125(2624): 64–66 Care Research and Policy, Georgetown Uni- (1997). versity, Washington, DC (1997). 36. Rodgers, J, Lutz, S, “Cost of Caring: Key Driv- 23. Id. ers of Growth in Spending on Hospital Care,” 24. Linna, M, Hakkinen, U, “Reimbursing for report prepared for the American Hospital the Costs of Teaching and Research in Finn- Association and the Federation of American ish Hospitals: A Stochastic Frontier Analysis,” Hospitals, Washington, DC: PriceWater- International Journal of Health Care Finance house-Coopers (2003). and Economics, 6: 83–97 (2006). 37. Id. 25. GAO, “Health Insurance: Characteristics and 38. Fisher, ES, Wennberg, DE, Stukel, TA, Gott- Trends in the Uninsured Population,” Wash- lieb, DJ, “Variations in the Longitudinal Effi - ington, DC (2001); Mills, R, “Current Popu- ciency of Academic Medical Centers,” Health lation Reports: Health Insurance Coverage,” Affairs, 19–32 (2004). US Department of Commerce, Washington, 39. Id. DC (2000). 40. Anonymous, “Teaching Hospitals’ Operating 26. Bellandi, D, “Hospital Prices Jump,” Modern Margins Drop,” Healthcare Financial Man- Healthcare, 12 (Feb 21, 2000). agement, 56(12): 24–25 (Dec 2002). A Comparative Analysis: Nonprofi t Teaching and For-Profi t Non-Teaching Hospitals 37

41. Id. Training,” Minnesota Medicine, 80: 19–22 42. Younis, supra, n.14; Younis and Forgione, (1997). supra, n.15. 57. Lehrer et al., supra, n.55. 43. “Report to the Congress: Selected Medicare Pay- 58. Pardes, H, “The Future of Medical Schools ment Issues,” Washington, DC: MedPAC (2001) and Teaching Hospitals in the Era of Man- DC(2000). aged Care,” Academic Medicine, 72: 97–102 44. Anonymous, “Hard Times for Teaching Hospi- (1997). tals,” Trustee, 53(5): 4 (May 2000). 59. Grosskopf, S, Margaritis, D, Valdmanis, V, 45. Vancil, DR, Shroyer, LW, “Creative Payment “Competitive Effects on Teaching Hospitals.” Strategy Helps Ensure a Future for Teaching European Journal of Operational Research, Hospitals,” Healthcare Financial Manage- 154: 515–525 (2004). ment, 52(11): 48–52 (Nov. 1998). 60. Rosko, MD, “Performance of Major Teach- 46. Foster, MJ, “On Estimating the Costs of the ing Hospitals During the 1990s: Adapting Products of Teaching Hospitals: A Review of the to Turbulent Times,” Journal of Health Care Literature,” Financial Accountability & Manage- Finance, 30(3): 34–48 (Spring 2004). ment, Oxford: 3(2): 175–195 (Summer 1987). 61. Grosskopf et al., supra, n.59. 47. Rosko, M, Carpenter, C, “The Impact of Intra- 62. Yafchak, R, “A Longitudinal Study of Econo- DRG Severity of Illness on Hospital Profi tabil- mies of Scale in the Hospital Industry,” Jour- ity: Implications for Payment Reform,” Journal nal of Health Care Finance, 27(1): 67–89 (Fall of Health Politics, Policy and Law, 19: 729– 2000). 751 (Winter 1994). 63. Rosko, supra, n.60. 48. Thorpe, K, “The Use of Regression Analysis 64. Carr, J, Feldstein, P, “The Relationship of Cost to Determine Hospital Payment: The Case of to Hospital Size,” Inquiry, 4: 45–65 (June Medicare’s Indirect Teaching Adjustment,” 1967). Inquiry, 25: 219–231 (1988). 65. Lave, J, Lave, L, “Hospital Cost Functions.” 49. Koenig, L, Dobson, A, Ho, S, Siegel, JM, et al., The American Economic Review, 379–394 “Estimating the Mission-Related Costs of (June 1970). Teaching Hospitals,” Health Affairs, 22(6): 66. Bays, C, “Specifi cation Error in the Estimation 112–124 (Nov./Dec. 2003). of Hospital Cost Functions,” The Review of 50. Rosko, supra, n.16. Economics and Statistics, 302–305 (1979). 51. Banks, DA, Paterson, M, Wendel, J, “Uncom- 67. Yafchak, supra, n.62. pensated Hospital Care: Charitable Mission 68. Rosko, M, “Understanding Variations in Hos- or Profi table Business Decision?” Health Eco- pital Costs: An Economics Perspective,” Annals nomics, 6(2): 133–143 (1997). of Operations Research, 67: 102–132 (1996). 52. O’Dougherty, SM, Cotterill, PG., Phillips, S, 69. Friedman, B, Pauly, M, “Cost Functions for a Ser- Richter, E, “Medicare Prospective Payment vice Firm with Variable Quality and Stochastic Without Separate Urban and Rural Rates,” Demand: The Case of Hospitals,” The Review of Health Care Financing Review, 14(2): 31–47 Economics and Statistics, 620–624 (1981). (Winter 1992). 70. Dranove, D, “What Do Multi-Hospital Sys- 53. Thorpe, KE, “Why Are Urban Hospital Costs tems Accomplish,” Strategic Management So High? The Relative Importance Patient Journal, 55–74 (1995). Source of Admission, Teaching, Competition, 71. Connor, R, Feldman, R, Dowd, B, “The Effects and Case Mix,” Health Services Research, of Market Concentration and Horizontal 22(6): 821–836 (Feb. 1988). Mergers on Hospital Costs and Prices,” Inter- 54. O’Dougherty et al., supra, n.52. national Journal of the Economics of Business, 55. Lehrer, L, Burgess, J, “Teaching and Hospital 5: 159–180 (1998). Production: The Use of Regression Estimates.” 72. Rosko, supra, n.68. Health Economics, 4: 113–125 (1995). 73. Rosko, M, “Performance of Academic Medi- 56. Kralewski, J, Kephart, K, Hakanson, S, “Untan- cal Centers during the 1990s,” paper pre- gling the Costs of Family Practice Residency sented at the 2002 American Public Health 38 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Association Annual Meeting, Philadelphia, Journal of Accounting Research, 13: 257–266 PA (Nov. 12, 2002). (Autumn 1975). 74. Langabeer, supra, n.2. 84. Cheung, JK, Heaney, J, “A Contingent-Claim 75. Id. Integration of Cost-Volume-Profi t Analysis with 76. Id. Capital Budgeting,” Contemporary Accounting 77. Guidry, F, Horrigan, JO, Craycraft, C, “CVP Research, 6: 738–760 (Spring 1990). Analysis: A New Look,” Journal of Managerial 85. Chan, YL, Yuan, Y, “Dealing with Fuzziness Issues, 10(1): 74–85 (Spring 1998). in Cost-Volume-Profi t Analysis,” Accounting and Business Research, 20(78): 83–95 (1990). 78. Horngren, CT, George, F, and Srikant, MD, 86. Yafchak, supra, n.62. Cost Accounting: A Managerial Emphasis, 8th 87. Schneider, A, “Cost-Volume-Profi t Models Ed., Englewood Cliffs, NJ: Prentice-Hall (1994). Containing Earnings-Based Bonus Expenses,” 79. Id. Accounting Enquiries, 2(August): 168–190 80. Langabeer, supra, n.2. (1992); Schneider, A, “How to Include Earn- 81. Id. ings-Based Bonuses in Cost-Volume-Profi t 82. Guidry et al., supra, n.77. Analysis,” Journal of Managerial Issues, 2: 83. Magee, RP, “Cost-Volume-Profi t Analysis; 231–240 ((Summer 1994). Uncertainty and Capital Market Equilibrium,” 88. Guidry, et al., supra, n.77. The Economics of Health Care Quality and Medical Errors

Charles Andel, Stephen L. Davidow, Mark Hollander, and David A. Moreno

Hospitals have been looking for ways to improve quality and operational effi ciency and cut costs for nearly three decades, using a variety of quality improvement strategies. However, based on recent reports, approximately 200,000 Americans die from preventable medical errors including facility-acquired conditions and millions may experience errors. In 2008, medical errors cost the United States $19.5 billion. About 87 percent or $17 billion were directly associated with additional medical cost, including: ancillary services, prescription drug services, and inpatient and outpatient care, according to a study sponsored by the Society for Actuaries and conducted by Milliman in 2010. Additional costs of $1.4 billion were attributed to increased mortality rates with $1.1 billion or 10 million days of lost productivity from missed work based on short-term disability claims. The authors estimate that the economic impact is much higher, perhaps nearly $1 trillion annually when quality-adjusted life years (QALYs) are applied to those that die. Using the Institute of Medicine’s (IOM) estimate of 98,000 deaths due to preventable medical errors annually in its 1998 report, To Err Is Human, and an average of ten lost years of life at $75,000 to $100,000 per year, there is a loss of $73.5 billion to $98 billion in QALYs for those deaths—conservatively. These numbers are much greater than those we cite from studies that explore the direct costs of medical errors. And if the estimate of a recent Health Affairs article is correct—preventable death being ten times the IOM estimate— the cost is $735 billion to $980 billion. Quality care is less expensive care. It is better, more effi cient, and by defi nition, less wasteful. It is the right care, at the right time, every time. It should mean that far fewer patients are harmed or injured. Obviously, quality care is not being delivered consistently throughout US hospitals. Whatever the measure, poor quality is costing payers and society a great deal. However, health care leaders and professionals are focusing on quality and patient safety in ways they never have before because the economics of quality have changed substantially. Key words: medical errors, quality, patient safety, quality-adjusted life year, QALY, Joint Commission, Institute of Medicine, Society of Actuaries, Milli- man, effi ciency, Medicare, accountable care organizations (ACOs), facility-acquired condition, cost savings.

ospitals have been looking for ways Charles Andel, MBA, BSRT, is Manager of Radiology to improve quality and operational Quality and Compliance at Loyola University Medical Heffi ciency and cut costs for nearly Center and a graduate of Loyola University Chicago’s three decades, using a variety of quality Quinlan School of Business MBA in Healthcare Manage- improvement strategies. The Joint Commis- ment program. He may be reached at [email protected]. sion implemented its Agenda for Change Stephen L. Davidow, MBA, APR, is a health care market- in 1986 to improve the systems, processes, ing and communications professional with a strong focus and, ultimately, the outcomes of care. How- on quality and patient safety. He is a graduate of Loyola University Chicago’s Quinlan School of Business MBA in ever, there has not been widespread adop- Healthcare Management program. He may be reached at tion of these principles, in part because the [email protected] or 708-284-2300. incentives were not substantial enough to Mark Hollander, MBA, is a fi nancial manager at the overcome the inertia of many hospital cul- Department of Veterans Affairs and graduate of Loyola tures and the US payment system. How- University Chicago’s Quinlan School of Business MBA ever, those hospitals and health systems in Healthcare Management program. that overcame that inertia have experienced David A. Moreno, MBA, is a biotechnology business ana- tremendous improvement in quality, fi nan- lyst and graduate of the Loyola University Chicago’s Quin- lan School of Business MBA in Healthcare Management cial performance, patient safety, and patient program. He may be reached at [email protected]. satisfaction. In a recent interview on PBS’s Nightly Business Report,1 Dr. Mark Chas- J Health Care Finance 2012; 39(1):39–50 sin, The Joint Commission’s president and Copyright © 2012 CCH Incorporated 39 40 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

CEO, said that only about a quarter of the extension taxpayers, billions of dollars— nation’s 6,000 hospitals are involved in (1) preventable readmissions and (2) health some sort of quality improvement effort. care facility–acquired conditions, such as Preventable medical harm has been an infections. Before we discuss the new incen- ongoing and vexing problem. Quality and tive system, let’s explore how widespread patient safety expert Dr. Lucian Leape medical errors are and how much they cost. from the Harvard School of Public Health, estimated more than 25 years ago that the problem’s human toll equaled 300 jumbo How Big a Problem Is Quality jets crashing every year in the United States. and Patient Safety? That’s nearly one a day. If that were the case, In 1999, the US Institute of Medicine the US aviation industry would have been (IOM) issued its landmark report, To Err shut down until a solution was found. Yet, Is Human, which stated that up to 98,000 interestingly, the Medicare program over all Americans died as a result of preventable these years reimbursed hospitals regardless medical errors in US hospitals (see Figure 1), of outcome. In fact, there are ICD-9 billing and up to one million more patients experi- codes for specifi c errors. enced some type of preventable error.2 An In a perversion of the fee-for-service sys- error is defi ned as an act that produces a pre- tem, hospitals fared better fi nancially when ventable adverse outcome compared to the patients needed follow-up care after an error natural progression of disease that leads to occurred. A hospital was encouraged by injury or death.3 the payment system to harm a patient just The Centers for Disease Control and Pre- enough without killing him or her and per- vention more recently noted that another form some additional services, for which 100,000 Americans die from infections.4 A it received additional payments. Not much quarter of Medicare benefi ciaries admitted incentive to improve care or save Medicare to a hospital are victims of medical harm, money. Of course, no hospital’s leader- according to a December 2010 report from ship consciously decided to hurt patients to the Offi ce of the Inspector General (and make more money, but the system did not that’s only patients age 65 and above or those encourage and reward better and more effi - on disability). Approximately 5,000 ben- cient care. efi ciaries per month suffer a “never event,” Recent national health reform legislation and 180,000 die from medical errors annu- (the Patient Protection and Accountable Care ally.5 Newer studies from Health Affairs in Act or PPACA) has several quality improve- April 2011 suggest that the rate of prevent- ment provisions including restructuring the able harm may be up to ten times higher than way health care is delivered in the United IOM estimates.6 Although 12 years have States through accountable care organiza- passed since the IOM report, experts are still tions (ACOs) and value-based purchasing. having a diffi cult time developing a concrete The Centers for Medicare & Medicaid Ser- picture of the problem but clearly the toll is vices (CMS) has for the fi rst time said it will high in terms of death, injury, and loss. stop reimbursing hospitals for two major An even greater challenge may be estimat- problems that cost the government, and by ing the economic impact poor quality and The Economics of Health Care Quality and Medical Errors 41

Figure 1. Leading Causes of Death in United States

1. Heart Disease 599,413 2. Cancer 567,628 3. Chronic Lower Respiratory Disease 137,353 4. Stroke 128,842 5. Accidents (Unintentional Injuries) 118,021 Preventable Medical Harm (Medical Errors) 98,000 6. Alzheimer’s Disease 79,003 7. Diabetes 68,705 8. Infl uenza/Pneumonia 53,692 9. Nephritis/Nephrosis 48,935 10. Intentional Self-Harm (Suicide) 36,909

Sources: CDC Web site, FastStats: Leading Causes of Death (Jan. 2012); National Vital Statistics Report, Deaths Final Data for 2009, vol. 60, no. 3; and for 98,000 statistic, IOM Report, To Err Is Human (1998).

unsafe care has in the United States because level? What is the value of a human life and there are so many factors involved—loss of that person’s relationship with his or her life or functionality, lost wages, impact on family members and, more broadly, in rela- family and dependents, law suits, ineffi cient tion to the patient’s community? Ask those and wasteful care as a result of poor facility family members and the answer is incalcu- operations, etc. lable. Although diffi cult to measure because Much of the national discussion about the value of an individual life is not exact, quality and patient safety focuses on the we have applied an economic approach direct medical costs associated with poor using quality-adjusted life years (QALYs) in care. The studies that we explore in this an attempt to develop one answer. Based on article do just that and that seems to be the the IOM fi gure of 98,000 deaths each year bulk of the leading literature on the subject. with an estimate of ten lost years of life at However, there is a signifi cant human cost $75,000 to $100,000 per year, there is a loss for the loss of human life or the impact it of $73.5 billion to $98 billion in QALYs for has on patients who are injured and must those deaths—conservatively. These numbers live with disability for the remainder of their are much greater than those we cite from lives. The focus of the health reform legisla- studies that explore the direct costs of medi- tion is on cost savings to the government by cal errors. And if the estimate of a recent improving care. Of course, the side benefi t is Health Affairs article is correct—preventable fewer harmed patients. death being ten times the IOM estimate—the It is easy to forget when reviewing study cost is $735 billion to $980 billion.7 after study, that what we are talking about are The various estimates on medical errors patients—real people—and their families. point out several failures of the US health care What does poor quality care cost on a human system. The recently enacted health reform 42 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

legislation has many provisions to improve entire US health care system may be too the quality and effi ciency of care provided diffi cult—and certainly not an estimate for to Medicare benefi ciaries.8 In this article, which one can budget. If PPACA is success- we explore several studies and estimates of ful in providing health insurance and access the economic impact preventable medical to care to 32 million more Americans, invar- errors have on the US health care system. iably there will be a numerical increase in Although there is no meaningful estimate as medical errors if nothing is done to improve to how much cost savings can be achieved quality of care. through better care under changes outlined in the PPACA, we will discuss the signifi cance The Economics of Medical Errors of these studies and some anecdotal exam- In 2008, medical errors cost the United ples of hospitals or health systems that have States $19.5 billion. About 87 percent or $17 improved their care and experienced signifi - billion were directly associated with addi- cant cost savings. In this article we explore tional medical cost, including: ancillary ser- different ways improved quality may help vices, prescription drug services, and inpatient reduce costs of the US health care delivery and outpatient care, according to a study spon- system. Quality improvement is a major fac- sored by the Society for Actuaries and con- tor in the PPACA legislation’s efforts to reign ducted by Milliman in 2010. Additional costs in costs in federal expenditures for health care. of $1.4 billion were attributed to increased We also explore the incentives and disincen- mortality rates with $1.1 billion or ten million tives for better quality care and what behavior days of lost productivity from missed work changes among health care providers can be based on short-term disability claims.11 expected, both for facilities and individual The report analyzed claims data to extrap- professionals, to achieve better quality care. olate an estimated 6.3 million medical Under PPACA, hospitals and other provid- injuries. Of those it is believed that, conserv- ers that deliver poor or substandard care will atively, 1.5 million medical injuries were no longer be able participate in the Medicare preventable errors. The study concluded that and Medicaid programs.9 Medicare is a sub- the most signifi cant errors were easily pre- stantial source of income for nearly every ventable if better policies and practices were hospital, nursing home, and physician, so the followed. Opportunity savings of 19.5 bil- impact could be signifi cant. lion would be available. We did not fi nd specifi c analyses of the For medical errors to be in the top ten proposed quality provisions’ economic causes of death we must refl ect on the value impact in the Congressional Budget Offi ce we put on life, as medical errors are easily (CBO) scoring for the PPACA legislation,10 preventable and caused by simple negligence. or in analyses by the Kaiser Family Foun- Milliman also reviewed two previous major dation, the Alliance for Health Reform, nor studies, which attempted to estimate the eco- the Heritage Foundation. No one questions nomic impact of medical errors. The fi rst was the fact that health care can be made better the Harvard Medical Practice Study,12 which in terms of operational effi ciency and higher estimated that all types of medical injuries quality. However, developing a meaning- totaled approximately $3.8 billion in New ful estimate for expected savings across the York in 1984, $50 billion nationally. The Economics of Health Care Quality and Medical Errors 43

The second study, Costs of Medical Inju- An additional 21 percent feel that a moder- ries in Utah and Colorado,13 reviewed a ate negative impact has occurred. Those sur- representative sample of 14,732 randomly veyed conveyed concern over medication selected discharges from 1992 and estimated safety, new purchasing procedures, equip- total costs for errors to be $662 million in ment lifespan, facility maintenance, prop- 1996 dollars—$308 million of that was erly qualifi ed staff, and staff shortages.16 related to preventable medical errors. Extrap- Those surveyed stated that medication olated nationally, that is approximately $37.6 safety was still the number one problem billion for all medical errors and $17 billion and explained its causes. First the elimina- for preventable errors. That study’s authors tion or reduction in time spent by key safety categorize medical errors into fi ve areas: personnel such as medication safety offi cers was reported by over 42 percent surveyed, 1. Operative; coupled with 33 percent reporting less clini- 2. Drug-related; cal pharmacist involvement in patient care 3. Diagnostic or therapeutic; units specifi cally. Intensive care units and a 4. Procedure-related; and new level of risk adverse behavior starts to 5. Other. appear. The reduction has also affected allot- ted time for nurse education, a key area of Postoperative complications were the most concern as more facilities use more part-time expensive, accounting for 35 percent of costs or registry nurses whose integration into the for medical errors and 39 percent of costs for health institution medication administration preventable medical errors. There are many procedures is not adequate. This has led to ways to measure errors and the economic drug administration short cuts and missed impact. The National Quality Forum and safety steps all increasing the risk of harm or National Priorities Partnership talks about the death to patients. $21 billion cost of medication errors.14 Cit- New purchasing procedures refer to ing the New England Healthcare Institute, the purchasing of multi-dose medications inpatient preventable medication errors cost instead of single dose vials and syringes. approximately $16.4 billion, while outpatient Yes, Sam’s Club and Costco have their place medication errors cost $4.2 billion, according in the medical market as medical institu- to the Center of Information Technology.15 tions look to cut cost with bulk medication. This has increased the number of medication The Economy’s Impact on errors fi vefold. The opposite effect, medica- Quality and Patient Safety tion shortages, is occurring as facilities try Although there is no defi nitive evidence to limit supply cost of fast-expiring medica- that the recent recession is having an impact tions that are normally expensive and not on health care quality and patient safety, often utilized. it may be having an affect. A recent study Next is the investment in medical equipment of over 800 nurses, administrators, and and technology. Life cycles for equipment physicians revealed that 20 percent believe are being extended as institutions try to limit a large impact on patient and staff safety has or cancel plans to purchase expensive new occurred due to the recession and its fall out. technology. This becomes one of the biggest 44 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

problems as stagnation in purchasing new will have a signifi cant impact on its ability technology slows or stops both development to survive. and research. Hospital profi ts are not adequate for re-invest in the technology areas that are not covered under PPACA. New Quality: Solutions hospitals, remodeling, and design improve- Quality and patient safety have histori- ments have come to a halt as credit and cash cally been a secondary issue for the majority fl ow have slowed. America’s rural and inner of the nation’s approximately 6,000 hospi- city hospitals are antiquated at best and are tals. Of course, there have been visionary a point of contention as they are not suited leaders who have seen quality, operational for best care practices and are in a state of excellence, and patient care as inextricably ill repair. linked. They provide examples for the rest of the country. In 1986, The Joint Com- mission launched its Agenda for Change Designing for Quality initiative to introduce quality improvement Stepping back from health reform’s spe- philosophy into its accreditation process. cifi c focus on preventable readmissions It was also a way to encourage US hospi- and facility-acquired conditions, quality is tals to adopt these principles in order to a much broader fi eld and incorporates the improve health care operations, quality, and safe design of medical facilities. According ultimately, patient safety. The PPACA legis- to the Agency for Healthcare Research and lation and the fi nancial penalties associated Quality, there is a correlation between how with poor care makes quality “job one” to a hospital is designed and quality of care borrow a slogan from Ford Motor Company and outcomes. New health care construc- in the 1990s. tion over the next ten years is expected to Besides providing better care, improved be $250 billion.17 Evidence-based design quality under the legislation is expected and incorporating it into hospital best prac- to be a major force in efforts to “bend the tices (Lean and Six Sigma)18 is expected to cost curve” for Medicare, as well as private reduce medical waste, improve quality out- insurance. Given the size of the Medicare comes, reduce medical errors, and improve program, the focus on quality will have a patient and employee satisfaction while signifi cant ripple effect throughout the instilling a culture of safety. Improvements entire US health system. But how much include noise suppression, additional light- money is at stake? ing (sunlight when possible), nature areas, Rather than try to determine an exact fi g- and music. The additional $12 million19 ure, we will review a number of reports that in upgrades per facility are expected to be explore the economic impact of quality health recovered within 12 months through opera- care and discuss the work and achievement tional savings and increased revenue. Given of three leading hospitals and health systems the state of the economy and ongoing staff that have implemented signifi cant quality shortages, information that provides trans- improvement efforts. Two institutions have parency about hospital performance and how received the prestigious Malcolm Baldrige it is linked to a hospital’s reimbursement Award for Quality with only ten other US The Economics of Health Care Quality and Medical Errors 45

hospitals the fi rst being SSM Healthcare in patients and assume the responsibility for St. Louis in 2002.20 patients from hospitals that do not meet the standard. This situation has the potential to alter market leadership quickly. Incentives and Penalties for Quality Care Brad Bowman, Director of the Health Historically, the Medicare program paid Care Advisory Practice at Pricewaterhouse- for whatever services health care provid- Coopers suggested that detailed quality ers charged, including provider errors. That performance data that identifi es a poor per- has been changing in recent years and in an forming hospital in a three or four hospital effort to improve quality of care and reduce town, would put the hospital at a signifi - costs to the Medicare program, CMS will no cant disadvantage especially when the local longer reimburse providers for preventable media publicizes the information.22 hospital readmissions. The initial focus will By 2017, up to 6 percent of hospital be on heart attack, heart failure, and pneumo- diagnosis-related group payments will be nia. Hospitals will see their reimbursement at-risk based on quality performance meas- rates go down for high rates of readmissions. ures. Voluntary quality reporting will begin Fines or penalties will start at one percent in 2011, something entirely new for physi- and reach 3 percent over the next three years. cians who accept Medicare patients. Physi- Recognizing that hospitals need to do a bet- cians will receive one percent bonuses going ter job of reducing hospital-acquired condi- down to 0.5 percent by 2014. By 2015, there tions such as infection, the government will will be a 1.5 percent penalty and by 2016, it fi ne hospitals with the highest rates, one per- will be 2 percent. And for the fi rst time, start- cent. Given the low average operating mar- ing in 2015, physicians will see mandatory gin of 5 percent for US hospitals,21 those that individual performance reports published on provide poor quality care will have diffi culty the CMS Web site, as has been done for hos- staying in business. In fact, those that do not pitals and nursing homes.23 improve and meet the national requirements Although the government’s plans seem will lose their ability to care for Medicare logical, the question remains, will the pen- patients. To turn up the heat a bit, Medicare alty/incentive approach work? The approach has said it will publicize which hospitals is not based on a demonstration project or are performing well and those that are not. past experience. However, there is evidence Depending on what actually happens, poor that at least physicians are responsive when quality and the government’s active public their personal compensation is tied to per- notifi cation could damage institutional repu- formance. In a study conducted at the Uni- tations and be the dominant force in shifting versity of Illinois at Chicago that used a market position or leadership. four-year sample of 59 physicians and 1.1 Although Medicare does not reimburse million encounters, a network of primary at the same level as most private insurance, care clinics shifted from salary to a com- it is a signifi cant portion of most hospitals’ pensation plan with a lower salary and business, so a loss of it would be disastrous. piece rates for encounters and procedures. Those hospitals that provide quality care Physicians increased the number of patient will continue to be eligible to treat Medicare encounters from 11 to 61 percent. They were 46 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

paid between $22 and $30 for each outpa- sector delivery reform that is working. tient visit. The authors also noted that physi- Twenty models are included in the statute cians of all medical specialties increased the but it allows for unlimited possibilities. To number of procedures done to one per visit— support these types of initiative, $10 bil- reimbursed at $5 per procedure. The authors lion is provided over ten years. Under the found that physicians respond strongly to Center, a new federal Coordinated Health even to marginal incentives when it is tied to Care Offi ce has been established to help their overall compensation.24 Whether CMS improve the coordination of care for benefi - penalties will motivate health care providers ciaries who are eligible for both Medicare to deliver better quality and less expensive and Medicaid (known as dual-eligible).26 care remains to be seen. The study suggests it doesn’t take much to change certain physi- Leading Examples of Hospitals and cian practice patterns. Health System Improving Quality In a study published in Medical Care and Review, Blue Cross and Blue Shield Intermountain Healthcare of Michigan implemented a pay-for- Since 1988, Intermountain Healthcare, performance program composed of just over based in Salt Lake City, Utah, has been $22 million in incentive payments to hos- applying quality improvement techniques pitals plus a 5 percent administrative cost. to health care delivery that were developed Almost 25,000 patients had better care and by W. Edwards Deming at the end of World had from 733 to 1,701 QALYs depending on War II, and adopted throughout Japanese the effectiveness of care provided. Based on industry. the study results, a QALY was estimated to Brent James, Director of Intermountain be $12,967 to $30,081, much less than most Healthcare’s Institute for Health Care Deliv- generally accepted estimates.25 ery Research, wrote in Health Affairs that Clearly, incentives can have an impact on quality improvement methods were applied how physicians and hospitals deliver care to reduce rates of elective induced labor, and what care they provide. Will the penal- unplanned caesarean sections, and admis- ties be the right motivation to improve health sions to newborn intensive care units. He care? There may be other proven ways to estimated that the initiative saved $50 million improve the care that can be replicated dollars annually. Nationally, it would save around the country. $3.5 billion.27 Another initiative focused on improving Center for Medicare and Medicaid the operation of mechanical ventilators that Innovation were used in treating acute respiratory dis- Given that the PPACA legislation’s goal tress syndrome. As a result, they improved is to reduce cost and improve effi ciency adherence to the guideline and reduced vari- and quality, the Center for Medicare and ation from 59 percent to 6 percent within Medicaid Innovation was established in four months. Patient survival increased from 2011 to test innovative payment and service 9.5 percent to 44 percent. Physician time delivery models. The Center will fund pro- involved in care dropped in half and the total jects at a local level building upon private cost of care dropped by 25 percent. The Economics of Health Care Quality and Medical Errors 47

In 1995, Intermountain analyzed its cost Another important measure of quality is savings from 65 such initiatives and found ICU ventilator-associated pneumonia. In 2006 $30 million—approximately 2 percent of they had no cases. In 2007 and 2008 they had its total clinical operations. These interven- one case per 1,000, which translated to one tions were only applied in one local practice. case per year. In 2009 and 2010 they had no Intermountain estimated that if this initiative cases. Using data from Thomson-Reuters, the was applied to the entire health system, there hospital dropped their actual/expected mor- would be total savings of $100 million to tality ratio signifi cantly. $150 million, approximately 6 to 10 percent Although already below what was expected of annual clinical costs. As a result of these for a hospital of their type based on acuity fi ndings, Intermountain developed a strate- (Level-1 trauma center) in 2003 (0.74), it gic plan to apply these methods throughout dropped to 0.25 in 2010. That translates to the health system. more than 550 patients living than would In a separate White Paper, Intermountain have otherwise have died. The last statistic Healthcare’s CEO, Charles Sorenson, MD, shared was an 83 percent decrease in medi- notes that the United States could reduce cal liability insurance expense. A specifi c national health care spending by as much dollar amount was not shared during the as 40 percent if Intermountain’s operational presentation or in an interview, but the sav- and clinical processes were used as a bench- ings were reported to be in the millions. mark and adopted nationwide.28 This is based on research by Dartmouth’s Paul Wennberg who focused on variation in medical care Poudre Valley Health System practices nationally. At the same meeting, Priscila Nuwash, President of the Center for Performance Excellence at Poudre Valley Health System Advocate Good Samaritan Hospital (PVHS), Fort Collins, Texas, discussed their Downers Grove, Illinois–based Advocate quality efforts, which also were recognized Good Samaritan Hospital is nationally rec- by a Baldrige Award.30 ognized for the quality of its operations and Again, while not specifi c on fi nancial sav- care. It won the Malcolm Baldrige Award for ings, two of PVHS’s hospitals are doing better Quality in 2010. Tamara Schaefer, Director of than competitors on HCAHPS data (federal Patient Safety, recently presented at the 2011 patient satisfaction data). In other PVHS Chicagoland Patient Safety Summit29 and dis- patient satisfaction data examining customer- cussed the importance of a culture of patient focused outcomes, PVHS had nine years of safety on improving care. Although she did consistent improvement. Also, staff voluntary not discuss the hospital’s performance on spe- turnover rates dropped from 19 percent to cifi c Partnership for Patients goals, she made approximately 7 percent from 2000 to 2010. a good case for improving quality. From There was a slight up tick the last year with February of 2009 to April 2011, the hospital the opening of a new hospital. increased hand washing from 30 percent to Another interesting sign of PVHS’s im- almost 100 percent—one of the major ways proved quality is the 59 percent increase in to reduce facility-acquired infections. health system discharges from 2000 to 2010 48 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

compared to 15 percent for their competi- settlements are made when appropriate to tion. PVHS experienced only a 19 percent help the patients and their families begin increase in population growth in their home the healing process. These analyses are also county compared to 40 percent growth in used to change systems and the way pro- their competitor’s home county. Obviously, cedures are done in order to prevent recur- a number of factors can affect these num- rences. Besides being the right thing to do bers but the increase in discharges is attrib- and helping those affected, a byproduct of uted, in part, to the improved quality of this approach is that UIC’s medical liability services. PVHS also uses Thomson-Reuters insurance costs have dropped 53 percent. data and they fall in the nation’s top 10 per- This approach is being evaluated further cent of hospitals for risk-adjusted mortality. through a $3 million Agency for Healthcare An interesting measure was their fi nancial Research and Quality demonstration pro- fl exibility index from Ingenix. It examines ject administered by UIC’s Institute among total margin, return on investment, replace- ten Chicago-area hospitals. Some legal and ment viability, equity fi nancings, days of patient safety experts think this approach cash on hand, cash fl ow to total debt, and may be an alternative to tort reform, which average age of plant. Although the trend has been struck down twice in Illinois by the line shows some ups and downs, from 2001 state’s Supreme Court.31 to 2010, they improved from approximately 7 percent to over 12 percent while locally and nationally there have been decreases. Conclusion

Quality care is less expensive care. It is University of Illinois at Chicago better, more effi cient, and by defi nition, less Medical Center wasteful. It is the right care, at the right time, Historically, medical professionals have every time. It should mean that far fewer been uncomfortable with the idea of admit- patients are harmed or injured. Obviously, ting that a mistake has been made or harm quality care is not being delivered consistently has been caused from an action they have throughout US hospitals. Although recently taken. They fear increased medical liability enacted health reform legislation does not costs, giving attorneys ammunition, censure, require hospitals to implement comprehen- and public rebuke. Counter to this tradi- sive quality improvement and patient safety tional way of thinking is the University of programs, incorporating operational quality Illinois at Chicago (UIC) Medical Center. improvement programs involving Lean or It has focused on developing a culture of Six Sigma have had a signifi cant affect at patient safety. In fact, through its Institute Intermountain Healthcare, Advocate Good for Patient Safety Excellence, its philosophy Samaritan Hospital, Poudre Valley Health has become an international model. Errors System, and the University of Illinois at are quickly identifi ed, disclosed to patients Chicago Medical Center. Interestingly, or their families, root-cause analyses are developing a culture of safety and quality conducted, and the results are shared with also can improve medical liability insur- those who have been affected, and fi nancial ance costs although that’s not the primary The Economics of Health Care Quality and Medical Errors 49

motivator for improving care. The Society says, then that cost could be $735 billion of Actuaries/Milliman report showed that to $980 billion—almost $1 trillion annu- medical errors cost the United States $19.5 ally. Whatever the measure, poor quality billion in direct medical costs. Other studies is costing payers and society a great deal. show the cost to be much higher. Looking at Time will tell if the Medicare program’s the totality of a human life, our own conserv- incentives will make the difference. How- ative calculation shows that medical errors ever, health care leaders and professionals cost $73.5 billion to $98 billion in QALYs are focusing on quality and patient safety and if applied to the most recent estimate in in ways they never have before because Health Affairs that says preventable medi- the economics of quality have changed cal harm is ten times what the IOM report substantially.

REFERENCES

1. Oct. 14, 2011. May Be Ten Times Greater Than Previously 2. For 98,000 statistic, To Err Is Human: Building Measured,” Health Affairs, Project Hope, a Better Health System, Institute of Medicine Bethesda, MD (Apr. 2011). (IOM), National Academy Press, Washington, 7. Id. DC (1999); CDC Web site, FastStats: Leading 8. The Staff of The Washington Post, “Landmark: Causes of Death, http://www.cdc.gov/nchs/ The Inside Story of America’s New Health fastats/lcod.htm/, last updated: Jan. 27, 2012, Care Law and What It Means for Us All,” accessed Aug. 30, 2012; National Vital Statis- Public Affairs Reports (2010). tics Report, Deaths Final Data for 2009, vol. 60, 9. Id. no. 3, http://www.cdc.gov/nchs/data/nvsr/ 10. Elmendorf, DW, CBO’s Analysis of Major nvsr60/nvsr60_03.pdf. Health Care Legislation Enacted in March 3. Shreve, J, et al., The Economic Measurement 2010, Congressional Budget Offi ce, State- of Medical Errors, sponsored by Society of ment before the subcommittee on Health, Actuaries Health Section, prepared by Milli- Committee on Energy and Commerce, US man Inc., Schaumburg, IL (June 2010). House of Representatives, Washington, DC 4. Klevens, RM, et al., “Estimating Health-Care (Mar. 30, 2011). Associated Infections and Deaths in U.S. 11. Shreve, supra, n.3. Hospitals, 2002: U.S. Pubic Health Service,” 12. Brennan, TA, et al., “Incidence of Adverse Public Health Reports, Association of Schools Events and Negligence in Hospitalized of Pubic Health, Washington, DC (Mar.-Apr. Patients: Results from the Harvard Medical 2007), http://www.cdc.gov/ncidod/dhqp/ Practice Study I,” New England Journal of pdf/hicpac/infections_deaths.pdf. Medicine, 324:370–376 (1991). 5. Levinson, DR, “Adverse Events in Hospitals: 13. Thomas, EJ, “Costs of Medical Injuries in Utah National Incidence Among Medicare Benefi - and Colorado,” Inquiry, 36: 255–264 (Fall 1999). ciaries,” Offi ce of Inspector General, Depart- 14. “Preventing Medication Errors: A $21 Billion ment of Health and Human Services (Nov. Opportunity,” Compact Action Brief: A Road- 2010), http://oig.hhs.gov/oei/reports/oei-06- map for Increasing Value in Health Care, 09-00090.pdf. National Priorities Partnership, convened by 6. Classen, DC, et al., “‘Global Trigger Tool’ the National Quality Forum, Washington, DC Shows That Adverse Events in Hospitals (Dec. 2010). 50 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

15. Id. 25. Nahra, TA, “Effectiveness of Hospital Pay-for- 16. “ISMP Survey: Economy Having Negative Performance Incentives, Medical Care and Impact on Medication Safety,” Oncology Times, Review, Sage Publications: Thousand Oaks, CA. 32(3) (Feb. 10, 2010). The online version of this article can be found at: 17. “Transforming Hospitals: Designing For Safety http://mcr.sagepub.com/content/63/1_suppl/ and Quality U.S. Department of Health and 49S DOI: 10.1177/1077558705283629 Human Services, Agency for Health Care (Jan. 18, 2006, online supplement Feb. 2006). Research and Quality,” www.ahrq.gov, AHRQ 26. Landmark, supra, n.8. Pub. No. 07-0076-1 (Sept. 2007). 27. James, BC, Savitz, LA, “How Intermountain 18. SSMHC First Health-Care Recipient of 2002 Trimmed Health Care Costs Through Robust Malcolm Baldrige National Quality Award. Quality Improvement Efforts,” Health Affairs, SSM Healthcare Web site news release, (2011): doi:10.1377/hlthaff.2011.0358–http:// accessed Oct. 22, 2011, update unknown, content.healthaffairs.org/content/ http://www.ssmhc.com/internet/home/ early/2011/05/17/hltaff.2011.0358.full.html. ssmcorp.nsf/6ca2af2859f73db98625710800 28. Sorenson, CW, “Healthcare in the U.S. and 6dce97/671952c33f1a497f862573ec003a5 Utah: A Clinician’s Perspective,” An Inter- 76b?OpenDocument. mountain Healthcare white paper, Salt Lake 19. Thomson-Reuters, Hospital Operating Trends City, Utah (2011). Quarterly, Aug. 2011, accessed Oct. 22, 2011, 29. Schaeffer, T, Presentation: Lifesaving Leadership: update unknown http:// thomsonreuters. Using a Leadership System to Create a Culture com/content/healthcare/pdf/articles/hosp_ of Patient Safety, Chicagoland Patient Safety oper_trends_quarterly_aug_2011. Summit, held at UIC Forum (Sept. 15, 2011). 20. Id. 30. Nuwash, P, Presentation: Innovation and 21. Kavilanz, P, “US to Hospitals: Clean Up Excellence in Service: Is Pursuing Safety and Your Act, CNNMoney Web site, posted Quality the Pathway to Baldrige? Chica- Apr. 29, 2010, accessed October 22, 2010, goland Patient Safety Summit, held at UIC http://money.cnn.com/2010/04/29/news/ Forum (Sept. 15, 2011). economy/healthreform_hospital_ﬁ nes/. 31. McDonald, T, Presentation: Protecting Patients 22. Landmark, supra, n.8. and Supporting Staff: The Seven Pillars Dem- 23. Id. onstration Project, Chicagoland Patient Safety 24. Shreve, supra, n.3. Summit, held at UIC Forum (Sept. 15, 2011). Complicated Billing Requirements Challenge Physical Therapy Industry, Creating Inefﬁ ciencies and Confusion

Annette R. Ciavarella

This article is designed to explain the subtle differences between the reimbursement requirements for coverage of physical therapy services in physician-based settings under the Medicare Benefi t Policy Manual Chapter 15—Covered Medical and Other Health Services.1 These billing challenges have a profound fi nancial impact on the physical therapy industry.2 This article includes: (1) a general background of the reasons surrounding the increased regulations in the physical therapy industry; (2) general defi nitions within the physical therapy industry; (3) a discussion of the confusing and complicated billing requirements for physical therapy services; (4) a discussion of the “incident to” billing requirements within the physical therapy billing requirements; (5) an explanation of differing rules or policies within the physical therapy billing requirements; and (6) a discussion of why these rules regarding physical therapy billing requirements are essential to the delivery of quality of care within the physical therapy industry. Key words: physical therapy, billing requirements.

General Background of the Reasons Offi ce of Counsel to the Inspector General. The Surrounding the Increased Regulations Work Plan includes projects planned in each in the Physical Therapy Industry of the department’s major entities and projects related to issues among all departmental pro- The Offi ce of Inspector General (OIG) has grams.”9 With the continuation of improper been closely scrutinizing payments for phys- billing procedures within the large physi- ical therapy services since 2002 when it dis- cal therapy industry (see Appendices A covered improper billing.3 The OIG’s reports and B), therapy services, and review of indicated more than 90 percent of claims payments were once again reviewed on the submitted to the Medicare program during OIG fi scal year 2011 Work Plan,10 and are the fi rst six months of 2002, failed to meet once more slated for review on the 2012 Medicare’s requirements for payments when Work Plan11 (see Appendix C). therapy services were provided in a physician’s offi ce, resulting in $136 million worth Defi nitions of improper payments paid by Medicare 4 (see Figure 1). Although “fraud” requires Qualifi ed Professional the intent to deceive or misrepresent,5 abuse only involves receipt of payment for items There are several defi nitions within the or services for which there is no entitlement Medicare Benefi t Policy Manual that warrant to that payment.6 This discovery of improper discussion. Under the Manual, a “qualifi ed payments was the catalyst for the subsequent professional” means a “physical therapist increased regulations for physical therapy 7 services. Annette R. Ciavarella, LLM, JD, CFE, CHC, is 8 The OIG uses a “Work Plan” to address President and CEO of Healthcare Compliance various projects “during the fi scal year by the Professionals, LLC. Offi ce of Audit Services, Offi ce of Evaluation J Health Care Finance 2012; 39(1):51–78 and Inspections, Offi ce of Investigations, and Copyright © 2012 CCH Incorporated 51 52 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 1. Improper Payments for Physical Therapy Billed by Physicians

Sample Projected Allowed Services Allowed Amount Type of Error Services Amount (Percent) (Millions) Not medically necessary 18 $455.65 26% $33.0 Undocumented: —Nonresponse 14 466.58 * * —Missing documentation 9 209.56 * * Total undocumented 23 $676.14 34% $49.0 Incomplete/No plan of care: —Incomplete plan of care 23 $802.40 34% $58.2 —No plan of care 16 $397.50 24% $28.8 Total incomplete/no plan of care 39 $1,199.90 57% $87.0 Overlapping errors (Both not medically (18) ($455.65)> <26%> <$33.0> necessary and incomplete/no plan of care) Total 62 $1,876.04 91% $136.0

Source: Medical Review of Physical Therapy Billed by Physicians January to June 2002. *Indicates the n for that cell is too small to reliably project. Totals may not equal the sum of individual rows due to rounding, in HHS OIG semi-annual report on fraud and abuse, May 1, 2006, http://oig.hhs.gov/oei/reports/ oei-09-02-00200.pdf.

… [and] may also include physical therapist and consequently cannot provide billable assistants (PTAs) … when working under the therapy services to Medicare.17 If the PTA ‘supervision’ of a qualifi ed therapist, within is not properly supervised, and claims are the scope of practice allowed by state law.”12 submitted for reimbursement to any federal (See Appendix D for State Practice Acts.) health care program, these claims would be This rule establishes the supervision require- improper and any monies paid by the federal ments for a PTA to be reimbursed for services health care program, would need to be re- by Medicare.13 According to Medicare’s def- paid to that program, and could also include initions, a PTA only meets the requirements civil monetary penalties (CMPs).18 of a “qualifi ed professional,” thereby allow- ing those services to be properly billed to and Covered Physical Therapy Services reimbursed by Medicare, when that therapist is “supervised” by a “qualifi ed therapist.”14 Medicare’s defi nitions dictate what physi- The supervision by the physical therapist cal therapy services are in fact covered and must also take place in the same room as the therefore reimbursable.19 The Manual defi nes PTA.15 PTAs must be supervised because covered physical therapy services as skilled they do not perform initial evaluations of rehabilitative services, which are medically the patients or develop treatment plans.16 necessary and reasonable and necessary skilled Therefore, if a PTA is not being supervised, procedures provided by a qualifi ed profes- that PTA is not a “qualifi ed professional” sional within the scope of his or her practice.20 Complicated Billing Requirements Challenge Physical Therapy Industry 53

A Tennessee physical therapy organization 2. Services that can be furnished by PTAs was held to have violated the federal False without a supervising therapist are not Claims Act for improperly billing Medicare rehabilitative and therefore not reason- for physical therapy services which were not able or necessary.29 medically necessary, and was ordered to pay $1.8 million.21 On May 2, 2012, a nationwide The Medicare Benefi t Policy Manual takedown by Medicare Fraud Strike force is quite clear: a PTA must be supervised.30 operations22 resulted in charges against 107 If a PTA is unsupervised, then those ser- individuals, including physical therapists.23 vices are considered non-rehabilitative and Their alleged participation in Medicare fraud thereby not reimbursable under Medicare.31 schemes involved approximately $452 million Moreover, the type, or level, of supervision in false billing.24 These individuals allegedly required for covered services under Medi- participated in schemes to submit claims that care depends on the following: were either medically unnecessary or never provided.25 These indictments should “remind 1. The location of the PTA; and those criminals that they risk prosecution and 2. The employment arrangement,32 which prison time every time they submit a false is discussed later in this article. claim.”26 Medicare explains, “Rehabilitative therapy Billing Confusion occurs when the skills of a therapist … are necessary to safely and effectively furnish Establishing Billing Privileges a recognized therapy service.”27 Medicare To bill Medicare directly, Medicare adds the following emphasis: requires every physical therapist (PT) to Services that can be safely and effec- enroll as a “private practitioner.”33 The tively furnished by … PTAs … with- PT is then referred to as a therapist in out the supervision of therapists are not private practice (TPP).34 To qualify to rehabilitative therapy services. If at any bill Medicare directly as a TPP, every PT point in the treatment of an illness it must also be employed in one of the fol- is determined that the treatment is not lowing practice categories: rehabilitative, or does not legitimately require the services of a qualifi ed pro- an unincorporated solo practice, unin- fessional … the services will no longer corporated partnership, unincorpo- wbe considered reasonable and neces- rated group practice, physician … sary. Services that are not reasonable or group or groups that are not profes- necessary should be excluded from cov- sional corporations [PCs], if allowed 28 erage under §1862(a)(1) of the Act. by state and local law.35 Physician These points reinforce the following … group practices may employ TPP facts: if state and local law permits this employee relationship. For purposes 1. Unsupervised PTAs are not qualifi ed of this provision, a physician…group professionals; and practice is defi ned as one or more 54 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

physicians … enrolled with Medicare more than the employment arrangement, who may bill as one entity.”36 and are insuffi cient research to determine proper billing procedures.45 To clarify, if PTs want to obtain and retain The OIG inquiries associated with Medi- billing privileges with Medicare to bill directly care payments for physical therapy direct for their services, they must fi rst: that more research is needed than query- ing peers.46 Comparisons with peers require 1. Complete the CMS-855-I application;37 specifi c details, including the site of that and employment arrangement as well as the type 2. Choose the “private practitioner” option of employment arrangement; specifi c super- on the application when enrolling.38 vision requirements must be investigated.47

The therapist will then be issued a National Supervision Requirements Provider Identifi er (NPI), the unique identifi cation number for health care providers, As detailed above, if the therapist is which enables them to bill Medicare directly employed in a physician group practice, and for their services.39 The therapist must also wants to bill Medicare directly for services be employed in one of the above practice performed, then that therapist must enroll types, all of which have different supervi- with Medicare as a TPP.48 However, if the sion requirements,40 and are discussed later TTP practice does not enroll as a Medi- in this article. Therefore, if the therapist was care provider, then that therapist may not employed by an incorporated41 physician bill Medicare directly, and must then be group practice, or partnership,42 or an institu- directly supervised (“direct supervision” is tional setting (e.g., skilled nursing facility), discussed later in this article) by an enrolled then that therapist would not be permitted to therapist (the same requirements for a PTA, directly bill Medicare (see Figure 2).43 as noted earlier).49 “These direct supervi- This is why merely comparing billing sion requirements apply only in the private practices and supervision procedures with practice setting and only for therapists and peers is insuffi cient.44 Comparisons with their assistants.”50 Supervision requirements peers require specifi c details, which include differ within other sites of service, and are

Figure 2. Employment Billing Categories

Employment Category Having the Employment Category NOT Having Ability to Bill Medicare Directly the Ability to Bill Medicare Directly Unincorporated solo practice Incorporated physician group practice Unincorporated partnership Incorporated physician partnership Unincorporated group practice Incorporated group(s) Physician group(s) that are not PCs, if permitted Hospital by state and local law Skilled nursing facility Complicated Billing Requirements Challenge Physical Therapy Industry 55

often at the core of the confusion, and ulti- state practice requirements are more strin- mately, civil monetary penalties (CMPs).51 gent, in which case those requirements The OIG, through the Secretary of Health must be followed.57 General supervision and Human Services (HHS), is authorized to requires the physician’s direction and con- seek CMPs, as well as, exclusion from partic- trol over the services performed.58 Although ipation in all federal health care programs.52 it is not necessary for the physician to be There was a very recent False and Fraudu- present during the treatment, that physician lent Claims CMPs case with the OIG.53 This remains ultimately responsible for the treat- case involved a physician from Mississippi ment as well as for the maintenance of any who agreed to settle his case for $25,500 equipment.59 “Direct supervision” requires for allegedly submitting improper claims to the presence of the physician in the actual Medicare by failing to directly supervise the ofﬁ ce building, though not necessarily in physical therapy services.54 Another recent the treatment room, and immediately avail- case came out of Minnesota.55 The organiza- able to render assistance during treatment if tion in this case had been submitting false necessary.60 Presence of the physician in the Medicare and Medicaid claims for over treatment room is required under “personal three years due to either: supervision.”61 See Figure 3.62 The supervision requirements in Figure 1. Unsupervised physical therapy ser- 3 are related to a therapist wishing to bill vices; or Medicare directly for his or her services, 2. Physical services that were not per- using the therapist’s own NPI. Adding to formed at all.56 the already confusing billing requirement, Medicare has additional billing require- “General supervision,” is required for ments related to a therapist wishing to bill PTAs in all settings except private practice Medicare using the employer’s NPI, rather (which requires direct supervision), unless than the therapist’s own NPI.63

Figure 3. Supervision Requirements

Supervision Site of Service Requirement Physical Presence Requirement Outpatient hospital General PT presence not required on premises Nursing facility General PT presence not required on premises Comprehensive outpatient General PT presence not required on premises rehabilitation facilities Outpatient rehabilitation facilities General PT presence not required on premises Home health agencies General PT presence not required on premises PT in private practice Personal PT present in the same room Physician office (“Incident to”) Direct Physician present in the office suite Physician office Personal PT present in the same room 56 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

“Incident To” Requirements provided incident to services of a physical therapist.”73 “Incident to” rules were developed by PTAs may not bill incident to either the Medicare to reimburse non-physician provider physical therapist’s service or the physician’s services that were “incident to” a physician’s service.74 However, if a physical therapist service.64 When a physical therapist who has and PTA are both employed by a physician acquired his or her own NPI provides services group, then the services of the supervised in the physician’s group practice where that PTA may be billed by the physician group therapist is employed, but that therapist bills as physical therapy services (using the physi- using the physician’s group practice NPI for cal therapist’s NPI), if the physical therapist the services provided by the therapist, rather is enrolled with Medicare.75 It is necessary than using the therapist’s own NPI, then all to learn and understand these “incident to” enrolled therapists must bill “incident to” the requirements because they are mandatory enrolled physician’s services, though this is billing requirements for reimbursement under not the typical billing practice.65 In this situ- Medicare guidelines as well as for some com- ation, the rules for both physical therapy ser- mercial payers.76 vices as well as the confusing “incident to” Medicare’s “incident to” billing procedure billing must be followed.66 is often required by private third-party pay- The OIG reports, mentioned earlier, deter- ers (e.g., managed care organizations, private mined that a portion of the improper pay- insurance carriers) that often do not credential ments they discovered was due to physicians physical therapists as providers in their group improperly billing under the complicated contracts.77 Many third-party payers are also and confusing “incident to” rule.67 Inaccu- Medicare Parts C and D payers (a program rate “incident to” billing is one reason physi- through which organizations contract with the cal therapy services are on OIG’s 2012 Work Centers for Medicare & Medicaid Services Plan.68 The OIG is also investigating whether (CMS) to provide coverage of health care ser- “incident to” billing is “vulnerable to overu- vices to Medicare benefi ciaries78) and as such tilization [which consequently exposes] “must follow Medicare guidelines”,79 includ- Medicare benefi ciaries to care that does not ing “incident to” requirements, regardless of meet professional standards of quality.”69 whether that particular patient is a Medicare To qualify for reimbursement under benefi ciary.80 The “incident to” billing proce- Medicare’s “incident to” rules, the supervis- dure is quite complex with several steps.81 ing physician must be on the premises on the A non-physician provider is billing inci- day and time of services, immediately avail- dent to a physician-provider’s services.82 able to render assistance if needed, and be “Incident to” services are defi ned as those documented as such in the patient record.70 services that are furnished incident to a phy- Auditors and investigators will compare both sician’s professional services in the phy- the physicians’ and therapists’ schedules to sician’s offi ce, which must meet certain determine billing integrity.71 As previously requirements as follows:83 mentioned, “incident to” services only apply to billing “incident to” the services of phy- • Services must be part of the patient’s sicians.72 “There is no coverage for services normal course of treatment, during which Complicated Billing Requirements Challenge Physical Therapy Industry 57

a physician personally performed an —The caregiver providing the service initial service and remains actively must represent a direct fi nancial involved in the course of treatment.84 expense to the physician-provider • The physician does not need to be phys- (e.g., a W-2 or leased employee, an ically present in the patient’s treatment independent contractor); and room while these services are provided, —The caregiver providing the service but the physician must provide direct must be directly supervised by the supervision, i.e., the physician must physician-provider. be present in the offi ce suite to render ° The physician directly supervis- immediate assistance, if necessary.85 ing the auxiliary personnel need — “The availability of the physician not be the same physician upon by telephone … does not constitute whose professional service the direct supervision.”86 “incident to” service is based. • Documentation should include the es- ° If the physician is in a group prac- sential requirements for “incident to” tice, any physician of the group service. may be present in the offi ce to • The physician is not required to see the supervise (see specifi cs immedi- patient each time the non-physician ately below).87 practitioner treats the patient. • There must be subsequent services by Medicare rules for “incident to” a physi- the physician of a frequency that refl ects cian’s service provided within a physician the physician’s continuing active par- group practice are generally the same as the ticipation in and management of the rules listed above, with a few differences:88 course of treatment. • These services must be all of the fol- 1. A physician (or a number of physi- lowing: cians) is physically present in the facil- —An integral part of the patient’s treat- ity to perform medical (rather than ment course; administrative) services at all times the —Commonly rendered without charge [group] is open; or included in the physician’s bill 2. Each patient is under the care of a [i.e., relatively insignifi cant supplies [group] physician; and and services that are not separately 3. The non-physician services are under payable by Medicare; or, under other medical supervision.89 circumstances, the supply or service would already be included in the Therefore, the physician ordering a particu- physician’s bill]; lar service need not be the physician who is —Of a type commonly furnished in a supervising the service in a group practice.90 physician’s offi ce or clinic (clinic In a group practice, direct supervision by a includes a physician group practice physician may be the responsibility of several but not an institutional setting); physicians on the premises.91 Thus, medical —Provided by a caregiver qualifi ed to management of all services provided in the provide the service; group practice is assured.92 Hence, services 58 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

performed by non-physician personnel are the therapist is billing for his or her services protected even though those services are per- as “incident to” one of the group physician’s formed in another department of the group services, the claim form will be completed practice.93 The “direct supervision” require- differently.106 ments for a group practice are no different.94 For those practices that have implemented Direct supervision of a physical therapist electronic medical records, how to properly by a physician in an offi ce setting requires electronically document “incident to” ser- the physician to “be present in the offi ce vices would be as follows: suite.”95 Although Medicare does not specifi - cally defi ne what a “suite” is, it has indicated • The physician who ordered the physical that a “suite” means the same offi ce building therapy service should have his or her where a physician is “immediately available information included on every line of to furnish assistance and direction through- service in the 2420E loop; out the performance of the procedure if so • The supervising physician’s informa- needed.”96 These requirements do not oblige tion should be included in loop 2310E; the physician to be present in the same room • The “rendering” provider would be the while the procedure or service is performed.97 treating physical therapist and his or her The physician need only be in the same information should be included in loop offi ce building.98 Direct supervision should 2420D.107 not be confused with Medicare’s “general” or “personal supervision” requirements.99 The supervising physician will now be Under “general supervision,” the physician identifi ed on the CMS 1500 claim form in is not even required to be in the building.100 the lower left corner of the page, and his Even if physicians successfully comprehend or her NPI will be indicated on that same the previously listed “incident to” require- form in the center of the page. The physical ments (which the OIG report indicates other- therapist would not be listed on the form at wise101), they may still encounter confusion all. However, the supervising physical ther- when submitting the actual claims.102 apist is still documented in the electronic medical record by following the procedures above. Dealing with Claim Forms A therapist with his or her own Medicare NPI, employed in a physician group, typi- Differing Rules or Policies cally bill services as his or her own, using the therapist’s NPI, with the therapist iden- Confusion Between Reimbursement tifi ed on the claim.103 Physician and other Differences qualifi ed professional provider services are There is further confusion when it comes fi led on the CMS 1500 claim form.104 The to reimbursement rates for physical therapy therapists are then identifi ed on the CMS (PT) services. PT services are reimbursed at 1500 claim form in the lower left corner of 100 percent of the applicable physician fee the page, and their NPI is indicated on that schedule,108 (up to the $1,860 therapy cap form in the center of the page.105 However, if established in 2010109). This reimbursement Complicated Billing Requirements Challenge Physical Therapy Industry 59

is the same as nurse practitioners (NPs) only between TPPs and NPs are just a few of the when their services are billed under the phy- sources contributing to the confusion sur- sician’s NPI as “incident to” that physician’s rounding billing requirements for physical services.110 The NP is reimbursed only at 85 therapy services. When confl icting messages percent of the fee schedule amount when are sent by different organizations, this leads those services are billed under the NP’s to even more confusion. NPI and not as “incident to” a physician’s 111 service. This percentage differential does American Physical Therapy not exist for physical therapists.112 PTs are Association always reimbursed at 100 percent of the phy- Medicare indicates that the American sician fee schedule.113 Physical Therapy Association (APTA) Signature Requirement Differences guidelines may be utilized; however, they may be used as guidelines only, and “not as There is another signifi cant difference policy.”121 Medicare goes on to clarify that between billing requirements for TPPs whenever a differing policy or law exists that and NPs regarding progress note signature is “more stringent” than Medicare rules, it is requirements. Rehabilitative services must the more stringent policy or rule that must be include a written plan of treatment from the followed.122 Medicare repeatedly mentions provider in the progress report.114 The plan that the “more stringent” policy or law must of care must include the type, frequency, be met,123 indicating that whenever there is and duration of the services, as well as the a harsher policy or law, that “more strin- diagnosis and goals the provider hopes to gent” regulation must be followed, regard- accomplish.115 The progress report provides less of whether the stricter rule is a state justifi cation for the medical necessity of law or the Medicare regulation.124 Some of this plan of treatment.116 The determination the confusion surrounding the supervision of the necessity of the services provided is requirements originates directly with the based on the documentation in the treatment presentation of the rules in the Medicare notes and progress report.117 The signature Benefi t Policy Manual. indicates the ownership of that documenta- As noted earlier, a physical therapist tion (i.e., the medical decision-making).118 employed within a group medical practice For Medicare payment purposes, informa- is a therapist in private practice.125 This can tion required in progress reports shall be be very confusing to physical therapists, written by a clinician who either provides or especially if seeking assistance from the supervises the services, or by the therapist individual State Practice Act. The APTA who provides the services and supervises also indicates that the requirements for a PTA.119 Medicare does require the super- supervision and direction of PTAs are based vising physician to sign the progress notes, on several considerations, which include: written by the NP, for services billed “incident to” that physician’s service; however, Medicare does not require the supervising 1. Site of service (group practice, institu- physician to sign the progress reports writ- tional, etc.); ten by the TPP.120 The above differences 2. The use of personnel; 60 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

3. The education, experience, and skill of Rules: Essential to Quality of Care the PTA; 4. The patient’s criticality, acuity, stabil- The above rules are indeed quite compli- ity, and complexity; and cated. It is understandable to think that the 5. Federal regulations.126 complication can lead to confusion regarding accurate billing procedures. It may be difﬁ cult PTs must therefore look to three major for providers to appreciate the importance of governmental bodies to discern physical the above rules and regulations. Providers may therapy billing requirements: also be feeling production ﬁ nancial pressures. However, these regulations were designed to 1. APTA; protect “the health of all Americans,”128 i.e., to 2. The State Practice Act; and obtain the highest quality of care for patients. 3. CMS. Quality of care is dependent on patient safety. And patient safety within organizations is However, the safest course is to comply “dependent on people and staff to adhere to with Medicare rules, because governing the appropriate principles and rules in the face bodies can vary.127 of production pressures.”129 Complicated Billing Requirements Challenge Physical Therapy Industry 61

Appendix A. Ratio of PTAs to PTs: By CMS Region (1999) 0.11 0.17 0.25 0.25 0.27 0.31 0.35 Ratio of PTAs to PTs: By CMS Region (1999) Ratio of PTAs 0.38 Total Atlanta Boston Chicago Denver Philadelphia San Francisco Seattle New York 0.27 Urban Institute analysis of 1999 State Occupation Employment and Wage Estimates, US Department of Estimates, Wage and Occupation Employment State of 1999 analysis Urban Institute

PT is physical therapist; PTA is physical therapist assistant. therapist is physical PTA therapist; PT is physical

0.50 0.45 0.40 0.35 0.30 0.25 0.05 0.00 0.20 0.15 0.10 PTA to PT Ratio PT to PTA Source: Labor. Note: 62 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Appendix B. Distribution of PTs and PTAs: By Setting (2001) 4.7% Other PT PTA 5.2% 5.1% schools Colleges/ secondary 10.1% 7.6% Home/ home care 12.0% 7.2% hospital 5.0% Rehabilitation 21.7% facility Skilled nursing 8.7% 17.1% facility Outpatient 19.7% ed settings. Distribution of PTs and PTAs: By Setting (2001) Distribution of PTs and PTAs: Acute hospital (inpatient) 13.8%13.8% 22.8% Private practice 25.5% Urban Institute summary of American Physical Therapy Association fall 2001 survey of members and non- of members survey 2001 Association fall Therapy American Physical summary Urban Institute of PT is physical therapist; PTA is physical therapist assistant; “other” includes health/wellness facilities, industry, industry, assistant; therapist facilities, includes health/wellness is physical “other” PTA therapist; PT is physical

5% 0%

30% 25% 20% 15% 10%

PTAs / PTs of Share research centers, and other unspecifi centers, research Source: respondents). PTA n = 3,413 (n = 4,898 PT respondents; members Note: Complicated Billing Requirements Challenge Physical Therapy Industry 63

Appendix C. OIG 2012 Work Plan

Every year, the Offi ce of Inspector Gen- is in the Social Security Act, § 1861(s) eral (OIG) within the Department of Health (2)(A). Medicare requires providers to and Human Services (DHHS) publishes its furnish such information as may be nec- Work Plan, a description of issues targeted essary to determine the amounts due to for government attention during the upcom- receive payment. (Social Security Act, ing fi scal year. For FY 2012, the OIG has § 1833(e).) (OEI; 00-00-00000; expected multiple issues in its Work Plan that are of issue date: FY 2013; new start) importance to physical therapists. The issues include: • Physicians: Impact of Opting Out of Medicare (New)—Review the extent • Physicians: Incident-To Services to which physicians are opting out of (New)—Review of physician billing Medicare and determine whether phy- for “incident-to” services to determine sicians who have opted out of Medi- whether payment for such services had care are submitting claims to Medicare. a higher error rate than that for non- We will also examine whether specifi c incident-to services. We will also assess areas of the country have seen higher CMS’s ability to monitor services billed numbers of physicians opting out and as “incident-to.” Medicare Part B pays its potential impact on benefi ciaries. for certain services billed by physicians Physicians are permitted to enter into that are performed by non-physicians private contracts with Medicare benefi - incident to a physician offi ce visit. ciaries. (Social Security Act, § 1802(b).) A 2009 OIG review found that when As a result of entering into private con- Medicare allowed physicians’ billings tracts, physicians must commit that they for more than 24 hours of services in a will not submit a claim to Medicare for day, half of the services were not per- any Medicare benefi ciary. (OEI; 07-11- formed by a physician. We also found 00340; expected issue date: FY 2012; that unqualifi ed non-physicians per- work in progress) formed 21 percent of the services that physicians did not perform personally. • Independent Therapists: Outpatient Incident-to services represent a pro- Physical Therapy Services—Review gram vulnerability in that they do not of outpatient physical therapy services appear in claims data and can be iden- provided by independent therapists to tifi ed only by reviewing the medical determine whether they were in com- record. They may also be vulnerable pliance with Medicare reimbursement to overutilization and expose Medicare regulations. Previous OIG work has benefi ciaries to care that does not meet identifi ed claims for therapy services professional standards of quality. Medi- provided by independent physical thera- care’s Part B coverage of services and pists that were not reasonable, medically supplies that are performed incident to necessary, or properly documented. Our the professional services of a physician focus is on independent therapists who 64 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

have a high utilization rate for outpatient Part B payments to physicians and sup- physical therapy services. Medicare will pliers. We will determine whether pay- not pay for items or services that are ment system controls are in place to not “reasonable and necessary.” (Social identify such payments and assess the Security Act, § 1862(a)(1)(A).) Docu- effectiveness of those controls. Medi- mentation requirements for therapy care Part B services must be reason- services are in CMS’s Medicare Benefi t able and necessary (Social Security Act, Policy Manual, Pub. 100-02, Ch. 15, § 1862(a)(1)(A)), adequately documen- § 220.3. (OAS; W-00-11-35220; vari- ted (§ 1833(e)), and provided consist- ous reviews; expected issue date: FY ent with Federal regulations (42 CFR, 2012; new start) § 410). A high cumulative payment is an unusually high payment made to an • Comprehensive Outpatient Reha- individual physician or supplier, or on bilitation Facilities—Review national behalf of an individual benefi ciary, over Medicare utilization patterns for Com- a specifi ed period. Prior OIG work has prehensive Outpatient Rehabilita- shown that unusually high Medicare tion Facility (CORF) services, identify payments may indicate incorrect billing CORFs in high-utilization areas, and or fraud and abuse. (OAS; W-00-12- determine whether they meet basic 35605; various reviews; expected issue Medicare requirements. Medicare paid date: FY 2012; new start) about $61 million for 35,000 benefi - ciaries who received CORF services in • Medicare Payments for Part B 2009. Previous OIG work identifi ed Claims with G Modifi ers—Review CORF services that did not meet Medi- Medicare payments made from 2002 care reimbursement standards because to 2010 for claims on which providers they were not medically necessary or used certain modifi er codes indicat- lacked documentation that they were ing that Medicare denial was expected. provided. OIG has also raised concern We will determine the extent to which about potentially inappropriate rental Medicare paid claims having such mod- arrangements between physician land- ifi ers. We will also identify providers lords and CORFs. Federal regulations and suppliers with atypically high bill- require that CORFs maintain locations ing related to the modifi ers. Providers that provide safe and suffi cient space for may use GA or GZ modifi ers on claims the scope of all services offered (42 CFR they expect Medicare to deny as not rea- § 485.62). We will conduct site visits of sonable and necessary. (CMS’s Claims CORFs. (OEI; 05-10-00090; expected Processing Manual.) They may use GX issue date: FY 2012; work in progress) or GY modifi ers for items or services that are statutorily excluded. A recent • Physicians and Other Suppliers: High OIG review found that Medicare paid Cumulative Part B Payments (New)— for 72 percent of pressure-reducing Review payment systems controls that support surface claims with GA or GZ identify high cumulative Medicare modifi ers, amounting to $4 million in Complicated Billing Requirements Challenge Physical Therapy Industry 65

potentially inappropriate payments. that consistently submitted claims found (OEI; 02-10-00160; expected issue to be in error in a 4-year period. In this date: FY 2012; work in progress) review, we will select the top error- prone providers based on expected dol- • Payments for Services Ordered or lar error amounts and match selected Referred by Excluded Providers— providers against the National Claims Review the nature and extent of Medi- History fi le to determine the total dol- care payments for services ordered or lar amount of claims paid. We will then referred by excluded providers (those conduct a medical review on a sample who have been barred from billing fed- of claims. Providers must submit accu- eral health care programs) and examine rate claims for services provided to CMS’s oversight mechanisms to iden- Medicare benefi ciaries. (CMS’s Medi- tify and prevent payments for such ser- care Claims Processing Manual, Pub. vices. No payments shall be made for 100-04.) (OAS; W-00-11-35565; vari- any items or services furnished, ordered, ous reviews; expected issue date: FY or prescribed by excluded individuals or 2012; new start) entities. (Social Security Act, §§ 1128 and 1156, and 42 CFR § 1001.1901.) • Payments for Physical, Occupational, (OEI; 00-00-00000; expected issue and Speech Therapy Services (Med- date: FY 2013; new start) icaid)—Determine the extent to which payments for Medicaid physical, occu- • Medical Claims Review at Selected pational, and speech therapy services Providers—Review Medicare Part A comply with State standards and limits and Part B claims submitted by error- on coverage. Previous OIG studies found prone providers to determine their that some therapy services provided validity, project our results to each pro- under Medicare were billed incorrectly. vider’s population of claims, and rec- Through a review of selected States, we ommend that CMS request refunds on will determine whether Medicaid has projected overpayments. Previous OIG similar program integrity issues. States work illustrated a methodology for may provide physical, occupational, identifying error-prone providers using and speech therapy services to Medic- CMS’s Comprehensive Error Rate Test- aid benefi ciaries pursuant to the Social ing (CERT) Program data. Using this Security Act, § 1905(a), and regulations methodology, we identifi ed providers at 42 CFR § 440.110.130 66 JOURNAL OF HEALTH CARE FINANCE/FALL 2012 Regulation 24, art. 34, ch. Tit. 5, § 34-24-193 12 4, § 510, Art. 54.510(e) ACC, 1, § art. 19, Ch. AZ State 32-2001, Legislature Bd of Practice § 17-93-101, Act, 1(B)(2) of the § 1398.44 PT Regulations Revised Colorado 12, Tit. Statutes, 41-113(1) art. for (Continuous: of for- graduates eign PT schools: § 20-70a) 376, Ch. § 20-66(3); § 20-70(2)(b)(1) Supervision Requirements for PTAs Requirements Supervision nition Appendix D. State Practice Acts State Practice Appendix D. practicing. The board shall not issue any rules or regulations that require a PTA to be within sight to a PTA that require rules shall not issue any or regulations The board of a consulting PT or supervisor of that PT or while working under the direction with § 34-24-217(b). inconsistent or orders rules, regulations, issue any tel- by be provided that may with the PTA consultation for The PT shall be available or in writing. verbally, ephone, when the via telecommunications available The supervising PT is on call and readily interventions. treatment is providing PTA the to telecommunication or by in person available The supervising PT shall be readily The supervising PT shall provide patients. is treating at all times while the PTA PTA by supervisionperiodic rendered on-site and observation of the assigned patient care the PTA. a licensed physical by of or the participation in the work of a PTA The overseeing communica- of direct availability (A) continuous to: not limited but including, therapist, of a licensed PT on regularly (B) availability and a licensed PT; tion between the PTA in the and (ii) support the PTA of the PTA; the practice (i) review scheduled basis to: situ- emergency plan for and (C) a predetermined services; of the PTA’s performance licensed PT in the absence of including the designation of an alternate ations, licensed PT. regular Supervision Requirement Defi direction of a direction PT registered at least once per month supervision supervision of for- graduates eign PT schools) and imme- Direct supervisiondiate State Alabama Under the AlaskaArizona on-site Periodic Arkansas General supervision Direct Consultation. for The supervising and available PT is on-site California Adequate Coloradosupervision Direct Connecticut such unlicensed individuals are any Supervision where that is on the premises for (Continuous: State Arkansas

Supervision State Requirement Defi nition Regulation Delaware Direct supervision A licensed PT shall be on the premises; and evaluations and progress notes shall be Tit. 24, Del. C. co-signed by the supervising, licensed PT. In relation to a PTA with less than one year of § 2611(a) experience, a PT shall be on the premises at all times and see each patient. In relation to a PTA with one year or more of experience, the supervising PT must see the patient at

least once every sixth treatment day, and the PTA must receive on-site, face-to-face super- Complicated BillingRequirements Challenge Physical Therapy Industry vision at least once every 12th treatment day. The supervising PT must have at least one year of clinical experience. The PT must be available and accessible by telecommunications to the PTA during all of the PTA’s working hours. All supervision must be documented. District of Direct supervision The PT is physically present and immediately available for direction and supervision. Ch. 12, Columbia The PT will have direct contact with the patient/client during each visit that is defi ned as § 3-1212.04(C) all encounters with a patient/client in a 24-hour period. Telecommunications does not meet the requirement of direct supervision. Florida Direct supervision Except in a case of emergency, direct supervision requires the physical presence of the FL statute licensed PT for consultation and direction of the actions of a PT or PTA who is practic- 486.021(9) ing under a temporary permit and who is a candidate for licensure by examination. Georgia Direct supervision PT must be on the premises at all times. § 43-33-13.1 Hawaii Direct supervision The supervisor shall be present in the same building as the person being supervised Ch. 461J-3(e) and available for consultation and assistance. Idaho A licensed PT shall supervise and be responsible for patient care given by PTAs and 54-2218(1) supportive personnel. Illinois General supervi- General: The PT must maintain continual contact with the PTA including periodic per- 225 ILCS 90/2 sion for PTAs sonal supervision and instruction to insure the safety and welfare of the patient. Direct supervision for students Indiana Direct supervision The supervising PT or physician at all times shall be available and under all circumstances § 25-27-1 shall be absolutely responsible for the direction and the actions of the person supervised when services are performed by the PTA or holder of a temporary permit issued under IC 25-27-1-8(d). For the holder of a temporary permit issued under IC 25-27-1-8(d), unless the supervising PT or physician is on the premises to provide constant supervision, the holder of a temporary permit shall meet with the PT or physician at least once each working day to review all patients’ treatments. This meeting must include the actual presence of the PT or physician and the holder of a temporary permit. The patient’s care shall always be the responsibility of the supervising PT or physician. Reports shall be countersigned by the PT or physician who may enter remarks, revision, or additions. With respect to the supervision of PTAs under IC 25-27-1-2(c), unless the supervising PT or physician is on the premises to provide constant supervision, the PTA shall consult with the supervising 67 Continued ... 68 JOURNAL OF HEALTH CARE FINANCE/FALL 2012 .6402 Regulation 65, KS Ch. Statutes 2418F(2)(c) 45, 32, ch. Tit. (2) § 3111 § 13-206, 10, Tit. 04.02 ch. § 73-23-33 (f) 334.506(7); 334.650(2) State Practice Acts State Practice Continued... nition PT or physician at least once each working day to review all patients’ treatments. The treatments. all patients’ review to at least once each working day PT or physician days 14 (1) every each patient not less than: shall examine supervising PT or physician (2) the earlier facility; rehabilitation or comprehensive inpatients in either a hospital for the mentally for patients in a facility visits for therapy or 6 physical 90 days of every and patients; (DD) and school system disabled (MR) and developmentally retarded all other patients; visits for therapy physical 15 or every 30 days (3) the earlier of every is not face-to-face, consultation If this daily and progress. treatment patients’ review to A of 3 full-time PTAs. not supervise than the equivalent may more the PT or physician be in person, may betweenconsultation and the PTA a supervising PT or a physician is the deaf (TDD), so long as there for device a telecommunications or by telephone, by concerning communication patient care. interactive and supervision works under the direction of a licensed PTPTA and imme- of the PT who is on-site acting under the direction a PT or PTA by Oversight the support to personnel. available diately of the patient or PTA. session upon request scheduled treatment the next the patient by of the patient. and welfare the safety ensure supervision to adequate § 148A.6(2) or activities are procedures therapy and instruction when physical aid, direction, give to performed. in regulations as prescribed intervals, a licensed PT at regular by oversight Face-to-face a licensed PTA a patient by to of the services the board, provided by adopted supervision than 4 full-time under his or her direct more No licensed PT shall have PTAs. equivalent ed ed ect supervisionect in the facility. present be physically PT must A-D § 23 112, Ch. Supervision Requirement Defi and supervision supervision supervision supervision State IowaKansas Under direction Kentucky Personal Louisiana Not specifi Maine SupervisingMaryland to and available beeper or telephone by accessible A supervising PT shall be readily Directionsupervision Direct Massachusetts area within the treatment available and immediately present A licensed PT is personally Dir Michigan including periodic on-site a PT with PTA by Minnesota contact and written verbal Continuing Mississippi Not specifi supervision Direct is provided. while treatment intervention Missouri in-person on-site for Direct, Is available on-site Direct, 148 Supervision State Requirement Defi nition Regulation Montana Supervision The licensed PT must make an on-site visit to the client at least once for every 6 37-11-105 visits made by the PTA or once every two weeks, whichever occurs fi rst. The PT may not concurrently supervise more than 2 full-time PTAs, 4 PT aides, or two assistants and two aides. This supervision does not require the presence of the PTA. A student requires on-site supervision. Complicated BillingRequirements Challenge Physical Therapy Industry Nebraska General supervi- General supervision: either on-site or by means of telecommunication. 71-1370; 71-1368 sion: PTA Direct supervision: The PT must be physically present and immediately available and Direct supervi- does not include supervision provided by means of telecommunication. sion: Students Nevada Immediate A person is present and immediately available within the treatment area to give aid, NRS 640.016 supervision direction, and instruction to the person he or she is supervising. New IV. A PTA shall III. Direct personal supervision: the PT or the PTA is physically present and immedi- §§ 328-A:2, III-V; Hampshire work under a ately available to direct and supervise tasks that are related to patient/client manage- A:11, IV-V PT’s general ment. The direction and supervision is continuous throughout the time these tasks supervision. are performed. Telecommunications does not meet the requirement of direct personal A PTA shall supervision. document care IV. “Direct supervision’’ means that the PT is physically present and immediately avail- provided and able for direction and supervision. The PT will have direct contact with the patient/cli- shall report to a ent during each visit that is defi ned as all encounters with a patient/client in a 24-hour supervising PT period. Telecommunications does not meet the requirement of direct supervision. any status in a V. “General supervision’’ means that the PT is not required to be on-site for direction and patient requiring supervision, but must be available at least by telecommunications. a change in the plan of care. The supervising PT shall review and co-sign all notes during each reevaluation. V. A PT aide shall work under the direct personal supervision of a PT or a PTA. 69 Continued ... 70 JOURNAL OF HEALTH CARE FINANCE/FALL 2012 Regulation 13:39A-2.1 16.20.7.8 6741 §§ 6738; 90-270.24 (5); 90-270.31(a); 21 NCAC 48A.0105(8); 21 NCAC 48C.0102(h) §§ art. 61.5, Tit. 04 61.5-05-01-02; State Practice Acts State Practice Continued... nition (D) A PT supervising a temporary licensee must notify the New Mexico PT licensing A PT supervisingMexico notify the New a temporary licensee must (D) supervision for of a tempo- when he or she is no longer responsible in writing, board, rary licensee. the patient. regarding with the PTA conferences hold documented PT must The referring (E) room Same Direct: building Same On-site: regarding a patient’s treatment or a patient’s reaction to the treatment occurring during the treatment to reaction or a patient’s treatment a patient’s regarding procedure. treatment any requiring per week) 80 work hours totaling (full-time equivalency or 2 FTEs time PTAs, and licensed PTAs. temporary PTAs, supervision, including temporary PT, more do not exceed combined FTEs provided supervise PTAs A PT may 2 or more (B) per week. than 80 hours work supervising(vacation, a PT planning an absence from When another licensee, (C) supervise another PT to the for arrange education) must continuing of absence, leave licensee in his or her place. “On-site supervision” means the supervising in the department licensee is present or facil- supervision” “On-site being supervised, the person to available is immediately ity services provided, where are sessions in which students in aspects of treatment involvement continued and maintains in components of care. involved or PT aides are completing clinical requirements or by directly available be immediately means that the PT must supervision” “Immediate supervising a PT aide or student. a PTA to telecommunication for available and immediately on the premises present The PT is physically Direct: with the patient during contact each direct The PT will have and supervision. direction does not meet this requirement. Telecommunications visit. is the services provided, where in the facility are and present The PT is on-site On-site: being supervised, involve- the person continued and maintains to available immediately session in which supportive personnel aspects of each treatment ment in appropriate in components of care. involved are Supervision Requirement Defi PTAs supervision: supervi-On-site Students sion: On-site supervi-On-site PT aides sion: super- Immediate PTAs vision: PT aide On-site PTA supervision: New YorkNew Direct personal State New JerseyNew supervision Direct consequence any to respond to available readily on-site, be present The PT must MexicoNew Supervision and supervision than 2 full- the direction of more for not be responsible A PT may (A) North Carolina North Dakotasupervision: Direct Supervision State Requirement Defi nition Regulation Ohio Supervision: PTA PTA: A PTA may only be supervised by a PT and may not be supervised by §§ 4755-27-04(C) Direct supervi- any other person, including those persons licensed to practice in any other (2); 4755-27- sion: Student profession. 04(D)(2)

Student: The PT is on-site and available to immediately respond to the needs Complicated BillingRequirements Challenge Physical Therapy Industry of the patient. The PT shall have direct contact with the patient during each visit. Oklahoma Direct or general General: The responsible supervision and control of the practice of the licensed 435:20-7-1 supervision: PTA PTA by the supervising PT. The supervising therapist is regularly and routinely Immediate on-site, and every 3 months will provide a minimum of 1 co-treatment of face-to-face, supervision: real-time interaction with each PTA providing services with his or her patients PT aides These co-treatments will be documented in the medical record and on a supervision log, which is subject to inspection. When not on-site, the supervising therapist is on call and readily available physically or through direct telecommunication for consultation. On-site or direct: The PT is continuously on-site and present in the department or facility where services are provided, is immediately available to the person being supervised, and maintains continued involvement in appropriate aspects of each treatment session in which assistive personnel are involved in components of care. Oregon Supervision and (5) Supervise means to provide the amount of personal direction, assistance, advice, Admin. Rules direction: PTA and instruction necessary to reasonably assure that the supervisee provides the Ch. 848; Div. 20; Direct, on-site patient competent PT services, given the supervisor’s actual knowledge of the § 848-020- supervision: supervisee’s ability, training and experiences. Additionally, supervision of: (a) A 0000(5)(a) PT aides treatment-related task requires that the supervising PT or PTA be in the same building and within sight or earshot of the aide who is performing the treatment-related task, such that the supervising PT or PTA is immediately available at all times to provide in person direction, assistance, advice, or instruction to the aide or the patient. A PT may delegate supervision of an aide to a PTA. Pennsylvania Direct, on-prem- The physical presence of a PT on the premises so the PT is immediately available to § 40.1 ise supervision provide supervision, direction, and control. Rhode Island Supervision The PT is at all times responsible for supportive personnel and students. § 5-40-1

Continued ... 71 72 JOURNAL OF HEALTH CARE FINANCE/FALL 2012 SL 1996, ch. 231, ch. SL 1996, § 2, 36-10-18.2 §§ 63-13-103(10); 63-13-103(11) §§ 322.3(b)(2); 322.3(14) §§ 58-24(b)- (8); 102(4), 58-24b-401(2); 58-24b-404 § 26-38-2086(2) Regulation 40-45-20(2) State Practice Acts State Practice Continued... nition PTA: The supervising PT will be readily available to the PTA being supervised. When When being supervised. the PTA to available The supervising PT will be readily PTA: the supervising PT will be immediately setting, in an offsite is practicing the PTA scheduled and will be regularly patient conferences telecommunications; by accessible documented. in the and present on-site (1) be continuously must The PT or PTA Aides/students: available (2) be immediately department services being performed; where or facility are aspects of each treat- in appropriate involvement continued and (3) maintain assist; to the supervisee. by is being performed ment session in which a component of treatment when PT services being are available be on call and readily The PT must PTAs: provided. respond. to available and readily The PT is on the premises Aides: electronic or by telephone, by in person, available be immediately The PT must PTAs: assist the PTA. to communication the where at the facility present be continuously must the therapist Aides/students: and regularly assist the person; to available immediately services; is providing person the person. by in the services being provided involved in an off- is practicing the PTA When the PTA. to available the PT will be readily Aides: telecommunications. by accessible the PT will be immediately setting, site PT must be (1) continuously on site; (2) present in the department (2) present where or facility on site; be (1) continuously PT must assist in the services to being available (3) immediately services being performed; are aspects of each treat- in appropriate involvement continued and (4) maintain performed; the supervisee the patient. ment session where is treating that good and and the PTA the patient, PT, ensure to of PTA and practice records, is rendered. treatment safe Supervision Requirement Defi On-site super- On-site Aides, vision: students supervi-On-site PT aides sion: PTAs sion: super- On-site aides, vision: students PTAs supervi-On-site aides sion: On-site On-site supervision State Tennessee PTA Supervision: TexasUtah PTA Supervision: supervi- General Vermont Supervision: South Carolina South Dakota Supervision the work, and review Supervisiondirect, observe, is the responsibility of PT to Supervision State Requirement Defi nition Regulation Virginia Direct supervi- Direct: A PT or PTA must be physically present and immediately available and is fully §§ 54.1-3479(C); sion: student responsible for the PT services being performed. Cannot supervise more than 3 train- VA Bd of PT131 General supervi- ees at any one time.

sion: PTAs General: A PT shall be available for consultation. Complicated BillingRequirements Challenge Physical Therapy Industry Washington Direct supervi- Direct: The PT must (a) be continuously on-site and present in the department or facility RSW: 18.74.130, sion: students, where assistive personnel or holders of interim permits are performing services; (b) 150, 180 aides, PTAs be immediately available to assist the person being supervised in the services being Indirect supervi- performed; and (c) maintain continued involvement in appropriate aspects of each treat- sion: PTAs ment session in which a component of treatment is delegated to assistive personnel. May only supervise 2 assistants at a time. Indirect: The supervisor is not on the premises, but has given either written or oral instructions for treatment of the patient and the patient has been examined by the PT at such time as acceptable health care practice requires and consistent with the particular delegated health care task. West Virginia General supervi- General: The PT must be available at least by telecommunications. §§ 30.20-2(h); sion: PTAs Direct: The physical presence of the PT in the immediate treatment area where the 16-1-2.2 Direct supervi- treatment is being rendered. sion: Aides Immediate treatment area: the area within the PT’s direct line of sight or within audible distance of the PT and the ability of the PT to immediately respond to calls for assistance from the patient or PT aide. Wisconsin Direct supervi- Direct: PT must provide direct, on-premises supervision on a daily basis; have primary § 448.56(6); § 8. sion: PTAs, aides, responsibility for patient care rendered by supervisee; and be available at all times for PT 5.01 and .02 students consultation. There may not be more than 2 PTAs or 4 aides being supervised at one General supervi- time. sion: PTAs General: There may not be more than four PTAs or four aides being supervised at one time. Wyoming On-site super- PT must be continuously present in the facility where the supervised services are § 33-25-101(a)(9) vision: aides, provided, is immediately available to the person being supervised, and maintains students continued involvement in each treatment session. 73 74 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

REFERENCES

1. http://www.cms.gov/Regulations-and- 13. http://www.aota.org/DocumentVault/ Guidance/Guidance/Manuals/downloads// Surveys/LCD-Outpatient.aspx. bp102c15.pdf. The rules discussed in this 14. http://www.cms.gov/manuals/Downloads/ article also apply to Occupational Therapists bp102c15.pdf, 220 - Coverage of Outpa- and Occupational Therapists Assistants. tient Rehabilitation Therapy Services (Physi- 2. HHS OIG semi-annual report on fraud and cal Therapy, Occupational Therapy, and abuse, May 1, 2006, http://oig.hhs.gov/oei/ Speech-Language Pathology Services) Under reports/oei-09-02-00200.pdf. Medical Insurance (Rev. 88, Issued: 05-07- 3. HHS OIG semi-annual report on fraud and 08, Effective: 01-01-08, Implementation: abuse, May 1, 2006, http://oig.hhs.gov/oei/ 06-09-08), A; http://www.aota.org/Document reports/oei-09-02-00200.pdf. Vault/Surveys/LCD-Outpatient.aspx. 4. HHS OIG semi-annual report on fraud and 15. 42 C.F.R. §§ 410.32(b)(3)(iii), 410.60(c)(2). abuse, May 1, 2006, http://oig.hhs.gov/oei/ 16. Report to Congress: Standards for Supervision reports/oei-09-02-00200.pdf. of PTAs and the Effects of Eliminating the “Per- 5. One has committed fraud if one “know- sonal” PTA Supervision Requirement on the ingly and willfully executes or attempts to Financial caps for Medicare Therapy Services, execute a scheme … to defraud any health- http://www.cms.gov/TherapyServices/ care beneﬁ t program or to obtain by means downloads/61004ptartc.pdf. of false or fraudulent pretenses, representa- 17. http://www.aota.org/DocumentVault/ tions, or promises any of the money or prop- Surveys/LCD-Outpatient.aspx. erty owned by … any healthcare beneﬁ t 18. http://oig.hhs.gov/fraud/enforcement/ program,” Health Insurance Portability and cmp/index.asp. Accountability Act 1996 (18 U.S.C., Ch. 63, 19. http://www.cms.gov/Regulations-and- § 1347). Guidance/Guidance/Manuals/downloads// 6. CMS Glossary, http://www.cms.hhs.gov; bp102c15.pdf. http://www.medicare.gov. 20. http://www.cms.gov/Regulations-and- 7. HHS OIG semi-annual report on fraud and Guidance/Guidance/Manuals/downloads// abuse, May 1, 2006, http://oig.hhs.gov/oei/ bp102c15.pdf; http://www.cms.gov/manuals/ reports/oei-09-02-00200.pdf. Downloads/bp102c15.pdf, Rehabilitative Ther- 8. http://oig.hhs.gov/reports-and-publications/ apy (220.2 - Reasonable and Necessary Outpa- workplan/index.asp#current. tient Rehabilitation Therapy Services), C. 9. http://oig.hhs.gov/reports-and-publications/ 21. “Benchmark Physical Therapy to pay $1.8 workplan/index.asp#current. Million for improper Medicare Billings” 10. OIG Archives, http://oig.hhs.gov/reports-and- (6/30/10), Benchmark violated the federal FCA publications/archives/workplan/index.asp. by submitting claims for PT services which 11. http://oig.hhs.gov/reports-and-publications/ were not medically necessary (http://www. archives/workplan/2012/Work-Plan-2012. chattanoogan.com/2010/6/30/178983/ pdf. Benchmark-Physical-Therapy-To-Pay- 12. http://www.cms.gov/manuals/Downloads/ 1.8.aspx). bp102c15.pdf, 220 - Coverage of Outpa- 22. http://www.justice.gov/opa/pr/2012/ tient Rehabilitation Therapy Services (Physi- May/12-ag-568.html. Cases are being inves- cal Therapy, Occupational Therapy, and tigated by Medicare Strike Force Teams, FBI Speech-Language Pathology Services) Under agents, HHS-OIG and State Medicaid Fraud Medical Insurance (Rev. 88, Issued: 05-07- Control Units. 08, Effective: 01-01-08, Implementation: 23. http://www.justice.gov/opa/pr/2012/May/ 06-09-08), A. 12-ag-568.html. Complicated Billing Requirements Challenge Physical Therapy Industry 75

24. Ibid. bp102c15.pdf (§ 230.4: Services Furnished 25. Ibid. by a Therapist in Private Practice). 26. Ibid, FBI Assistant Attorney General Breuer. 35. Federal law does permit this employment 27. http://www.cms.gov/manuals/Downloads/ arrangement. The Federal self-referral, “Stark” bp102c15.pdf, Rehabilitative Therapy Law, and anti-kickback statutes as well as (220.2 - Reasonable and Necessary Outpa- Medicare payment policies all permit physi- tient Rehabilitation Therapy Services), C. cians to employ and to refer patients to physical 28. http://www.cms.gov/manuals/Downloads/ therapists within their group practice. Federal bp102c15.pdf, 220 - Coverage of Out- laws clearly draw an important distinction patient Rehabilitation Therapy Services between prohibited referrals and legitimate (Physical Therapy, Occupational Therapy, employment relationships. Referrals, § 1877 and Speech-Language Pathology Services) [42 U.S.C. § 1395nn] (b)(2). Additionally, it is Under Medical Insurance (Rev. 88, Issued: important to determine if such an arrangement 05-07-08, Effective: 01-01-08, Implementa- is permitted under individual state law. tion: 06-09-08), A, emphasis added. 36. http://www.cms.gov/manuals/Downloads/ 29. LCD for Outpatient Physical and Occu- bp102c15.pdf (230.4: Services Furnished pational Therapy Services (http://www. by a Therapist in Private Practice). See aota.org/DocumentVault/Surveys/LCD- provider enrollment at www.cms.hhs.gov/ Outpatient.aspx); Comments and Responses MedicareProviderSupEnroll, Pub. 100-08, Regarding Draft Local Coverage Deter- Medicare Program Integrity Manual, chapter mination: Outpatient Physical and Occu- 10, section 12.4.14; American Physical Ther- pational Therapy Services, http://www. apy Association; Culver, L, “Understanding cms.gov/medicare-coverage-database/ APTA Positions on the Use of Personnel,” Mag- lcd_attachments/26884_33/Outpatient_ azine of Physical Therapy, 8(9): 28–33 (2000). Physical_Occupational_Therapy_Services_ 37. Healthcare Payment Systems, An Introduction, Comm_Resp_art_pub_Jul_08.pdf. Duane C. Abbey, p. 83. 30. http://www.cms.gov/Regulations-and- 38. http://www.cms.gov/Regulations-and- Guidance/Guidance/Manuals/downloads// Guidance/Guidance/Manuals/downloads// bp102c15.pdf. bp102c15.pdf (§ 230.4: Services Furnished 31. LCD for Outpatient Physical and Occu- by a Therapist in Private Practice). pational Therapy Services (http://www. 39. http://www.cms.gov/NationalProvIdentStand/. aota.org/DocumentVault/Surveys/LCD- 40. http://www.cms.gov/Regulations-and- Outpatient.aspx); Comments and Responses Guidance/Guidance/Manuals/downloads// Regarding Draft Local Coverage Deter- bp102c15.pdf (§ 230.4: Services Furnished mination: Outpatient Physical and Occu- by a Therapist in Private Practice). pational Therapy Services, http://www. 41. An organization that has ﬁ led articles of cms.gov/medicare-coverage-database/ incorporation with the Secretary of State. lcd_attachments/26884_33/Outpatient_ 42. An organization that has registered their Physical_Occupational_Therapy_Services_ Statement of Partnership Authority with the Comm_Resp_art_pub_Jul_08.pdf. Secretary of State. 32. http://www.cms.gov/Regulations-and- 43. http://www.cms.gov/Regulations-and- Guidance/Guidance/Manuals/downloads// Guidance/Guidance/Manuals/downloads// bp102c15.pdf (§ 230.4: Services Furnished bp102c15.pdf (section 230.4: Services Fur- by a Therapist in Private Practice). nished by a Therapist in Private Practice). 33. http://www.cms.gov/Regulations-and- 44. http://oig.hhs.gov/oei/reports/oei-09-02- Guidance/Guidance/Manuals/downloads// 00200.pdf. bp102c15.pdf (§ 230.4: Services Furnished 45. http://oig.hhs.gov/oei/reports/oei-09-02- by a Therapist in Private Practice). 00200.pdf. 34. http://www.cms.gov/Regulations-and- 46. http://oig.hhs.gov/oei/reports/oei-09-02- Guidance/Guidance/Manuals/downloads// 00200.pdf. 76 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

47. http://oig.hhs.gov/oei/reports/oei-09-02- supervisory visit at least once every 30 days 00200.pdf. or more frequently if required by state or 48. http://www.cms.gov/manuals/Downloads/ local laws or regulation. bp102c15.pdf, 230.4 - Services Furnished 58. 42 C.F.R. Ch. IV (10-1-03 Edition), § 410.32(b) by a Therapist in Private Practice (TPP), (Rev. (3)(i). 106, Issued: 04-24-09, Effective: 07-01-09, 59. Ibid. Implementation: 07-06-09), A. 60. 42 C.F.R. Ch. IV (10-1-03 Edition), § 410.32(b) 49. http://www.cms.gov/manuals/Downloads/ (3)(ii). bp102c15.pdf, 230.4 - Services Furnished 61. 42 C.F.R. Ch. IV (10-1-03 Edition), § 410.32(b) by a Therapist in Private Practice (TPP), (Rev. (3)(iii). 106, Issued: 04-24-09, Effective: 07-01-09, 62. http://www.cms.gov/TherapyServices/ Implementation: 07-06-09), B. downloads/61004ptartc.pdf. 50. Ibid. 63. http://www.cms.gov/manuals/Downloads/ 51. http://www.cms.gov/Regulations-and- bp102c15.pdf, 230.5 – Physical Therapy … Guidance/Guidance/Manuals/downloads// Services Provided Incident to the Services of bp102c15.pdf (section 230.4: Services Fur- Physicians and Non-Physician Practitioners nished by a Therapist in Private Practice). (NPP), (Rev. 36, Issued: 06-24-05, Effective: 52. http://oig.hhs.gov/fraud/enforcement/ 06-06-05, Implementation: 06-06-05). cmp/index.asp. 64. S.S. Act § 1862; 42 U.S.C. 1395y. 53. http://oig.hhs.gov/fraud/enforcement/ 65. http://www.cms.gov/manuals/Downloads/ cmp/false_claims.asp. bp102c15.pdf, 230.5 – Physical Therapy … 54. Ronald V. Myers, Sr., M.D. (Myers), Missis- Services Provided Incident to the Services of sippi, agreed to pay $25,500 for allegedly Physicians and Non-Physician Practitioners violating the Civil Monetary Penalties Law. (NPP), (Rev. 36, Issued: 06-24-05, Effective: The OIG alleged that Myers caused improper 06-06-05, Implementation: 06-06-05). claims to be submitted to Medicare from 66. Ibid. Select Care, Inc. and Primary Physical Medi- 67. HHS OIG semi-annual report on fraud and cine, Inc., for physical therapy and related abuse, May 1, 2006, http://oig.hhs.gov/oei/ health care items or services that Myers did reports/oei-09-02-0200.pdf. not personally render or did not directly 68. “PT in Motion News Now, HHS to Investi- supervise (2-22-12); http://oig.hhs.gov/fraud/ gate ‘Incident To’ Billing in 2012” (Oct. 13, enforcement/cmp/false_claims.asp. 2011), http://www.apta.org/PTinMotion/ 55. Star Tribune, “Whistleblower gets green light NewsNow/2011/10/13/OIGWorkPlan/. to build false-claims case, by Dan Brown- 69. http://www.apta.org/PTinMotion/ ing (2/24/12), http://www.startribune.com/ NewsNow/2011/10/13/OIGWorkPlan/. local/140316793.html. 70. 42 C.F.R. § 410.27(f). 56. Star Tribune, “Whistleblower gets green light 71. This list is a reminder of what to include when to build false-claims case, by Dan Brown- responding to a third party’s request for ther- ing (2/24/12), http://www.startribune.com/ apy records: http://www.palmettogba.com/ local/140316793.html. Palmetto/Providers.nsf/files/Responding_ 57. http://www.cms.gov/Regulations-and- to_an_Request_for_Outpatient_Therapy_ Guidance/Guidance/Manuals/downloads// Records_Checklist.pdf/$FIle/Responding_ bp102c15.pdf (section 230.4: Services Fur- to_an_Request_for_Outpatient_Therapy_ nished by a Therapist in Private Practice); In Records_Checklist.pdf. clinics, rehabilitation agencies, and public 72. http://www.cms.gov/manuals/Downloads/ health agencies, 42 C.F.R. § 485.713 indi- bp102c15.pdf, §1861(s)(2)(A) of the Social cates that when a PTA provides services, Security Act, 42 C.F.R. §§ 410.10(b), 410.26, either on or off the organization’s premises, Pub. 100-02, ch. 15, § 60. those services are supervised by a quali- 73. Ibid. ﬁ ed physical therapist who makes an onsite 74. Ibid. Complicated Billing Requirements Challenge Physical Therapy Industry 77

75. https://www.cms.gov/manuals/Downloads/ 92. CMS Manual System, Pub. 100-02 Medicare bp102c15.pdf, 230.5. Benefi t Policy, http://www.hcca-physician- 76. Gosfi eld, A, “The Ins and Outs of ‘Incident-To’ conference.org/past/2004/302/CMS%20 Reimbursement,” Family Practice Manage- Manual%20System.pdf. ment, (Nov.–Dec. 2001); http://www.aafp.org/ 93. https://www.cms.gov/manuals/ fpm/2001/1100/p23.html. Downloads/bp102c15.pdf, 230.5. 77. Ibid. 94. 42 C.F.R. § 410.32(b)(3)(ii). 78. http://www.medicare.gov/navigation/ 95. 42 C.F.R. § 410.32(b)(3)(ii). medicare-basics/medicare-benefi ts/part-c. 96. 42 C.F.R. § 410.32(b)(3)(ii). aspx?AspxAutoDetectCookieSupport=1. 97. http://www.cms.gov/manuals/Downloads/ 79. http://www.medicare.gov/navigation/ bp102c15.pdf. medicare-basics/medicare-benefi ts/part-c. 98. CMS Manual System, Pub. 100-02 Medicare aspx?AspxAutoDetectCookieSupport=1. Benefi t Policy, http://www.hcca-physician- 80. http://www.medicare.gov/navigation/ conference.org/past/2004/302/CMS%20 medicare-basics/medicare-benefi ts/part-c. Manual%20System.pdf. aspx?AspxAutoDetectCookieSupport=1. 99. “General supervision” means the procedure 81. Medicare Carriers Manual Claims Process, is furnished under the physician’s overall http://www.cms.gov/Regulations-and- direction and control, but the physician’s Guidance/Guidance/Transmittals/ presence is not required during the per- downloads//R1764B3.pdf. formance of the procedure. Under general 82. Medicare Carriers Manual Claims Process, supervision, the training of the non-physi- http://www.cms.gov/Regulations-and- cian personnel who actually perform the Guidance/Guidance/Transmittals/ diagnostic procedure and the maintenance downloads//R1764B3.pdf. of the necessary equipment and supplies are 83. http://www.cms.gov/manuals/Downloads/ the continuing responsibility of the physi- bp102c15.pdf, §§60.1, 60.2 and 60.3. cian (42 C.F.R. § 410.32(B)(3)(i)); and 84. Medicare Carriers Manual Claims Process, “Personal supervision” means a physician http://www.cms.gov/Regulations-and- must be in attendance in the room during Guidance/Guidance/Transmittals/ the performance of the procedure (42 C.F.R. downloads//R1764B3.pdf. § 410.32(b)(3)(iii)). 85. http://www.cms.gov/manuals/Downloads/ 100. 42 C.F.R. § 410.32(B)(3)(i). bp102c15.pdf, §§60.1, 60.2 and 60.3. 101. HHS OIG semi-annual report on fraud and 86. 42 C.F.R. § 410.26. abuse, May 1, 2006, http://oig.hhs.gov/oei/ 87. http://www.cms.gov/Regulations-and- reports/oei-09-02-00200.pdf. Guidance/Guidance/Manuals/downloads// 102. Gosfi eld, supra, n.76. bp102c15.pdf. 103. http://www.cms.gov/Regulations-and- 88. http://www.cms.gov/Regulations-and- Guidance/Guidance/Manuals/downloads// Guidance/Guidance/Manuals/downloads// bp102c15.pdf. bp102c15.pdf. 104. Healthcare Payment Systems, An Introduc- 89. http://www.cms.gov/Regulations-and- tion, Duane C. Abbey, p. 5. Guidance/Guidance/Manuals/downloads// 105. Medicare Benefi t Policy Manual CMS Pub. bp102c15.pdf. 100-2, 60.1. 90. Medicare Carriers Manual Claims Process, 106. CMS Billing and Coding Guidelines, http:// http://www.cms.gov/Regulations-and- www.cms.gov/medicare-coverage-database/ Guidance/Guidance/Transmittals/ lcd_attachments/32007_1/120811_00153_ downloads//R1764B3.pdf. L32007_OPHTH026_CBG_010112.pdf. 91. CMS Manual System, Pub. 100-02 Medicare 107. CMS Billing and Coding Guidelines, http:// Benefi t Policy, http://www.hcca-physician- www.cms.gov/medicare-coverage-database/ conference.org/past/2004/302/CMS%20 lcd_attachments/32007_1/120811_00153_ Manual%20System.pdf. L32007_OPHTH026_CBG_010112.pdf. 78 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

108. CMS Internet Only Manual (IOM) Pub- 119. Ibid. lication 100-04, Chapter 12, Section 120. http://www.cms.gov/manuals/Downloads/ 30.6.1.b, and 100-02 http://www.cms.gov/ bp102c15.pdf, 220.3, D. Regulations-and-Guidance/Guidance/ 121. http://www.cms.gov/manuals/Downloads/ Manuals/downloads//clm104c12.pdf. bp102c15.pdf, 220 - Coverage of Outpa- 109. CHC Medicare Explained Health Law Pro- tient Rehabilitation Therapy Services (Physi- fessional Series, Wolters Kluwer Law & cal Therapy, Occupational Therapy, and Business, p. 143 (2010). Speech-Language Pathology Services) Under 110. CMS Internet Only Manual (IOM) Publication Medical Insurance (Rev. 88, Issued: 05-07- 100-04, Chapter 12, Section 30.6.1.b, and 08, Effective: 01-01-08, Implementation: 100-02 http://www.cms.gov/Regulations-and- 06-09-08). Guidance/Guidance/Manuals/ 122. Ibid. downloads//clm104c12.pdf. 123. http://www.cms.gov/Regulations-and- 111. CMS Internet Only Manual (IOM) Publication Guidance/Guidance/Manuals/downloads// 100-04, Chapter 12, Section 30.6.1.b, and bp102c15.pdf. 100-02 http://www.cms.gov/Regulations- 124. Ibid. and-Guidance/Guidance/Manuals/ 125. http://www.cms.gov/Regulations-and- downloads//clm104c12.pdf. Guidance/Guidance/Manuals/downloads// 112. http://www.cms.gov/Regulations-and- bp102c15.pdf (§ 230.4: Services Furnished Guidance/Guidance/Manuals/downloads// by a Therapist in Private Practice). clm104c12.pdf. 126. Culver, supra, n.36. 113. CMS Internet Only Manual (IOM) Publication 127. Gosﬁ eld, supra, n.76. 100-04, Chapter 12, Section 30.6.1.b, and 128. http://www.hhs.gov/about/. 100-02 http://www.cms.gov/Regulations- 129. Youngberg, BJ, “Principles of Risk Manage- and-Guidance/Guidance/Manuals/ ment and Patient Safety,” Jones & Bartlett downloads//clm104c12.pdf. Learning, LLC: p. 11 (2011). 114. 42 C.F.R. § 410.61. 130. http://www.fearonlevine.com/Resource. 115. Ibid. aspx?id=7ca3287d-efe3-48a8-9be5- 116. https://www.cms.gov/Regulations- ee38a3ae3f64, http://oig.hhs.gov/reports- and-Guidance/Guidance/Transmittals/ and-publications/archives/workplan/2012/ downloads//R88BP.pdf. Work-Plan-2012.pdf. 117. Ibid. 131. http://www.dhp.state.va.us/PhysicalTherapy/ 118. Ibid. pt_faq.htm. Americans with Disability Act: Financial Aspects of Reasonable Accommodations and Undue Hardship

Sandra K. Collins and Eric P. Matthews

The Americans with Disability Act (ADA) is a signiﬁ cant piece of discrimination legislation that merits ongoing managerial exploration. This civil rights legislature indicates that employers are expected to provide reasonable accommodations to employees with reported disabilities. The statute also indicates that employers can refuse to offer a reasonable accommodation if doing so creates an undue hardship on the organization. However, health care managers should exercise extreme caution when using undue hardship as a defense against providing reasonable accommodations to employees with disabilities. This point should be duly noted by health care managers given that studies indicate that lawsuits alleging disability discrimination are on the rise. This is unfortunate given the costs of reasonable accommodations are typically miniscule. Key words: American Disability Act (ADA), reasonable accommodations, undue hardship, disability discrimination.

he Americans with Disability Act consideration. Health care managers may (ADA) is a challenging piece of fi nd it helpful to make offering disabled T legislation for managers in every industry to fully understand. Ambiguous Sandra K. Collins, MBA, PhD, is an associate terms such as “reasonable accommodation” professor and program director of the Health Care and “undue hardship” may make it diffi - Management Program at Southern Illinois Univer- cult for health care managers to accurately sity Carbondale. She has approximately 18 years and uniformly apply the principles of this of experience in the management of health care paramount civil rights law. Although the related facilities and has authored numerous publi- legislation can be perplexing, health care cations concerning human resource management and aspects in the health care fi eld. She lectures at the managers should carefully explore the stat- state, national, and international level on a variety ute given that the number of cases alleging of management related topics and can be reached at disability-related discrimination may be on [email protected]. the rise.1 Eric P. Matthews, PhD, is an assistant professor and As the ADA laws are somewhat vague, program director of the Radiography Program in conducting an in-depth exploration may still the College of Applied Sciences and Arts, Southern leave health care managers perplexed as to Illinois University Carbondale (SIU). He has taught within the School of Allied Health at SIU for nine the expectations of the statute. Reasonable years; he has over 19 years of experience in diagnos- accommodation and undue hardship are two tic imaging, ranging from large trauma centers and fundamental aspects within the ADA statute teaching hospitals to outpatient imaging facilities. which health care managers should fully He has published a wide range of articles and given understand. Increasing awareness of these numerous presentations on topics ranging from medi- topics is perhaps a positive step in assur- cal history to modern medical ethics and law. ing their organizations and employees are J Health Care Finance 2012; 39(1):79–86 given appropriate levels of protection and Copyright © 2012 CCH Incorporated 79 80 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

workers an equal employment opportunity were necessary and have worked diligently while simultaneously guarding their organi- to introduce more protective mechanisms for zations from unsavory and costly legal con- those with disabilities.6 In 2008, the Act was sequences their goal.2 further revised providing more emphasis on the episodic illnesses that are expected to A Historical Overview of ADA accompany the infl ux of soldiers returning home from the current war.7 The ADA obvi- Overseen by the Equal Employment ously continues to remain an evolving work Opportunity Commission (EEOC), the in progress. ADA specifi cally forbids employers from The ADA is often criticized because it discriminating against people with disabili- lacks the type of specifi city that provides ties if they are otherwise qualifi ed for the employers with solid guidance pertaining position. ADA protection is afforded to all to legal compliance. Allegations of discrim- aspects of employment including the appli- ination based on disability are reviewed by cation process, employee selection and hir- the EEOC on a case-by-case basis so as to ing, compensation, training, disciplining, allow qualifi ed individuals with disabilities promotion, and termination.3 Also known to compete for a variety of positions. Yet, at Title VII and/or the Civil Rights Act of the vagueness of the ADA makes it diffi cult 1964, originally legislation on the issue of for health care managers to know exactly discrimination sought to ensure that every- how to comply with ADA regulations.8 one, (without discrimination based on race, color, religion, gender, or national origin), would be given the right to pursue the work Legal Issues Surrounding Disabilities of their choice. Segregating and classifying employees or applicants which prevented Health care managers will fi nd it nec- them from equal employment opportunities essary to appropriately consider the vast were specifi cally addressed and forbidden array of complexities that are associated in Title VII legislation.4 with ADA legislature. This may include an Those with disabilities were mentioned in intense review of departmental policies and a peripheral fashion within the Title VII leg- procedures which are linked to all applica- islation, but the Rehabilitation Act of 1973 ble employment transactions. These range brought forth more specifi cations in greater from employee recruitment and job descrip- detail. The Rehabilitation Act became a cat- tions to the disciplining and termination of alyst to address perceptions associated with employees. The entire employment process employing disabled workers. It also paved should be closely evaluated to assure the the way for stricter guidelines and conse- absence of discriminatory practices.9 quences to be imposed against employ- Adhering to the guidelines provided by ers who discriminated against those with the ADA in terms of employing individu- disabilities.5 als with disabilities is a vital management Since then, the ADA has undergone mul- competency for any health care manager tiple revisions. In 1990, legislators again since the health care fi eld accounted for deemed that further amendments to Title VII more than 13 million jobs in 2004 and is ADA: Financial Aspects of Reasonable Accommodations and Undue Hardship 81

expected to provide over 19 percent of all of the job. Examples might include new jobs between 2004 and 2014.10 There- changing the counter height for indi- fore, an exploration of the most paramount viduals using wheelchairs or designing and signifi cant expectations regarding the computer keyboards that allow for var- employment of individuals with disabilities iations of hand/arm position or move- is a necessary action health care managers ment; and should initiate. Among the most relevant 3. Provide equal access to workers with issues for managers to fully grasp in terms disabilities in terms of benefi ts and of the ADA are the legal responsibilities sur- privileges of employment.13 For exam- rounding reasonable accommodations and ple, the salary for a position should be undue hardship.11 based on areas such as the individual’s skill and abilities, experience, educa- Reasonable Accommodation tion, and knowledge. It should not be based on disability. Therefore, an When a disabled employee is not capa- employee with a disability should not ble of performing the functions of his or her be offered a lower salary than that of job, the ADA requires that employers fi nd another non-disabled employee if their a reasonable accommodation for the disa- positions are comparable. bled worker if he or she is otherwise qualifi ed for the position. The term “reasonable The ADA would expect for employers to accommodation” means altering the job create a reasonable accommodation for an environment or the application method so individual with a disability who is otherwise that a qualifi ed individual is still capable of qualifi ed for the job.14 Being qualifi ed to per- doing the job and therefore is provided an form a job means that the individual must equal employment opportunity. Reason- satisfy the vital skills, experience, education, able accommodation guidelines given in the and other job related qualifi cations of the ADA statute include: position with or without a reasonable accommodation. If there is a portion of the job that 1. Modifi cation of application methods the individual cannot perform due to his or so as to make sure disabled workers her disability, it is paramount that the health have an equal opportunity to engage care manager be able to effectively demon- in the competitive job market. Exam- strate why that portion of the position can- ples of this might include providing not be modifi ed and is truly job related and/ interpreters to assist in the interview or consistent with business necessity. The process for those with hearing impair- EEOC seemingly scrutinizes these claims ments to make sure any questions or heavily. In essence, it is not prudent for statements within the application pro- health care managers to overlook employ- cess are absent discriminatory intent or ing a disabled employee because he or she perception;12 cannot enter the data into the computer fast 2. Modifi cation of the physical require- enough if that skill is not a requisite neces- ments of the position so disabled sity of the position. Health care managers individuals can perform the functions need to be able to legally defend any and all 82 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

job requirements which are associated with of the job as a whole or a restructuring of the each position.15 job assignments and duties, to the modifi ca- However, there are no protections in the tion of work schedules or equipment. The ADA afforded to those with disabilities that accommodations need not be cost prohibitive are not otherwise qualifi ed for the position. or diffi cult.19 In other words, just because an individual applies for the position and he or she happens Undue Hardship to have a disability, a health care manager Under the ADA, an employer can be par- does not have to hire that individual if he or doned from the expectations associated with she is not otherwise qualifi ed for the position reasonable accommodation if the employer in question. Too many times, those in charge can demonstrate that it creates an undue of the hiring process make irrational employ- hardship on the organization.20 Undue hard- ment related decisions because they are fear- ship, another elusive term in the ADA legis- ful of discrimination allegations. If the health lation, has been successful as a defense for care manager can clearly document that the employers when the reasonable accommo- individual is not qualifi ed for the job, then dation creates a signifi cant organizational he or she can more easily demonstrate that the expense. Criteria for this defense indicates disabled individual was not selected for the that the reasonable accommodation must be position due to factors other than the disabil- either exceedingly diffi cult to implement, ity. The ADA only offers protection for dis- cause extensive organizational disruption, crimination if the hiring manager refuses to or create a signifi cant change to the mission hire someone because of his or her disability. and/or operations of the organization.21 There was no intention in the statute to indi- Using the undue hardship defense should cate that employers are required to hire any 16 not be entered into lightly by astute health unqualifi ed person regardless of disability. care managers. The following data demon- Although the aforementioned guidelines strate the estimated costs of providing rea- are helpful, deciding exactly what an employ- sonable accommodation from the employer’s er’s responsibility is in terms of reasonable perspectives: accommodation is still challenging. The foundational premise is that if the cost of the rea- • 46 percent of reasonable accommoda- sonable accommodation does not outweigh tions resulted in no cost to the employer; the benefi t of having the disabled employee, • 45 percent of reasonable accommoda- then the ADA would provide protection if the tions resulted in a one-time median cost reasonable accommodation is refused by the of $500;22 employer.17 The ADA mandates employers • 14 percent of employers indicate the look for avenues by which to meet organi- cost of providing reasonable accommo- zational requirements and consideration for dation was more costly than they origi- the disabled worker.18 Since there is latitude nally anticipated;23 in what a reasonable accommodation might • 50 percent of employers indicate be, it really is only limited by the imagination the cost of providing a reasonable and creativity of the health care manager. It accommodation was exactly what or could range anywhere from the modifi cation even less than originally anticipated.24 ADA: Financial Aspects of Reasonable Accommodations and Undue Hardship 83

With these percentages in mind, it is pos- involved with the litigation of a disability- sible to understand why claims of undue related discrimination case could easily hardship may be heavily scrutinized by the result in a six-fi gure expense. Remedies for EEOC. Therefore, it is seemingly obvious violations under the ADA can range from that the largest percentage of reasonable reinstatement and retroactive promotion of accommodations should be relatively feasi- the employee, back pay, and attorney fees. ble in terms of cost. Undue hardship might Punitive damages have even been awarded be a diffi cult case to prove given the afore- in some extreme cases.31 This is perhaps why mentioned statistics.25 This should motivate failing to properly identify an employee cov- health care managers to accurately consider ered by the ADA is listed as the number one reasonable accommodations for disabled costly return-to-work mistake made by some workers and help them realize that reason- employers in the United States.32 able accommodations may not be all that cost prohibitive. The most common forms of Identifying Disabilities reasonable accommodation include: There are numerous disabilities that may • Modifi cations to the physical work be protected under ADA legislation. Charges 26 environment; of disability discrimination have been fi led • Changes to the employee’s work sched- the most for disabilities associated with the 27 ule; or following: • Purchasing a product or piece of equipment (often a one-time purchase).28 • Bad backs; • Neurological impairments; When considering the cost factors associ- • Emotional and/or psychological impair- ated with providing reasonable accommo- ments; and dation, it may also be wise for health care • Physical impairments affecting the heart, managers to consider the benefi ts of offer- vision, and/or hearing. ing reasonable accommodations to disabled employers. The typical benefi ts include: Although those are the most common, even conditions such as high blood pressure, • The capacity to retain valuable em - allergies, morbid obesity, cosmetic disfi gure- ployees; ments, anatomical loss, and signifi cant color • Increased levels of employee productivity; blindness have been considered disabilities • Improved collegial interaction and over- and afforded protection under the ADA. Fur- all company morale; and thermore, protection for those with paraple- • Increased levels of overall organiza- gia, epilepsy, and HIV or AIDS, can also be tional productivity.29 considered as a disability along with numerous other conditions.33 The monetary value of these benefi ts, The ADA provides the following guide- although potentially intangible and diffi cult lines to assist in identifying disabilities: to calculate, could well exceed any costs of implementing reasonable accommodations 1. A physical or mental impairment which for disabled workers.30 Furthermore, the fees restricts one or more of life’s major 84 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

activities in a signifi cant way. Major receive protection under the ADA for her life activities may consist of taking care disability-related issues.37 of one’s self, completing physical tasks, Another example the EEOC provides walking, running, seeing, hearing, regarding disability involves an employee speaking, breathing, learning, think- who fractures his back. He provides his ing, sitting, standing, reaching, lifting, employer with a physician’s release which concentrating, sleeping, or interacting instructs the employer that the individual is with others.34 to refrain from any form of lifting for approx- 2. A historical fi le or record which can imately six weeks. However, the physician confi rm and verifi es the individual’s indicates the individual is expected to make specifi c impairment. This includes a full recovery and be able to resume his nor- conditions such as cancer, heart dis- mal job duties without any restrictions at the ease, or mental disorders.35 end of this timeframe. Due to the short-term 3. The individual is perceived to have nature of this employee’s condition, which an impairment. The ADA has certain is not substantially limiting, he would most conditions that it has determined may likely not be able to receive protection under be considered disabilities although the the ADA for his condition.38 individual may not have challenges The ADA also provides a list of those con- with the performance of his or her ditions that are not considered to be disabili- daily life activities. These are typically ties. Conditions distinctively excluded from impairments that the public perceives ADA shelter include the following: to be disabilities, such as something like a disfi gurement due to severe • Homosexuality and bisexuality. The burns caused by a fi re. The mere per- ADA excludes these issues as disabili- ception of disability may subject the ties, but there may be some state and individual to disability discrimination local legislation that affords protection even though the individual can func- against discrimination due to sexual tion like those absent a disability.36 orientation;39 • Gender-identity disorders that are not Two examples, as provided by the EEOC, created by a physical impairment or dis- offer guidance that may prove helpful in order related to sexual behavior such as terms of distinguishing between a condition transvestitism or trans-sexualism;40 the ADA would consider a disability and • Voyeurism;41 one that would not be considered a disabil- • Compulsive gambling, pyromania, klep - ity. The fi rst example involves an employee tomania;42 with multiple sclerosis. The disease causes • Certain disorders resulting from current her to experience chronic and episodic illegal drug usage such as psychoactive weakness. As her disease intensifi es, she substance abuse;43 and fi nds it diffi cult to stand or walk for days • Drug addicts who are not longer or weeks, often forcing her to be dependent dependent and demonstrate no further on a wheelchair. Considered substantially drug issues are not considered disabled limited, this individual would most likely under the ADA.44 ADA: Financial Aspects of Reasonable Accommodations and Undue Hardship 85

Conclusion is to provide equal employment opportunities to those with disabilities while protect- The ADA is a complex and legally ing their organizations from allegations of charged piece of legislation that continues disability discrimination.45 Understanding to evolve with the needs of the workforce. foundational expectations of the ADA, such Elusive and vaguely deﬁ ned, the guide- as reasonable accommodation and undue lines provided by the ADA allow health hardship, may be a positive beginning. Cost care managers to begin a close evalua- factors often may be of little to no organization of the expectations associated with tional expense; however, the consequences employing individuals with disabilities. of neglecting or misunderstanding ADA The goal for astute health care managers legislation may be astronomical.46

REFERENCES

1. Fallon, LF, Jr., McConnell, CR, Human at http://www.eeoc.gov/facts/health_care_ Resource Management in Health Care: Princi- workers.html. ples and Practice, Jones and Bartlett: Sudbury, 11. Collins, supra, n.1. MA (2007); Collins, SK, “The Americans with 12. Id.; Reasonable Accommodation Procedures, Disability Act (ADA): How ‘Reasonable’ Is accessed on Jan. 6, 2010 at http://www. Reasonable Accommodation?” in The Health da.usda.goc/oo/target/subjects/at/addlcap/ Care Manager’s Legal Guide, 51–70, McCo- foreword.htm.; Gary Dressler, Human Resource nnell, CR, Ed. (2010). Management, 9th Ed., Upper Saddle River, NJ: 2. Collins, supra, n.1. Prentice Hall (2003). 3. Id.; Jones, NL, The Americans with Disabili- 13. Collins, supra, n.1; Reasonable Accommoda- ties Act (ADA): Overview, Regulations, and tion Procedures, supra, n.12. Interpretations, New York: Novinka Books 14. Collins, supra, n.1; United States and National (2003). Rehabilitation Hospital, Answers to Ques- 4. Supra, n.1. tions Commonly Asked by Hospitals and 5. Id. Health Care Providers: ADA, the Americans 6. Collins, supra, n.1, “Good News: ADA with Disabilities Act (ADA) (1994). Amendments Can’t Be Invoked Retroactively,” 15. US and National Rehabilitation Hospital, Minnesota Employment Law, 2009, accessed supra, n.14. Oct. 15, 2009 at http://www.thehrspecialist. 16. Collins, supra, n.1; Id. com/artivle.aspx?articleid=28699. 17. Dressler, supra, n.12. 7. Id. 18. Collins, supra, n.1; US and National Rehabili- 8. Collins, supra, n.1; Jones, supra, n.3. tation Hospital, supra, n.14. 9. Collins, supra, n.1; McConnell, CR, Umiker’s 19. Collins, supra, n.1; McConnell, supra, n.9; Management Skills for the New Health Care Dressler, supra, n.12. Supervisor, 4th Ed., Boston: Jones and Bartlett 20. Collins, supra, n.1; McConnell, supra, n.9; (2006). Fisher, R, Evaluating Social Policy (1995). 10. Questions and Answers About Healthcare 21. Collins, supra, n.1; McConnell, supra, n.9; Workers and the Americans with Disabili- Fisher, supra, n.20; United Nations Gen- ties Act (Q&A) (Dec. 14, 2009), accessed eral Assembly (UN), Concept of Reasonable 86 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Accomodation, accessed on Jan. 6, 2010 at 30. Hernandez et al., supra, n.22; Blanck, PD, http://www.unorg/esa/socdev/enable/rights/ “Communicating the Americans with Dis- ahc7bkgrndra.htm. abilities Act, Transcending Compliance: 1996 22. Ofﬁ ce of Disability Employment Policy Follow-Up report on Sears, Roebuck and (ODEP), Making Workplace Accommoda- Co.,” Iowa City, Iowa (1996). tions: Reasonable Cost, Big Beneﬁ ts (2009); 31. Kuhn, D, Stout, D, “Reducing Your Workforce: Hernandez, B, McDonald, K, Lepera, N, Sho- What You Don’t Know Can Hurt You,” Strate- hana, TM, Wang, A, Levy, J, “Moving Beyond gic Finance, 85(11), 40–45 (2004). Misperceptions: The Provision of Workplace 32. Ring, K, “10 Costly Return-to-Work Mistakes,” Accommodations,” J. of Social Work in Disabil- Buildings, 104 (4), 42–44 (2010). ity & Rehabilitation, 8(3/4): 189–204 (2009). 33. Collins, supra, n.1; The SHRM Learning Sys- 23. UN, supra, n.21; Hernandez et al., supra, tem: Workforce Planning and Employment: n.22. Module 2 (2005). 24. Hernandez et al., supra, n.22; Dixon, K, 34. Fried, B, Fottler, M, Human Resources in Kruse, D, VanHorn, C, “Restricted Access: A Healthcare: Managing for Success, (3rd Ed.) Survey of Employers About People with Dis- (2008). abilities and Lowering Barriers to Work,” John 35. Id. J. Heldrich Center for Workforce Develop- 36. Reasonable Accommodation Procedures, ment, New Brunswick, New Jersey (2003). supra, n.12; The SHRM Learning System, supra, 25. Collins, supra, n.1; Goldstein, T, Simonds, n.33. C, Sanders, C, “Succeeding Together: Peo- 37. Q&A, supra, n.10. ple with Disabilities in the Workplace,” 38. Id. accessed on Jan. 6, 2010 at http:www.csun. 39. Collins, supra, n.1; Fried and Fottler, supra, edu/~sp20558/dis/reasonable.html. n.34. 26. Supra, n.24. 40. Fried and Fottler, supra, n.34. 27. Id. 41. Collins, supra, n.1; Dressler, supra, n.12. 28. Hernandez et al., supra, n.22; Hendricks, DJ, 42. Supra, n.39. Batiste, L, Hirsh, A, “Cost and Effectiveness of 43. Id. Accommodations in the Workplace: Prelimi- 44. Id. nary Results of a Nationwide Study,” Disabil- 45. Collins, supra, n.1. ity Studies Quarterly, 25(4) (2005). 46. ODEP, supra, n.22; Kuhn and Stout, supra, 29. Supra, n.22. n.31; Ring, supra, n. 32. A Real Options Approach to Clinical Faculty Salary Structure

Marc J. Kahn and Hugh W. Long

One can use the option theory model originally developed to price ﬁ nancial opportunities in security markets to analyze many other economic arrangements such as the salary structures of clinical faculty in an academic medical center practice plan. If one views the underlying asset to be the portion (labeled “salary”) of the economic value of the collections made for the care provided patients by the physician, then a salary guarantee can be considered a put option provided the physician, the guarantee having value to the physician only when the actual salary earned is less than the salary guarantee. Similarly, within an incentive plan, a salary cap can be thought of as a call option provided to the practice plan since a salary cap only has value to the practice plan when a physician’s earnings exceed the cap. Further, based on analysis of prior earnings, the Black-Scholes options pricing model can be used both to price each option and to determine a ﬁ nancially neutral balance between a salary guarantee and a salary cap by equating the prices of the implied put and call options. We suggest that such analysis is superior to empirical methods for setting clinical faculty salary structure in the academic practice plan setting. Key words: physician salary, real option theory, Black-Scholes.

inancial derivatives, as suggested by transactions in the context of derivative- their name, are derived from under- equivalents brings an important perspective F lying assets, either real (physical) or to risk management. fi nancial, and in complex contemporary mar- The value of a derivative depends on the kets can be combined and compounded in a value of the underlying asset. Fire insur- dizzying array of interactive layers. Deriva- ance is a very basic example of a derivative. tives such as collateralized mortgage obliga- The value of a fi re insurance policy depends tions and credit default swaps are currently on the value of the remaining asset (house) viewed with disdain because of their associa- after a house fi re. If the fi re damage is exten- tion with the recent economic crisis where sive, the insurance policy has a high value. they magnifi ed various fi nancial risks imbed- If the fi re damage is trivial, the policy may ded in the reckless design or creation of the be virtually worthless. Using the fi re insur- underlying assets (e.g., no-down-payment ance example, derivatives can be thought mortgages to persons having little income), of as insurance or “hedges” against adverse and their subsequent concentration in indi- economic conditions. Two basic fi nancia l vidual fi nancial institutions (e.g., AIG). derivatives routinely traded in organized Ironically, the original intent of derivatives was to provide many market participants Marc J. Kahn, MD, MBA, is the Sr. Associate Dean with opportunities to reduce risk. When at Tulane University School of Medicine and a Pro- derivatives are viewed in the light of their fessor in the Department of Medicine. He can be risk-reduction potential, they are an impor- reached at [email protected]. tant vehicle for understanding the manage- Hugh W. Long, PhD, JD, is a Professor at Tulane ment of risk in organizations in general, and University School of Public Health and Tropical Med- in certain aspects of the health care and medicine in the Department of Global Health Systems and Development. ical care sector in particular.

Derivatives are an important part of every- J Health Care Finance 2012; 39(1):87–96 day ﬁ nancial transactions and viewing such Copyright © 2012 CCH Incorporated 87 88 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

markets are “put” and “call” options. The theory was developed. Real option analy- owner (holder) of a put option has the right sis was originally applied to capital budg- (but not the obligation) during a particular eting (the acquisition or liquidation of period of time to sell (put) the underlying “real” (physical) assets), but now extends asset at a predetermined price to the writer to the analysis of any fi nancial opportunity of the option (or the writer’s successor); the that can be described consistent with the owner (holder) of a call option has the right assumptions of the options model. Hence, (but not the obligation) during a particular not only real asset management, but also period of time to buy (call) the underlying management of liquid assets, working capi- asset at a predetermined price from the writer tal, contracts, and fi nancing are all poten- of the option (or the writer’s successor). tial candidates for this type of analysis and Because of these rights of sale or pur- valuation. Such opportunities have been chase, options have an intrinsic value (price) valued traditionally by taking into account related to the actual market value (price) of the profi tability of the arrangement, con- the underlying asset. Although several meth- verting accrual accounting numbers into ods have been used to value (price) options, cash fl ows, and adjusting for the differen- the Fischer Black and Myron Scholes pric- tial timing of cash fl ows by calculating their ing model published in 19731 building on present values using the cost of capital and work by Robert Merton, is the most widely required rates of return. Real options are an accepted. (That work was recognized in alternative to using net present value analy- 1997 by the award of the Nobel Prize in Eco- sis, particularly in situations where the vari- nomics to the two surviving members of the ability in the cash fl ows of the underlying trio, Robert Merton and Myron Scholes). asset is diffi cult, largely beyond the con- Any asset can be the “underlying” asset for trol of management. Although both meth- an option, and the attractiveness of options ods can be used in the health care setting, lies in the fact that the market value (price) real options have seldom been described of the underlying asset varies through time, in decisions made in departments of aca- including in particular the period of time dur- demic medicine.2 In this article, we use real ing which an option on that asset is in effect. option theory to describe academic faculty That underlying asset’s price (value) vari- salary models with particular attention to ability can refl ect myriad factors, but when salary guarantees and caps, and use the all is said and done, the owner of that asset Black-Scholes model to predict the price of and potential buyers of that asset are bearing guaranteed salaries and salary caps. the risk that whenever the time comes to buy or sell that asset, the value may be higher or lower (which could be either a good thing or Methods a bad thing depending upon which side of a potential transaction one is on). Model Description The term “real” applied to options analy- Figure 1 shows the typical payoff for the sis refers to the use of option pricing theory holder of a put option on a share of common to value fi nancial opportunities other than stock. As shown in the fi gure, a put option the traded stocks and bonds for which the has no value for the holder as long as the A Real Options Approach to Clinical Faculty Salary Structure 89

Figure 1. Payoff for a Put Option Note that the payoff is $0 for share price (ST) greater than the

strike price (K). Payoff is only positive when ST is less than K.

Payoff

Strike Price (K)

share price (ST) is above the exercise price option only has value when the share price (K) of the put. For example, if the exercise is higher than the strike price. For example, price is $100 and the share price is $120, a call option representing the right to buy then the put has no value since the holder of the share for $100 only has value when the the put would not sell the share to the writer share price is greater than $100. At that point, of the put for $100 when it could be sold the holder of the call option can purchase the in the open market for $120. On the other asset at a price less than market value. hand, if the share price is $80 with the same Clinical faculty physician salaries at exercise price of $100, then the put is worth academic medical centers, like most sala- $20 as the put allows the holder of the option ries of professionals, are strongly inﬂ u- to sell the share for $100 despite the actual enced by the supply of and demand for the share price being only $80. particular talents needed for the position. Figure 2 shows the same analysis for a They are also rationally related to the eco- call option. Contrary to a put option, a call nomic value of the medical care delivered 90 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 2. Payoff for a Call Option

Note that the payoff is $0 for share price (ST) less than the strike price (K).

Payoff is only positive when ST is greater than K.

Payoff

Strike Price (K)

by the physician, particularly that portion the center or the physician. These would of the value the employing academic medi- include, to mention just a few: cal center can realize. In particular, that portion is what the faculty practice plan • The demographic and epidemiological can collect, largely from claims ﬁ led with characteristics of the area from which third parties and to a degree from billings patients are drawn; to patients, for the care delivered. Viewed • The mix of actual CPT codes delivered; as an asset, this value of what will be col- • The mix of those patients payers (e.g., fee- lected for the care delivered can be highly for-service, capitated; Medicare, Medic- variable, depending on many factors, very aid, Tricare, workers’ compensation, etc.; few of which are controllable by either Aetna, the Blues, United Health, etc.); A Real Options Approach to Clinical Faculty Salary Structure 91

• Which services are covered and which payoff represented by the difference between are not; the guarantee (put) and the actual calculated • Changes in payment rates through time salary (share price). Because a salary guaran- (e.g., varying deductibles and copay- tee provides a potential benefi t to the physi- ments; and cian, it can be priced. Importantly, the price of • Expectations regarding the delivery of a salary guarantee put option is not merely the uncompensated care. difference between the salary guarantee and previous earnings. Rather, because earnings Hence, if the cost of practice and profes- vary over time, the price of the put option is sional liability were estimated to be, for partially dependent on the volatility of both example, 55 percent of collections, the “sal- productivity and collections, which can be ary” of a clinical faculty physician might be estimated from past experience and used to set at an amount equal to approximately 45 generate the statistical likelihood that the phy- percent of expected collections for the care sician’s calculated salary will be less than the expected to be delivered by the physician. salary guarantee, and that value (to the phy- Fixing the salary at an “expected” amount, sician) can be calculated using mathematical of course, creates risks for both the physi- principles that are part of option theory. cian and the faculty practice plan because of In our model, a salary cap can be thought the variability in the expected amount that of as a call option. In fi nancial terms, the may be available to collect and the amount physician has written (is “short”) the call that may actually be collected. and the practice plan is “long.” The call Refl ecting the fact of these risks, it is com- option provides a benefi t to the employer mon for clinical faculty physician salaries to as it places a maximum value on the poten- include both salary guarantees and salary caps. tial compensation provided the physician. Suppose then that the underlying asset, The employer can “call” the underlying using our example, is the medical care pro- asset (the economic value of the collections ductivity of the physician, the value of which from the physician’s care activity) at the cap is defi ned as 45 percent of collections. A guar- amount even if (and only if) the calculated anteed minimum annual salary can be thought salary exceeds that amount. As with a salary of as a put option on that underlying asset. guarantee, the price (representing here the More specifi cally, in fi nance terms, a physi- value of the cap to the employer) of the sal- cian with a salary guarantee is “long” on the ary cap can be calculated using mathemati- put and the faculty practice plan has written cal principles that are part of option theory. (is “short”) the same option, meaning that the According to the Black-Scholes option pric- physician can, at the end of the year, sell to the ing model, put and call options can be priced practice plan the physician’s productivity for in a manner that depends on fi ve variables: the year at the set minimum price, regardless of its defi ned value. As such, if the calculated • The share price (calculated salary); value of the physician’s earnings exceed the • The exercise price of the option (salary guarantee, the put has no payoff. Contrarily, guarantee or cap in our model); if the calculated value of the physician’s ser- • The continually compounded risk-free vices is less than the guarantee, the put has a interest rate; 92 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 3. Equations Used in the Salary Real Options Pricing Model

Volatility Formula

n 1 2 volatility = ∑()x − x −1 i n i=1

x = natural log of cash fl ow returns (ln (earnings month 2/earnings month 1)) – x = mean natural log of cash fl ow returns n = number of intervals (11 for a 12 month period)

Black-Scholes Model

−rT c = S0N(d1)−Ke N(d2)

−rT p= Ke N()−d2 − S0N()−d1 2 ln + +σ ()S0 K ()r 2 T d1 = σ T

d2 = d1 −σ T

N(x) = cumulative probability distribution function for a standardized normal distribution to be less than x p = put option price c = call option price K = strike price (salary guarantee)

So = Stock Price (earned salary prior year) r = continuously compounded risk-free interest rate (3% in our model) ␴ = volatility of prior earnings T = time to maturity (one year for salary guarantee)

• The time to maturity of the option (one D. Also included in Figure 4 are the annual year in our model); and totals of earned salary, the volatility thereof • The volatility of the stock (volatility of calculated per the formula in the Figure 3, prior calculated salary in our model). and the price of a $150,000 salary guarantee (put option) calculated from the Black- The Black-Scholes model is most com- Scholes model. Our model predicts the price monly used by ﬁ nancial experts to price of the salary guarantee to be highest for phy- options traded on the Chicago Board sician A. This is intuitive since physician A Options Exchange. Both the Black-Scholes has a salary guarantee that is much higher model and a simple method for calculating than the physician C’s total calculated sal- volatility of physician-calculated salaries ary the previous year. Physicians B and C, are provided in Figure 3. both have earnings during the prior year that exceed the guaranteed salary. However, Results even though physician C had higher earnings the prior year, the price of his guaran- Figure 4 shows monthly earned salaries teed salary is higher than that calculated for for four theoretical physicians, A, B, C, and physician B. This is because the monthly A Real Options Approach to Clinical Faculty Salary Structure 93

Figure 4. Monthly Earnings for Four Physicians (A, B, C, and D) Guaranteed salary is $150,000 for each example. Monthly earnings are derived as random numbers between $10,000 and $16,000.

Physician A’s Physician B’s Physician C’s Physician D’s Time Period Earned Salary Earned Salary Earned Salary Earned Salary July $13,968 $15,058 $11,364 $12,032

August $10,423 $11,058 $15,719 $12,300

September $13,400 $10,308 $10,632 $12,400 October $10,895 $12,040 $12,728 $12,004

November $11,007 $12,884 $11,999 $12,543

December $12,017 $14,035 $14,499 $12,578

January $10,826 $15,044 $13,002 $12,222

February $14,246 $10,265 $15,337 $12,378

March $10,715 $10,191 $10,203 $12,007

April $12,265 $14,317 $15,470 $12,068

May $11,426 $15,656 $14,672 $12,309

June $10,151 $15,064 $13,870 $12,223

Total $141,339 $155,920 $159,495 $147,064

Volatility 18.95% 19.36% 25.42% 2.32%

Put price $5,190.19 $745.88 $887.34 $13.25

Salary cap having $159,387.97 $178,725.64 $189,143.80 $153,426.53 value equal to put earnings for the prior year for physician C makes stocks and bonds less desirable. In are more volatile. Somewhat paradoxically, contrast, physician D has the lowest salary options have a higher price as the volatility volatility among the examples. As such, of the underlying asset (earnings) increases physician D’s salary guarantee is not only because volatility adds to the likelihood that the lowest of the four, but, in particular, is the option will be “in the money.” less than that of physician A in spite of the This positive relationship between vola- fact that physician D had higher actual earn- tility and price is unique for options. Stocks ings than physician A. and bonds tend to have a lower price with Since the value of the put option (salary increased volatility, since price uncertainty guarantee) accrues to the physician and the 94 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

Figure 5. Collection Volatilities Calculated for Tulane University Medical Group by Department

Department Calculated Volatility Family Medicine 11.9% Internal Medicine 23.6% Neurology 33.1% Obstetrics/Gynecology 16.4% Pediatrics 24.4% Psychiatry 57.7% Surgery 23.8%

value of the call option (salary cap) accrues guarantees to physicians, the risk is shifted to the practice plan, an equitable arrange- from the clinician to the plan. When the plan ment would equate the values of the two is large, having clinicians from many differ- options. Using simple algebra and the Black- ent departments, the risk becomes diversi- Scholes model for the price of a call option, fi ed when shifted to the plan. As such, the the salary cap that would have a call option risk is less subject to fl uctuations in indi- price identical to that of the put option (sal- vidual earnings. This diversifi cation of risk ary guarantee) is also shown in Figure 4. is similar to an individual investor having a Although the previous discussion involves portfolio of diverse investments in order to theoretical earnings, we have used actual protect against market fl uctuations that may account collection data from the Tulane Uni- have great impact on any single investment. versity Medical Group to calculate collection Although clinical faculty salary guar- volatilities for individual departments (Fig- antees have a value and can be priced as ure 5). As shown in the Figure 5, the actual shown by our analysis, physicians are sel- volatilities vary by specialty. Specialties with dom “charged” for this benefi t. This results business with more seasonality have higher in fi nancial tension. Obviously, the cost of volatility. Primary care specialties like family a salary guarantee must be balanced with medicine have lower volatility due to the rela- another source of revenue to prevent defi cit tively constant nature of their income streams. fi nancing. We suggest three possible scenar- ios that could be used to resolve this tension. Conclusions Likely, in many instances, a faculty practice incentive plan may in effect balance the Why salary guarantees exist as a part of put option provided the physician as a salary a compensation plan is a fundamental ques- guarantee with a call option that the physi- tion. On the surface, a salary guarantee can cian provides back to the faculty plan in the be used as an enticement to get a physician form of a salary cap. As shown in our prior to join a practice. From a fi nancial per- analysis, a salary cap provides a direct ben- spective, when a practice plan offers salary efi t to the practice plan since any earnings A Real Options Approach to Clinical Faculty Salary Structure 95

over the cap are kept by the plan. In options a physician’s salary guarantee is easily met terms, the call option of a salary cap is “in and therefore the put option has little value. the money” for the practice plan when the Finally, the tension created by salary guar- physician earns more than the cap. Although antees may be relieved if the price of the not typically calculated as such, within a fac- option is paid from sources other than clini- ulty practice plan, it is likely that a combina- cal income. Such would be the case if reve- tion of puts and calls in the form of salary nues from tuition, endowment, philanthropy, guarantees and salary caps balances out to or research were used to support the clini- avoid fi scal defi cits. cal enterprise. In fact, it is not uncommon to As shown in Figure 4, to remain fi nancially use hospital revenues from graduate medical neutral, physician A would have the lowest education to support physician salaries.3 salary cap since physician A has the lowest A recent report has suggested that among earnings during the prior year and the most medical school tenured faculty in clinical expensive put option. In contrast, physician departments, 44 percent have a salary guar- C would have the highest salary cap based antee linked to tenure.4 In this report, the on earnings from the previous year and high majority of the salary guarantee represented volatility of prior earnings. In short, the cost a guaranteed base salary. Further, the num- of the salary guarantee to the practice plan ber of medical schools offering a guaran- can be balanced by the benefi t of a salary cap teed salary has been decreasing over time.5 and option theory can be used to calculate In addition to issues related to the rules for the exact salary cap that would provide the clinical faculty in obtaining tenure, we sug- same risk (same option price) as the salary gest that some of the erosion of salary guar- guarantee. As an investment tool, balancing antees among clinical faculty is due to the the benefi t of a put option with the price of inherent cost of a salary guarantee for the a call option is called a “zero premium col- medical school, and diffi culties in estimat- lar.” A zero premium collar allows an investor ing these costs. We suggest that our analysis to hedge against losses with the put that is in is a reasonable step in estimating such costs. effect paid for with the limitation in value pro- Our analysis must be considered in the con- vided by the call. Because the put and call are text of the limitations of the Black-Scholes priced identically, the result is a fi nancially analysis. First, Black-Scholes bases future neutral position as shown in our salary model. expectations on prior events. Specifi cally, Another possible resolution to the tension our analysis is based on monthly earnings provided by a salary guarantee would be if for the year prior (although longer histori- the salary guarantee is signifi cantly less that cal horizons could also be used). Second, the the physician’s historical earnings. From a Black-Scholes equation works very well for fi nance perspective, the price of an option fl uctuations in share price (monthly salary) decreases as the likelihood that the option that are incremental and smooth. In fact, the will be “in the money” decreases. For exam- equations stem from mathematical equations ple, a physician with a salary guarantee of used to describe the transfer of heat across $150,000 who has historically earned well a solid where small areas become heated in over $200,000 has not been provided with a an incremental fashion. As such, our model very costly option. We suggest that at times, would become more precise if weekly or 96 JOURNAL OF HEALTH CARE FINANCE/FALL 2012

daily salary were used rather than monthly have been described in such terms.6 We sug- earnings. In fact the exponential function gest that analysis like ours can provide a (ex) contained in the Black-Scholes equation mathematical basis for the implementation implies changes that are small and continu- of an incentive plan that includes salary guar- ous. Black-Scholes works less well when antees and caps. We further suggest that our changes are large and sudden. We suggest analysis is superior to salary predictions that that fl uctuations in clinical faculty earnings are merely empirical. are typically smooth. Only with sudden cata- Real option theory can be used to assess strophic events such as disability-causing the medical school practice plan cost of sal- physician illness, near total disruption of a ary guarantees and the model presented can health care system, or closure of a sole facil- be used to balance such costs with the fi nan- ity providing care would this analysis falter. cial benefi t of a salary cap. We suggest that Although academic physician salary struc- such analysis can provide a reliable estimate ture has not been described previously in of salary projections and would be useful the literature in terms of real options theory, for institutional physician salary modeling managed care contracting and regulation within incentive plans.

REFERENCES

1. Black, F, Scholes, M, “The Pricing of Options 4. Bunton, SA, “The Relationship Between Tenure and Corporate Liabilities,” Journal of Political and Guaranteed Salary for U.S. Medical School Economy, 81: 637–654 (1973). Faculty,” AAMC Analysis in Brief, 9 (2010). 2. Williams, DR, Hammes, PH, “Real Options 5. Bunton, SA, Mallon, WT, “The Continued Reasoning in Healthcare: An Integrative Evolution of Faculty Appointment and Tenure Approach and Synopsis,” Journal of Health- Policies and U.S. Medical Schools,” Acad. care Management, 52: 170–186 (2007). Med., 82: 281–289 (2007). 3. Cohen, JR, Fox, S, “Developing a New Faculty 6. McLean, RA, Magiera, FT, “Options Analysis Practice Plan with a Model for Funds Flow of Managed Care Contracting and Regulation: Between the Hospital and the Plan, Acad. Theory and Evidence,” Health Serv. Manage Med., 78: 119–124 (2003). Res., 13: 170–177 (2000).