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

Joanne Mitchell-George, Senior Managing Editor William O. Cleverley, PhD Dom Cervi, Marketing Director Professor Ohio State University 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 Department of Accounting, The University of Services, 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 40:3, Spring 2014 Contents

1 Hospital Diversiﬁ cation Strategy Steven R. Eastaugh 14 Public Hospitals in Peril: Factors Associated with Financial Distress Zo Ramamonjiarivelo, Robert Weech-Maldonado, Larry Hearld, and Rohit Pradhan 31 Using a Social Entrepreneurial Approach to Enhance the Financial and Social Value of Health Care Organizations Sandra S. Liu, Jui-fen Rachel Lu, and Kristina L. Guo 47 Health Care Reform and the Development of Health Insurance Plans: The Case of the Emirate of Abu Dhabi, UAE Samer Hamidi, Dr.PH, Sami Shaban, PhD, Ashraf A. Mahate, PhD, and Mustafa Z. Younis 67 The Impact of Star Physicians on Diffusion of a Medical Technology: The Case of Laparoscopic Gastric Bypass Surgery Laura Shinn 86 Health Policies and Intervention Strategies Aliye T. Mosaad and Mustafa Z. Younis 101 Determinants of Differentials in Pneumonia Mortality in the UK and France Rizwan ul Haq, Patrick Rivers, and Muhammad Umar From the Editor— About This Issue

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

Steven R. Eastaugh

To determine the impact of health system restructuring on the levels of hospital diversifi cation and operating ratio this paper analyzed 94 teaching hospitals and 94 community hospitals during the period 2008–2013. The 47 teaching hospitals are matched with 47 other teaching hospitals experiencing the same fi nancial market position in 2008, but with different levels of preference for risk and diversifi cation in their strategic plan. Covariates in the analysis included levels of hospital competition and the degree of local government planning (for example, highly regulated in New York, in contrast to Texas). Moreo- ver, 47 nonteaching community hospitals are matched with 47 other community hospitals in 2008, having varying manager preferences for service-line diversifi cation and risk. Diversifi cation and operating ratio are modeled in a two-stage least squares (TSLS) framework as jointly dependent. Institutional diversifi cation is found to yield better fi nancial position, and the better operating profi ts provide the fi rm the wherewithal to diversify. Some services are in a growth phase, like bariatric weight-loss surgery and sleep disorder clinics. Hospital managers’ preferences for risk/return potential were considered. An institution life cycle hypothesis is advanced to explain hospital behavior: boom and bust, diversifi cation, and divestiture, occasionally leading to closure or merger. Key words: product line selection, diversifi cation, risk, operating profi t margin, regulation

Political leaders often promote compre- Successful general hospitals are increas- hensive health system reform without any ingly emulating specialty hospitals by plac- forethought into how it will impact organi- ing focus on a narrow number of specialized zations small and large. In this study we departments. Specialty hospitals are emulat- consider the largest of health care organiza- ing global business strategies for specialized tions, the hospital, and how shifts in govern- “focus factories,” for example, China, South ment policy and local fi nancial conditions Korea, Japan, Italy, and Germany.1 The hos- impact diversifi cation decisions. Diversifi - pitals, general and specialty, are increasingly cation is a key strategy to support the core selecting narrow posture decisions for what business, the inpatient hospital mothership, services to offer. For example, you might and a number of satellite service ventures. specialize in weight loss surgery or sleep dis- Interest in service product-line diversifi ca- orders, and also select a trio-combo strategy tion is evidenced by the popularity of con- like neonatal intensive care unit plus pedi- ferences and courses on hospital corporate atric inpatient special unit plus organized planning and marketing. Diversifi cation is a strategic management issue in the development of a marketing plan. For example, Dr. Steven R. Eastaugh was a professor at Cornell the basic way to build market share involves University and GWU for 37 years. He now runs the development of more markets for cur- Eastaugh Econometrics and can be reached at eas- rent service, and more services for current [email protected] markets. Many fi nance faculty argue that Acknowledgements: The author acknowledges the hospital that fails to diversify will be the assistance of Ed Roberts, Thomas Davy, Kevin relegated to a “plodder” or negative growth Smolich, Craig Rosenfeld, Linda Kleckner, and Erika existence. The plodder is not positioned to Schouten. benefi t from insurance exchanges or value J Health Care Finance 2014; 40(3):1–13 based purchasing (VBP). Copyright © 2014 CCH Incorporated 1 2 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

pediatric outpatient department. In some The new lines of business we will con- highly uninsured states, Obamacare (the sider in this empirical study are all related Affordable Care Act passed March 2010) ventures (that is, only expanding exist- might most signifi cantly boost diagnostic ing health product lines of service). Pub- testing volume. lic image and profi ts often improve after Such market posture decisions typically a diversifi cation expansion. Four teaching precede marketing mix decisions. Poorly hospitals reported profi table diversifi ca- performing general hospitals often fi xate tion efforts providing dental care for the on the initial posture decision stage, and do handicapped and alcohol detoxifi cation and little new or different. Any attempt to have counseling services to teenagers. Obamac- a general hospital emulate a general store are expansion in the coming decade might may lead to strategic mediocrity. Hospitals expand such opportunities. Some services make the choice about how broad or nar- such as bariatric weight-loss surgery are row a diversifi cation effort should be made in a growth period of development, while (posture decision), but almost never con- other services face maturation and decline. sider the strategic decision about whether Today, any institution must collect enough to use a marketing penetration or market- retained earnings to render health services skimming pricing strategy. Indeed, pric- in the future and to rebuild and modern- ing decisions are less frequently discussed ize facilities. The product portfolio of the in the hospital sector relative to private hospital can be diversifi ed to include health industry because the consumer is so highly promotion and health education activi- insured.2 ties, such as programs designed to improve Hospital boards realize that the recent patient compliance. growth in detailed restructuring plans The sample consists of 94 teaching hos- must be tempered by concern for admin- pitals and 94 nonteaching community hos- istrative costs. The better managed facili- pitals in 12 states. Service-specifi c shifts in ties do not just pursue imitative entry or diversifi cation are listed in Figure 1. The “bandwaggoning.” For example, if CGH fastest growth area, bariatric weight-loss medical center goes into sports medicine surgery, was added during the study period we follow with a “me too entry” into the to 21 hospitals. Sleep disorder clinics have same service-line. Well managed hospi- been added at 18 hospitals. The most dis- tals analyze posture variables, and meth- turbing trend is that 24 of the hospitals ods to offer a favorable degree of product (Figure 1, Line 6) closed their emergency differentiation relative to the competition. departments in the study period 2008–2013. The situation gets “sticky” if the hospital No hospital opened an emergency room. begins to heighten preexisting levels of In the last decade 945 US hospitals have competition with their own physicians. closed their emergency department accord- The hospital, as a centralized service dis- ing to the American College of Emergency tribution innovator, has the potential to Physicians 2013.4 With the onset of the reduce the time-cost to physicians and the Affordable Care Act this trend towards travel time-cost to patients in the provision closing emergency departments should of many health services.3 slow or stop. Hospital Diversifi cation Strategy 3

Figure 1. Service-Speciﬁ c Changes in Hospital Diversiﬁ cation 2008–2013 (94 pairs of hospitals)

Net 5-Year Type of Service Deletions Additions Change Quality-enhancing services Diagnostic radioisotope facility 3 7 4 Blood bank 1 5 4 Histopathology laboratory 1 3 2 Premature nursery 7 5 (2) Emergency department 24 0 (24) Pharmacy with pharmacist 5 8 3 Complex services Neonatal intensive care unit 4 5 1 Computer-assisted detection mammography 2 9 7 Electroencephalography 4 2 (2) Pediatric inpatient unit 4 8 4 Cardiac rehabilitation 5 10 5 Respiratory therapy department 6 7 1 Cardiac intensive care unit (CCU) 4 9 5 Dental Services 4 8 4 Laser optic treatment 3 15 12 Therapeutic radioisotope 5 5 0 Rehabilitation inpatient department 6 12 7 Psychiatric inpatient unit 13 1 (12) Radium therapy 5 3 (2) X-ray, therapeutic 1 10 9 Community services Bariatric weight-loss surgery 0 21 21 Podiatric services 8 14 6 Speech pathology services 7 8 1 Occupational therapy department 5 8 3 Skilled nursing or long-term care unit 2 9 7 Sleep disorder unit 0 18 18 Alcohol/chemical dependency outpatient services 9 15 6 Rehabilitation outpatient department 7 7 0 Inpatient hemodialysis 7 8 1 Physical therapy department 6 12 6 Clinical psychology services 4 7 3 Psychiatric outpatient unit 14 12 (2) Psychiatric consultation and education services 10 10 0

AU: Does this number Background and Objectives The central problem in modeling hospital (1) mean this calcula- diversifi cation and its impact on the operat- tion is In this study, hospital diversifi cation ing ratio is that the relationship is interde- step 1? (DIV) is calculated as follows: pendent. For example, diversifi cation might AU: What improve the operating ratio in a causal fash- does this No. of services available in the hospital ion, but the baseline operating ratio might number (33) (1) DIV = refer to? Maximum Number of Services (33) be inversely proportional to the hospital’s 4 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

perceived need to diversify. Conversely, a less closely related. The situation is analo- hospital might need a certain baseline oper- gous to designing an evaluation of a new ating ratio to afford to fi nance a diversifi - drug, in which patients whose prognosis cation effort or to gamble. Diversifi cation appears to be about the same at the begin- might also occur because of expectations for ning of the study (2008) are paired. The 10 change in the operating ratio. In other words, variables in the nearest available matching the level of diversifi cation and level of oper- process5 are listed in Figure 2. The facili- ating ratio are hypothesized to be jointly ties experience a different dosage level of dependent or endogenous within the context diversifi cation over the study timeframe, but of two simultaneous equations: they start with the approximate same fi nancial performance. The researcher asks how (2) Operating ratio, M = f (diversifi ca- AU: Does the operating ratio changed in response to this (2) refer tion, exogenous factors infl uencing changes in the dose of diversifi cation. The to the next market share, and reimbursement, analogy to the drug evaluation breaks down step in the disturbance term) calculation? in one important respect: the hospitals are in control of their own dosage of diversi- AU: Does (3) Diversifi cation, DIV = f (M, fi cation; that is, expansion is endogenous this (3) refer (within system control) within some exog- to the next exogenous factors impacting hospital step in the competition, physician demand, payers enous (external, not predetermined by the calculation? demands to contain capital expansion, hospital) regulatory and market constraints. disturbance term) Consequently, the evaluation design is similar to a pre-test/post-test situation. However, It is important to note that if we are to the hospitals are not in experimental isola- study the interdependent endogenous vari- tion, and a number of other intervening variables DIV and M, and a number of other ables must be monitored during the fi ve-year exogenous variables (I1, I2, … I7, I8), it would period. The hope is that if one monitors all be inaccurate and misleading to study a typi- the market characteristics that could have an cal multiple regression equation: intervening effect (IV1, IV2, IV3…IVm), and if the matching process adequately captures AU: Does (4) DIV = f (M, (I , I , … I , I )) this (4) refer 1 2 6 7 other intrinsic characteristics (V1-V2), then to the next the impact of diversifi cation on operating step in the In a simultaneous equation situation we calculation? are dealing with the regression of a set, or, in ratio M can be measured as: this case, a pair of dependent variables (DIV, AU: What (6) M = f (DIV, I , I , I , I , I ) does this M), regressed upon a set of independent var- 1 2 3 4 5 (6) number iables. A simultaneous situation is: The advantage of this approach, rather stand for? AU: Does than collecting data to estimate a more- this (5) refer (5) (DIV, M) = f (I1, I2, … I6, I7) to the next complicated regression equation, is that step in the The 188-hospital sample is matched by one need not worry about losing too many calculation? bed size, ownership, and teaching status. By degrees of freedom in measuring attributes

pairing the hospitals, we hope that they may V1-V2 with Iz+n additional number of inde- behave more nearly alike than do hospitals pendent variables. Hospital Diversifi cation Strategy 5

Figure 2. 10 Variables Used in the Nearest Available Matching of Hospitals into 94 pairs

Variable Category Variable Diversifi cation DIV Output severity Percent families below poverty line Surgical intensity (operations per admission) Outpatient intensity (visits per patient treated) Output complexity Number of services in three increasingly complex categories in Figure 1 system of classifi cation: Quality-enhancing services Community services Complex services Financial health M, operating ratio Other institutional factors Hospital size fi xed in the short run Age of plant

There are 10 basic hypotheses under con- Given that the hospital pairs start the time sideration in our analysis. Because a hospi- period with the same approximate bed size, tal may need a suffi cient operating margin to diversifi cation (no pairs were more than diversify and a diversifi cation effort should 0.042 different in DIV), operating ratio (no Au query: improve the reimbursement portfolio pairs were more than 0.015 different in M), are outline numbers (assuming one selects above-average areas output complexity, output severity, plant necessary? in which to diversify) and consequently the age, and teaching affi liation status in base See later operating margin, I hypothesize: year 2008, I postulate that only three addi- comments. tional exogenous variables impact operat- 1.0 A non-recursive simultaneous ing ratio M: model, where DIV is a function of margin M, and, further, M is a func- 1.1 A positive infl uence of exogenous bed tion of DIV. In our model DIV and closings of neighboring hospitals on M are endogenous, because they are (increasing) the operating ratio of our to a large extent dependent on other paired sample hospitals. variables in the system and under 1.2. A positive infl uence of the exogenous the control of management. (One variable MCARE, the annual fraction could alternatively have postulated of the county catchment area over a recursive model, in which M does 65, was postulated to have a positive not directly or indirectly infl uence infl uence on M. DIV, while DIV is an indirect or 1.3 Every variable in our analysis will direct determiner of M. Instead we be defi ned as the paired difference are testing the strength of a nonre- between paired hospitals. Conceptu- cursive model, in which in terms ally, 94 hospital pairings producing of a path diagram, M and DIV are 470 observations over the fi ve-year allowed to form a closed loop.) study period is equivalent to 470 fi rst 6 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

differences of logs. When matching and increased diversifi cation. The measure hospital pairs, instead of having 188 of noncompetition is the Herfi ndahl index, observations with 10+m independ- which summarizes the entire size distribu- ent variables (where m is a very large tion of fi rms in a market. The Herfi ndahl number of additional variables needed index can range from 1.0, by defi nition, for to measure the constructs captured in a natural monopoly situation (for example, the matching process, one is left with a very rural hospital) to a low of 0.00 in our 470 observations and 5 to 6 variables sample. Because the Herfi ndahl measure is per equation. an index of noncompetition, I hypothesize: 1.4 negative infl uence of the exogenous 2.1 A higher Herfi ndahl index, as a proxy AU: Why variable POOR, the annual fraction does this of the catchment area below the fed- for less competition, will be associ- set of ated with less diversifi cation; that is, H hypotheses eral poverty line, is postulated to have start with a negative infl uence on M. This vari- should have a negative impact on DIV. 2.1 when able is taken as a proxy measure for the the fi rst set From a hospital manager’s vantage started with likelihood of bad debt. point, the direction of change (increasing or 1.0? The natural logarithmic specifi cation (ln) decreasing) of this index for noncompetition for the long-run operating ratio in year t, may be more important than the absolute with a disturbance term u , equals: t level of competition. One might postulate AU: What β β β that the change in institutional pressures to does this (7) ln Mt = 1 ln DIVt + 2 ln Et + 0 + ut, (7) number diversify over time might depend more on refer to? where the set of exogenous explanatory the recent direction of change in the level of variables (E) includes: competition; that is, if competition is on the upswing, a manager might be less averse to • BEDCLOSE = reduction in number of risk and more likely to gamble on a diversifi - beds per capita in the catchment area cation proposal. However, if the level of the since 2007; competition is stable or declining, the man- • MCARE = fraction of population eligi- ager may be less likely to gamble. In other ble for Medicare in the area; words, the impact of our anti-competitive • POOR = fraction of population that is index H on DIV may vary as a function of poor; and whether H has been decreasing in the recent • T = time trend variable (2008 = 1, 2009 past. I hypothesize: = 2, etc.).

AU: Does There are fi ve hypotheses associated 2.2 The coeffi cient for the H index is this refer with the second equation in our simulta- smaller and less signifi cant when to the equation neous model. The academic literature on H has been declining (competition labeled (2), diversifi cation in the hospital fi eld is rather increasing) in the last two years, rela- above? uneven and anecdotal. The most universally tive to the coeffi cient for H when the mentioned variable in the recent literature index has been stable or increasing. is competition.6 One purpose for this study H is postulated to have a less negative is to test the link between high competition infl uence on DIV when competition Hospital Diversifi cation Strategy 7

has been increasing (H declining), per- labeled the “fat and happy” theory of “not haps because the manager’s aversion to rocking the boat” if the boat appears to be risk-taking would be diminished in the improving without further restructuring. I increasingly competitive environment. hypothesize:

A number of analysts have suggested 2.4 The coeffi cient for M is smaller and less that the exogenous pressures to diversify signifi cant when M has been increasing are generated, in part, from specialists.7 I in the last two years, relative to the coef- hypothesize: fi cient for M when the operating margin has been stable or on the decline. 2.3 The ratio of specialist physicians per 10,000 population in the area has a This hypothesis was suggested by hospi- positive impact on diversifi cation. tal managers in the data collection process. The administrators managing consistently in AU: This The eighth hypothesis I wish to test is the “bad year” (Figure 3, Column A) com- may con- fuse some whether M has a differential impact on DIV, plained that the few hospitals able to run in readers depending upon whether the fi nancial posi- the black (Figure 3, Column C) were the because the tion of the hospital is improving or not. The hypotheses only hospitals that had the fi nancial where- are num- argument is somewhat analogous to hypoth- withal to afford to diversify aggressively and bered 1.0, esis 2.2, in that if the hospital’s fi nancial posi- consequently to improve or preserve M over 1.1,…2.1, 2.2 …. It’s tion is on the upswing, the manager might be time in an era of tight reimbursement. The easy to less likely to take a gamble on a diversifi ca- last point suggests two fi nal null hypotheses lose track of how tion proposal. This null hypothesis could be for inclusion in this analysis. many there are. Can they be numbered Figure 3. Operating Ratio (M) Trends for the Sample of 188 Matched consecu- tively as 1, Hospitals 2, 3?

Number of Matched Hospitals Operating Ratio (M)a <0.985 0.985–0.999 ≥1.0 Median Year (Bad Year) (Typical Year) (Good Year) Value M 2008 66 41 81 0.999 2009 67 43 78 0.998 2010 76 40 72 0.994 2011 79 42 67 0.991 2012 80 43 65 0.990 2013 82 43 63 0.989

a Operating ratio equals operating revenues divided by operating expenses, or 1.0 plus the operating margin. Bottom-line operating margins include nonoperating revenue in the calculation. Adding in nonoperating margins makes the average bottom-line margin 3.62 percent as of March 31, 2013. 8 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

3.1 Time is postulated to have a nega- The log-linear form will be used; thus, all AU: tive impact on operating ratio M and the estimated coeffi cients represent short- Change a potential positive impact on the term elasticity. The diversifi cation equation number- ing? See purely technology-driven component has the following specifi cation, in time t comment of diversifi cation (DIV). The argument with disturbance term vt: 14? that hospital rate regulation reduced γ γ the average operating ratio has been (8) n DIVt = 1 ln MUPt + 2 ln MSDt + AU: What well documented.8 γ ln F + γ + v does this 3 t 0 t (8) number 3.2 Teaching status may interact with the refer to? management philosophy of the institu- in which the set of variables F includes: tion and consequently have an impact θ • HDOWN = Ht = competition increas- on the equation for DIV and M. There- ing. fore, we will run the analysis for the θ • HSG = (1− ) (Ht) = competition sta- entire sample and then rerun the analy- ble or declining (where H = Herfi ndahl sis for the 47 pairs of hospitals that are θ index, and = 1.0 when Ht

Figure 4. Estimated Operating Margin Equations for Teaching and Nonteaching Hospitals Under Two Estimation Methods: Single Equation OLS and TSLS, Normalized Variable ln M

Type of Estimation ln ln ln ln Hospitala Method DIV BED-CLOSE MCARE POOR T Constant R2 Teaching TSLS 0.018c 0.018c 0.037d – 0.090b – 0.009c 0.010 0.861 OLS 0.024b 0.020c 0.040d – 0.091b – 0.009c 0.011 0.844 Non-teaching TSLS 0.017c 0.012d 0.030 – 0.079b – 0.007c 0.004 0.831 All pairs OLS 0.027b 0.012d 0.029 – 0.077b – 0.007c 0.004 0.814 TSLS 0.032b 0.028b 0.039d – 0.093b – 0.009b 0.009 0.892 OLS 0.034b 0.028b 0.039d – 0.092b – 0.009b 0.009 0.880 a The three clusters of hospitals have 243, 243, and 481 degrees of freedom, respectively. b Statistically signifi cant at the 1 percent level. c Statistically signifi cant at the 5 percent level. d Statistically signifi cant at the 10 percent level, using appropriate t-test (See References, n.9.) has its predicted positive impact on operating (rejecting hypothesis 3.1), are reported in ratio, and this is statistically signifi cant in all Figure 5. Hypothesis 2.1 is supported, as a six cases. Moderate support is presented for larger Herfi ndahl index is associated with less hypotheses 1.1 and 1.2, in that the coeffi cients diversifi cation. However, the negative impact are positive for BEDCLOSE and Medicare of H on DIV was not statistically signifi cant variables, but results are only signifi cant at for the 35 percent of cases in which compe- the 0.10 level. Strong support is provided for tition was increasing (HDOWN), strongly hypotheses 1.3 and 3.1. There seems to be supporting hypothesis 2.2. All else being a steady deterioration in the operating ratio equal in the equation, the 56 percent of hos- under the infl uence of the payment system, pitals not experiencing an increase in compe- consistent with reports that hospitals are fac- tition (HSG) were able to avoid adding two ing barebones reimbursement.10 new services in the fi ve-year period relative In our sample, the −0.007 negative impact to the other facilities (HDOWN, competi- on M per year can be compensated for by tion up). Needless to say, neither the popular +0.0028 per new service added. This suggests nor the economic concept of competition is that diversifi cation is one means to compensate adequately represented by the H index, but for increasingly tight reimbursement. Another the results are suggestive. The impact of analysis, using this same sample of hospitals, declining H or increasing DIV is approxi- indicates that the downward impact on M mately equivalent in teaching and nonteach- can be equally compensated by +0.0031 gain ing hospitals. It might be speculated that as in department labor productivity.11 In other the hospital industry’s noncompetitive posi- words, limited diversifi cation and productivity tion subsides in a neighborhood, contrary to enhancement can, when done together, com- hypothesis 3.2, the managers of teaching and pensate for declines in M caused by payers. nonteaching hospitals are equally willing to The results for the DIV equation, exclud- take risks to maintain (or improve) fi nancial ing the statistically insignifi cant T variable position and market position. 10 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 5. Estimated Diversiﬁ cation Equation for Teaching and Nonteaching Hospitals Under Two Estimation Methods: Single Equation OLS and TSLS, Normalized Variable ln DIV

Type of Estimation ln ln ln ln Hospitala Method MUP MSD HDOWNe HSG SPECMD Constant R2 Teaching TSLS 0.799c 0.812c – 0.059 – 0.702c – 0.124 – 0.014 0.815 OLS 0.790c 0.819c – 0.061 – 0.697c – 0.119 – 0.010 0.802 Non-teaching TSLS 0.721c 2.248b – 0.082 – 0.629c 0.270d – 0.020 0.762 OLS 0.719c 2.419b – 0.079 – 0.637c 0.272d – 0.018 0.751 All pairs TSLS 0.758c 1.601b – 0.069 – 0.678c 0.109 – 0.016 0.821 OLS 0.756c 1.675b – 0.070 – 0.711b 0.104 – 0.014 0.810

a The three clusters of hospitals have 241, 241, and 479 degrees of freedom, respectively. b Statistically signifi cant at the 1 percent level. c Statistically signifi cant at the 5 percent level. d Statistically signifi cant at the 10 percent level. e H (Herfi ndahl index) declining means competition is increasing, whereas HSG stable or growing implies that competition is stable or declining.

There is little support for hypotheses 2.3. diversifi cation decisions are critical. Com- Consistent with the fi nding in hypothesis 3.2 petition is often promoted as a force that is the fi nding that hypothesis 2.4 is only sup- increases diversifi cation. An army of lawyers ported in the case of nonteaching hospitals. and consultants are available in all states to Surprisingly, the inverse of postulate 2.4 sell hospitals on a diversifi cation plan. The appears to be operating in the case of teach- plan is sometimes developed on a “turn-key” ing hospitals. That is, teaching hospital man- basis (that is, all hospitals get approximately agers only gamble on diversifi cation if they the same strategy advice with only marginal are experiencing improved fi nancial health changes in where to diversify). Some naive (M increasing), but nonteaching hospital hospital managers buy the “unique” plan of managers are more willing to gamble on the more unscrupulous (atypical) consult- diversifi cation if their operating ratio is sta- ants: “diversify; it’s fail-safe, and you cannot ble or declining. Nonteaching hospital man- help but make money.” Such administra- agers are less concerned with whether they tors would do well to study the failure rate are “fat and happy” (high M) relative to the of diversifi cation programs in industry. Not other sample hospitals. Further research is enough hospitals do a complete analysis of needed to ascertain the differences between future patient volume, cash fl ows, overhead, the three types of hospital management confi xed start-up costs,12 and the potential politi- cerning risk aversion and preferences for cal and economic problem of undermining what services are best for diversifi cation. existing programs and clinics. Diversifi cation is often sold with a simple qualitative Discussion and Conclusions argument that the more diverse the product line, the greater the number of affi liated At a time when hospital management is physicians and, therefore, the greater the becoming more complex, market mix and demand for each of the hospital’s services. Hospital Diversifi cation Strategy 11

AU: Is One might speculate that small-scale The evolving preference for salaried posi- there a gambles with diversifi cation require tions among younger physicians is changing word missing less-than-average managerial patience to see the hospital medical staff dynamic. One ele- here? The the project to conclusion because the maxi- ment missing from my study is the manner in previous phrase used mum regret (loss) is low, and large-scale gam- which physicians interact with management “mana- bles may require less-than-average patience and trustees. If the major hospital serving the gerial patience.” because the initial commitment has been so poor was to close in a small city or town, It’s not large that management fears the long-run con- one could imagine a range of responses. In clear when sequences of timidly “cutting out losses” and our area one hospital acted as a conservative referring to the pulling back on diversifi cation (retrenching). status quo “defender.” A second hospital in large-scale As more fi nancially distressed fi rms continue the area was characterized as the experimen- gambles who has to to take more risks, this will reconfi gure the tal playful “prospector” hospital willing to be patient. historic association between risk and return. make small-scale gambles to improve their Future researchers should explore whether bad fi nancial position. The third hospital hospital trustees, bolstered by rising consumer labeled itself as an “entrepreneurial-analyst” demand for care, seek to build a more-bal- fi rm willing to make only a few large-scale anced corporate portfolio of service offerings. commitments to new market niches and shel- Health care is the largest sector in the US tering their substantial fi nancial reserves. economy, approaching 19 percent of gross Future research should explore whether domestic product (GDP). Risk is the critical the hospital-based physician community issue for all sectors in our economy. Excessive splits into two basic divisions: (1) the diag- risk selection has been called into question for nostic division containing equal numbers of bankers and large investors in recent years. prospectors and defenders; and (2) the ther- The quantitative results of this small study apeutic division consisting of a multitude must be regarded as tentative. However, the of defenders and a few entrepreneurial ana- results suggest a potentially interesting inter- lysts willing to “convince management to AU: Do action between planning, diversifi cation, and make that one big-scale bet.” The hospital you mean that hospi- competition. The fi nancial ability and desire that grows through 2025 may be the insti- tals have of a hospital to diversify might change by fi ts tution with the vision to follow the entre- overdiver- and starts. Hospital diversifi cation behavior preneurial analysts into health promotion, sifi ed so they must may be controlled by two confl icting motiva- geriatric daycare, home health care, and reduce tions, growth and retrenchment. Some hos- service delivery keyed to particular human service lines and pitals face the risk of having over-diversifi ed problems, such as alcoholism, hypertension, depart- and reducing service-lines and departments. marital dysfunction, and eating disorders. ments to cut costs? Conversely, other managers may believe that The odds against diversifi cation are high This “things must get worse before they can get in many regulated markets. Indeed, some sounds as better” and move into newly vacated mar- hospitals in moderately good fi nancial though reducing kets as they poach enhanced market-share. health relative to their neighbors will fi nd service The subject of manager risk selection and the odds for success in broadly diversifying is part of overdiver- co-specialization of complementarities is are unattractive. For the fi nancially well- sifying. a popular topic in the burgeoning fi eld of managed or fi nancially destitute institution organizational theory and behavior.13 teetering on the edge of a catastrophe, the 12 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

urge to gamble a little or a lot, respectively, necessity impedes innovation or breed rapid may lead to diversifi cation of services. In diversifi cation because the operating margin the fi rst case, the management and trustees provides the venture capital for the hospital might be more interested in “having some to experiment. In a dynamic situation, the fun” and “doing some good” by building fi rst derivative of competition (increasing or a balanced corporate portfolio of service decreasing) and fi nancial health (improving offerings. In the second case, the manage- or deteriorating) may be more crucial to the ment might have to convince the trustees diversifi cation issue than the absolute level that “if we want to stay afl oat in this reg- of either competition or fi nancial health. ulated environment, we better fi nd some The hospital is an important civic enter- money-making service areas to subsidize prise in our society. Hospitals with broad our money losers.” Incentive bonus com- and lateral interests in the health and well- pensation plans are one way to assure that being of their consumers are likely to grow the hospital sector can attract high-quality through diversifi cation into other medical risk takers. Private hospitals initiated pay- arenas rather than expand into unrelated for-performance (P4P) payment incentives lines of business. The popularity of hospital 20 years before the current Medicare P4P restructuring plans has to be tempered by program. Forward-thinking managers will concern for direct and indirect administra- consider diversifi cation and a productivity- tive cost. Hospitals that place too much faith enhancement P4P internal payment formula in short-run fi nancial strains might profi t to improve the hospital’s fi nancial posi- from observation of the retail sector. Over- tion. Managers that reposition their service reliance on retrenchment (closure of prod- product-line mix, and adjust incentives, will uct lines) can drive the organization into the capture market share. ground. In 2012 an Apple executive (Ron Most hospital managers believe that com- Johnson) took over as CEO of J.C. Penny, petition prompts diversifi cation. The pic- closed 30 percent of product lines (khaki ture is most complex when one factors in pants and pantyhose, for example), and was fi nancial health and institutional capacity. fi red when revenues declined 22 percent. A hospital with a good operating margin can Retrenchment to fi nance high-brow diversi- either breed complacency because lack of fi cation can cost the CEO his job.

REFERENCES

1. Eastaugh, S, “Hospital Cost and Speciali- Emergency Departments, Washington, DC: zation,” Journal of Health Care Finance, ACEP (2013). 35(4):26–32 (2011). 5. Eastaugh, S, Health Care Finance and Econom- 2. Eastaugh, S, “Response to Market Reform,” ics, Sudbury, MA: Jones and Bartlett (2006). International Journal of Applied Economics, 6. Brekke, K, Cellini, R, “Competition and Qual- 8(1):1–16 (2011). ity in Health Care Markets,” Journal of Health 3. Feldstein, M, Hospital Costs & Health Insur- Economics, 29(4):508–523 (2010). ance, Cambridge, MA: Harvard (1999). 7. Courtemanche, C, “Does Competition Affect 4. ACEP, American College of Emergency Phy- Hospital Output,” Journal of Health Econom- sicians, ACEP Annual Survey of Hospital ics, 29(5):765–773 (2010). Hospital Diversifi cation Strategy 13

8. Eastaugh, S, “Hospital Productivity and Infor- 11. Eastaugh, S, “Health Information Technology mation Technology,” Journal of Health Care Impact on Productivity,” Journal of Health Finance, 36(4):27–37 (2010). Care Finance, 39(2):64–80 (2012). 9. Maddala, G, Introduction to Econometrics, 12. Eastaugh, S, “Total Cost,” Healthcare Finan- New York: Macmillan (2000). cial Management, 67(2):66–71 (2013). 10. Greenspun, H, Bercik, W, “Cost-outcomes 13. Clark, J, Huckman, R, “Specialization in the Focus is Essential,” Healthcare Financial Man- Hospital Industry,” Management Science, agement, 67(2):96–102 (2013). 58(4):708–722 (2012). Public Hospitals in Peril: Factors Associated with Financial Distress

Zo Ramamonjiarivelo, Robert Weech-Maldonado, Larry Hearld, and Rohit Pradhan

As “safety net providers,” public hospitals have played a major role in health care delivery, especially in serving the indigent and the uninsured. For several decades, public hospitals have been operating in a challenging environment, and some of them have experienced fi nancial diffi culties. The purpose of this study was to explore the organizational and environmental factors associated with public hospitals’ fi nancial distress. This study used a national sample of public hospitals based on longitudinal panel data from 1997 to 2009, resulting in a sample size of 7,257 hospital-year observations. The Altman Z-score method was applied to assess hospitals’ fi nancial condition. The signifi cant fi ndings from a random- effects logistic regression model with state and year fi xed-effects indicated that higher Medicare HMO penetration was associated with fi nancial distress. Organizational variables such as health network, size, occupancy rate, and outpatient mix decreased the odds of fi nancial distress; and membership in a multihospital system increased the odds of fi nancial distress. Key words: fi nancial distress, public hospitals, Altman Z-score, munifi cence, dynamism, complexity

Public hospitals play an important role in Consequently, some public hospitals have the health care delivery system. The prin- experienced fi nancial distress resulting in cipal mission of public hospitals is to pro- acquisitions by other hospitals, ownership vide healthcare for the indigent, the needy, status conversions, and even closures.5 As and the uninsured; therefore, they act as the such, the number of US public hospitals “provider of last resort” or “safety net” for declined from 1,761 to 1,105 (35 percent the community.1 In 2008, approximately 25 decline) between 1975 and 2008.6 percent of patients served by public hospitals were Medicaid patients compared to 17 percent for private, not-for-profi t hospi- Zo Ramamonjiarivelo, PhD, is an Assistant Profes- tals. In addition, public hospitals provided sor of the Department of Health Administration at more care for the uninsured (8 percent) Governors State University. She can be reached at compared to private not-for-profi t hospi- [email protected]. tals (5 percent).2 Furthermore, 16 percent Robert Weech-Maldonado, PhD, is a Professor & of public hospitals’ operating costs in 2009 L.R. Jordan Endowed Chair of the Department of were uncompensated relative to 6 percent of Health Services Administration at the University operating costs of all hospitals in the United of Alabama at Birmingham. He can be reached at [email protected]. States.3 Larry Hearld, PhD, is an Assistant Professor of the Although public hospitals carry the heav- Department of Health Services Administration at the iest burden in serving some of the most University of Alabama at Birmingham. He can be vulnerable populations,4 they have the least reached at [email protected]. operational fl exibility. Public hospitals rely Rohit Pradhan is an Assistant Professor of the largely on the availability of funds from Department of Health Policy and Management at the government entities, which may decrease University of Arkansas for Medical Sciences. He can sharply during economic crises. Further- be reached at [email protected]. more, they have limited ability to raise J Health Care Finance 2014; 40(3):14–30 additional funds from the capital markets. Copyright © 2014 CCH Incorporated 14 Public Hospitals in Peril: Factors Associated with Financial Distress 15

Conceptually, a fi rm can be considered in hospitals are open systems that interact with fi nancial distress when there is a high risk and are affected by their external environ- of bankruptcy or failure, or when “the liq- ments, it is important to include both organ- uid assets of the fi rm are not suffi cient to izational and environmental factors in the meet the current requirements of its hard prediction of fi nancial distress.15 contracts.”7 Hard fi nancial contracts refer to This study contributes to the hospital the fi rm’s obligations to its creditors such fi nancial distress literature in several ways. as bondholders, suppliers, and employees.8 First, this is the fi rst empirical study that There is no agreed-upon operational defi ni- examines the impact of both organizational tion of fi nancial distress. Most prior studies and environmental factors on public hospi- have used multiple fi nancial ratios in assess- tals’ fi nancial distress based on a nationally ing fi nancial distress or bankruptcy/failure representative sample and longitudinal data. risk.9 McCue defi ned fi nancial distress as Studying fi nancial distress of public hospi- having both an average negative cash fl ow tals is important given the unique operat- to total beds (cash fl ow ratio) and negative ing and fi nancial environment faced by this net income to total beds (profi tability ratio) group of hospitals. As such, this study can over a two-year period.10 Kim defi ned fi nan- provide important managerial and policy cial distress as having a negative fi nancial insights. Second, this is the fi rst study of strength index (FSI), based on a hospital public hospitals’ fi nancial distress applying performing worse than the industry median the Altman Z-score model. This model has on four fi nancial indicators: total margin, been widely used in the fi nance literature to days cash on hand, percent debt, and age of determine fi nancial distress, and it is calcu- the physical facility.11 lated based on a weighted composite score Relatively few studies have used empiri- of four fi nancial ratios: liquidity, fi nancial cally derived thresholds to determine leverage, profi tability, and capital struc- fi nancial distress of hospitals. Bazzoli and ture ratios. Thus, it takes into account “the Andes, based on a three-year average, interrelationships of many different fi nan- identifi ed fi nancially distressed hospitals cial aspects, similar to that of a balanced as those with at least six of eight fi nancial scorecard.”16 ratios below the median of hospitals with a BBB credit rating, which implies a high Conceptual Framework and Hypotheses risk for bond default.12 On the other hand, Langabeer13 examined the fi nancial distress This study applies the resource depend- of teaching hospitals using the Altman’s ence theory (RDT), which posits that “the Z-score model,14 empirically derived using key to organizational survival is the abil- multiple discriminant analysis of the bank- ity to acquire and maintain resources.”17 ruptcy of service and retail fi rms. Resources are viewed as the inputs The purpose of this study is to investigate that organizations need to produce out- the impact of environmental and organiza- puts and the environment refers to the tional factors on public hospitals’ fi nancial “organization’s source of inputs and sink distress, using a more comprehensive meas- of outputs.”18 In other words, the focal ure of fi nancial distress. Because public organization’s environment includes other 16 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

entities from which it acquires resources Prior empirical studies have found that and to which it sells products and services. hospitals operating in more munifi cent This defi nition implies that organizations environments have better fi nancial per- are neither self- suffi cient nor self-reliant;19 formance. Munifi cence, defi ned as the however, the availability of resources is proportion of people aged 65 and older in uncertain. Uncertainty is one of the char- the hospital’s market, was found to have acteristics of the environment, and it refers a positive association with fi nancial per- to the fl uctuation of resource availability formance.24 Additionally, Kim found that and the magnitude of challenges the organ- environmental scarcity in terms of a high ization faces in acquiring key resources.20 unemployment rate increased the prob- Thus, scarcity of resources, combined with ability of fi nancial distress among hospi- their uncertain supply, makes resource tals located in metropolitan statistical areas acquisition a critical element to organiza- (MSAs).25 tional survival. Organizations operating in more munifi - In addition, RDT posits that organiza- cent environments tend to exhibit higher tions seek to maximize their power and fi nancial performance because they have independence relative to other organiza- greater access to resources. As a result, tions by acquiring critical resources. Organ- organizations can focus their efforts on izations can survive in their environment increasing productivity, rather than spend- as long as they obtain access to resources. ing resources to acquire other resources. As a result, organizations are not passive, Therefore, we hypothesize that: they can transform the environment; the various strategic moves that organizations Hypothesis 1: Public hospitals operat- undertake to acquire resources will change ing in more munifi cent environments the environmental landscape.21 Three are less likely to experience fi nancial dimensions of the external environment distress than public hospitals operat- have become commonly used in empirical ing in less munifi cent environments. studies: environmental munifi cence, envi- Environmental Dynamism ronmental dynamism, and environmental complexity.22 Environmental dynamism is the extent to which changes in the environment are unpredictable and occur more frequently.26 Environmental Munifi cence In health care, such changes include tech- Environmental munifi cence refers to the nological change, rapid growth in the size abundance of resources in the environment and number of organizations operating in to support the operational needs of organiza- the same industry,27 and changes in regulations. Munifi cence has been operationalized tions and policies. For example, the dyna- as per-capita income, overall population mism of the health care environment may growth, growth rate of the elderly popula- be affected by the uncertainties with respect tion, growth in total sales, growth in total to the implementation of the Patient Protec- employment, and the number of physicians tion and Affordable Care Act (PPACA) of in the county.23 2010.28 Public Hospitals in Peril: Factors Associated with Financial Distress 17

Prior studies have found that fi rms to increase the organization’s effi ciency and operating in more dynamic environments productivity, if the environment was less exhibit worse fi nancial performance com- complex. Therefore, we hypothesize that: pared to those operating in more stable environments.29 The high rate of change Hypothesis 3: Public hospitals operat- associated with dynamic environments can ing in more complex environments are affect the process of planning and resource more likely to experience fi nancial dis- acquisition. Furthermore, the fl uctuation tress than public hospitals operating in of resource supply due to environmental less complex environments. change makes it more diffi cult to acquire resources. Therefore, we hypothesize that: Organizational Size

Hypothesis 2: Public hospitals operat- Organizational size has been found to ing in more dynamic environments are be positively associated with fi nancial per- 35 more likely to experience fi nancial dis- formance. Similarly, larger hospitals have tress than public hospitals operating in been found to have lower risk of fi nancial 36 more stable environments. distress compared to smaller hospitals. Larger organizations can reduce costs Environmental Complexity through economies of scale.37 Furthermore, Environmental complexity refers to the larger organizations may have greater bar- level of heterogeneity.30 An environment is gaining power over suppliers, which can 38 heterogeneous when it contains a large num- result in lower supply costs. In addition, ber of different types of entities that the focal large organizations are able to accumulate organization needs to interact with to acquire slack resources, which fi rms can set aside for 39 resources.31 Prior studies have found that future environmental challenges. There- higher environmental complexity is asso- fore, we hypothesize that: ciated with lower fi nancial performance.32 Hypothesis 4: Larger public hospitals Brecher and Nesbitt found that higher com- are less likely to experience fi nancial petition was associated with lower fi nancial distress than smaller public hospitals. performance among not-for-profi t hospitals in New York.33 Similarly, Kim reported that higher environmental complexity, in terms Teaching Status of higher competition and greater HMO Although teaching hospitals generally penetration, was associated with increased operate in complex environments and often fi nancial distress among private not-for- have to provide higher amounts of uncom- profi t hospitals.34 pensated care, affi liation with a medical When there are many organizations in school may confer several benefi ts. Hos- the same industry that compete for the same pitals affi liated with medical schools have key resources, it is more diffi cult to acquire the reputation of providing higher quality these resources. In addition, competing for care for complex procedures. Furthermore, key resources requires the consumption of they deliver highly specialized services and other resources that could have been used are well respected in their communities.40 18 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

In many cases, teaching hospitals have a the most validated and widely accepted data monopoly of ground breaking technology for hospital fi nancial analysis.46 The LAUS to treat specifi c conditions.41 Good reputa- data fi le contains estimates of monthly and tion and a monopoly of sophisticated pro- annual averages of total employment, total cedures have permitted teaching hospitals unemployment, and unemployment rates at to charge fees up to 10 percent higher than various geographical levels including metro- nonteaching hospitals.42 Besides, teaching politan areas, cities, census regions and divi- hospitals oftentimes have larger endowment sions, as well as counties. funds than nonteaching hospitals,43 which Sample can serve as fi nancial protection against adversities. Tennyson and Fottler44 and You- This study used a national sample of gov- nis and Forgione45 found that teaching status ernment, nonfederal, acute care, general, was associated with higher fi nancial perfor- and surgical hospitals in the US. To derive mance. Therefore, we hypothesize that: the analytic sample, several exclusion criteria were applied. First, hospitals that were Hypothesis 5: Teaching public hospi- either converted to a skilled nursing facility tals are less likely to experience fi nan- (N= 6), an ambulatory care facility (N=1), cial distress than nonteaching public or a critical access hospital (CAH) (N=520) hospitals were excluded. Since CAHs have a different reimbursement policy, it was deemed appro- Methods priate to exclude them for this study. Second, we excluded hospitals that were acquired or This study used longitudinal panel data merged (N= 9) during the study period. Third, from 1997 to 2009 of US public hospitals. hospitals without complete fi nancial reports during certain years of the study period were Data Sources excluded from the analytic sample (N= 43). Four data sources were used: (1) the Amer- Therefore, the fi nal analytic sample con- ican Hospital Association (AHA) Annual sisted of 608 public hospitals with a total of Survey, (2) the Bureau of Health Profession’s 7,257 hospital-year observations. Area Resource File (ARF), (3) the Medicare Cost Report (MCR) from the Centers for Variables Medicare & Medicaid Services, and (4) the The dependent variable was dichotomous Local Area Unemployment Statistics (LAUS) (1 = yes; 0 = no) that indicates whether a from the Bureau of Labor Statistics. The AHA hospital is in fi nancial distress in a given data fi le contains organizational information year. Financial distress is an indicator of the such as ownership status, number of hospital fi nancial health of the hospital, and it can be beds, teaching status, multihospital system used to predict the likelihood of a hospital affi liation, and information with respect to meeting its debt obligations. The Altman the number of clinical and nonclinical staff. Z-score model, designed to detect fi nancial The ARF data fi le contains demographic distress, was used to determine whether a and economic information on counties. The public hospital was in fi nancial distress.47 MCR data fi le contains fi nancial data; it is For this study, the Altman Z-score model Public Hospitals in Peril: Factors Associated with Financial Distress 19

designed for service and retail fi rms applied X4 was originally defi ned as the ratio of in Langabeer48 was used to estimate public book value of equity to the book value of hospitals’ fi nancial condition. The discrimi- debt. This ratio represents the capital struc- nant function is formulated as follows: ture of the hospital; it is the extent to which the assets are fi nanced by debt and equity,

Z = 6.56 X1 + 3.26X2 + 6.72X3 + 1.05X4 respectively. Public hospitals do not have equity, but the book value of equity or fund where: balance can be derived by subtracting total X = Net Working Capital / Total Assets 1 liabilities from total assets. In this study, X4 X2 = Net Assets / Total Assets is defi ned as the ratio of total fund balance to

X3 = Excess Revenue over Expenses / total liabilities. Total Assets Following Langabeer’s approach,51 we

X4 = Fund balance / Total liabilities classifi ed a hospital as being in fi nancial distress if its Z-score is less than 1.1; the hos- X1 is defi ned as the ratio of net working pital is not in fi nancial distress if its Z-score capital to total assets, and it is a measure of is greater than 1.1. Independent variables liquidity. Net working capital is defi ned as include environmental munifi cence, envi- the difference between current assets and ronmental dynamism, environmental com- current liabilities. Altman49 suggests this plexity, size, and teaching status. ratio is the most valuable liquidity ratio rela- Environmental munifi cence (hypothesis 1) tive to the current ratio and quick ratio. was operationalized with four county-level X2 was originally defi ned as the ratio of variables: per capita income, unemployment retained earnings to total assets, and it is a rate, percentage of people who are 65 years measure of fi nancial leverage. This study old or older, and the number of active physi- used net assets (total assets minus total lia- cians per 1,000 persons in the county. bilities) instead of retained earnings, given Environmental dynamism (hypothesis 2) that public hospitals do not accumulate earn- was measured as the yearly change in the ings from prior years. It measures the extent unemployment rate at the county level. The to which one dollar’s worth of total assets is unemployment rate refl ects the availability fi nanced by debt and net assets, respectively. of resources and its fl uctuation refl ects the Higher net assets to total assets ratio means variation of resources in the environment. that the hospital has more assets than debts. Environmental complexity (hypothesis 3) X3 was originally defi ned as the ratio of was operationalized with four variables: mar- earnings before income taxes to total assets, ket concentration, excess capacity, Medicare and it is a measure of both profi tability and HMO penetration, and metropolitan ver- productivity of the fi rm assets.50 Since public sus nonmetropolitan area. The Herfi ndahl- hospitals do not pay taxes, this study used Hirschman index (HHI) has been widely excess revenue over expenses, defi ned as used to measure market concentration.52 total revenues minus total expenses, in place It is defi ned as the sum of squared market of earnings before income taxes. This ratio shares (acute-care patient days for individual measures how much profi t a hospital makes hospital/total acute-care patient days of all out of one dollar worth of total assets. the hospitals in the county).53 HHI represents 20 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

perfect competition when it scores 0 and the essary to convert outpatient visits to their presence of a monopoly when it scores 1. inpatient days equivalent. Research has Excess capacity and Medicare HMO pen- suggested that outpatient services are less etration have also been used to measure the resource-intensive relative to inpatient ser- degree of environmental complexity in terms vices; resources consumed to provide one of competition.54 Competition is stronger outpatient service are equivalent to one-third when there is excess capacity. Hospitals of the resources consumed to provide a ser- compete for patients to fi ll the empty beds. In vice for one inpatient day. 56 Therefore, the this study, excess capacity was measured in inpatient days equivalent for outpatient vis- terms of the average number of unoccupied its was obtained by dividing total outpatient beds per hospital in the county. When there visits by three, and total equivalent inpatient is a greater Medicare HMO penetration, days was obtained by summing inpatient Medicare HMOs have power over hospitals days equivalent for outpatient visits and total with respect to price negotiations. Medicare inpatient days. Outpatient mix was defi ned as HMO penetration was measured as the num- inpatient days equivalent for outpatient visits ber of Medicare HMO enrollees relative to divided by total equivalent inpatient days. the total number of Medicare eligibles in the Occupancy rate refl ects the level of inpa- county. Variable metropolitan versus non- tient services utilization, and it was meas- metropolitan area was coded based on the ured at the hospital level as the ratio of total Rural-Urban Continuum Codes of 2003. It inpatient days to the total number of beds was coded 1 if the hospital is located in a times 365. Low occupancy rate has been metropolitan county and coded 0 if the hos- consistently associated with higher risk of pital is located in a nonmetropolitan county. hospital fi nancial distress.57 Hospital size (hypothesis 4) was measured Payer mix refl ects the proportions of as the total number of beds in the hospital. Medicare and Medicaid patients. Medicare Teaching status (hypothesis 5) was coded and Medicaid are important sources of rev- “1” if the hospital did not have a teaching sta- enues for public hospitals, and can contrib- tus and “0” if it had a teaching status. Hospi- ute to public hospital survival. Medicare mix tals can engage in medical education in three was measured as total Medicare inpatient different, but not mutually exclusive, ways: days divided by total inpatient days, and (1) membership of the Council of Teaching Medicaid mix as total Medicaid inpatient Hospitals and Health Systems (COTH); (2) days divided by total inpatient days. affi liation with a medical school; and (3) The American Hospital Association provision of residency programs. The hospi- defi nes a multihospital system as “two or tal was defi ned as having a teaching status more hospitals owned, leased, sponsored, if it met one or more of the above criteria.55 or contract managed by the central organi- The following variables were used as con- zation.”58 Multihospital membership is a trol variables: outpatient mix, occupancy dichotomous variable coded as “1” if the rate, payer mix, multihospital system mem- hospital was a member of a multihospital bership, participation in a health network, system and “0” if the hospital was not. and operating under contract management. A network is defi ned as “a group of hos- To determine outpatient mix, it was nec- pitals, physicians, other providers, insur- Public Hospitals in Peril: Factors Associated with Financial Distress 21

ers and/or community agencies that work the time varying errors.63 The analysis was together to coordinate and deliver a broad conducted using xtlogit function in STATA spectrum of services to their community.”59 version 11. Participation in a health network is a dichotomous variable coded as “1” if the hospital Results was engaged in such relationship and “0” if Figure 1 summarizes the descriptive sta- the hospital was not. tistics of the variables. There were a total of A hospital is under a contract management 7,257 hospital-years, of which 13 percent (N when an outside organization is fully respon- = 941 hospital-years) experienced fi nancial sible for its day-to-day operation.60 Contract distress. Twenty two percent of the public management is a dichotomous variable coded hospitals had teaching status, 32 percent were as “1” if the hospital was under contract man- members of multihospital systems, 15 per- agement and “0” if the hospital was not. cent were under contract management, and 25 percent were involved in health networks. Analysis About half of the hospitals were located in This study used random-effects logistic metropolitan areas and the other half located regression with year and state fi xed-effects in nonmetropolitan areas. We checked for to control for time-invariant unobservable potential multicollinearity by examining the factors that are thought to be uncorrelated correlations among independent variables, with each of the independent variables.61 and none of the correlations was above 0.80.64 Random-effects models are also appropriate Figure 2 summarizes the results of the for multilevel data structures such as public random-effects logistic regression analysis. hospitals nested within the county. In addi- Hypothesis 1 was not supported; environ- tion, panel data have a multilevel structure mental munifi cence was not associated with because the repeated measurements over a the fi nancial distress of public hospitals. certain period of time of each observation Hypothesis 2 was not supported; greater are clustered.62 The random-effects logistic environmental dynamism was not associ- regression model is described as: ated with higher odds of fi nancial distress. Hypothesis 3 was partially supported; greater β β * Log (Pit /1-Pit) = 0t + 1 Munifi cenceit environmental complexity, in terms of higher β * + 2 Dynamismit Medicare HMO penetration, was associated β * + 3 Complexityit with 5 percent greater odds of experiencing β * β * + 4 Sizeit + 5 Teaching fi nancial distress (OR = 1.05; p ≤ .000). β * Statusit + 6 Control Hypothesis 4 was supported. Larger hos- Variablesit + ai + uit pitals had signifi cantly lower odds of experiencing fi nancial distress compared to smaller

In which, Pit represents the probability hospitals (OR =.997; p ≤ .05); however, the that hospital “i” is in fi nancial distress in effect of size was negligible as indicated by β0t year “t”. is the intercept for year “t”, ai an odds ratio close to one. Hypothesis 5 was represents the time-invariant unobserved not supported; teaching status was not asso- factors that are uncorrelated with each of ciated with a lower likelihood of fi nancial the independent variables and uit represents distress. 22 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 1. Descriptive Statistics of All Variablesa

N Mean (%) Std. Dev. Dependent variable—financial distress Not in fi nancial distress 6,316 (87.03) In fi nancial distress 941 (12.97) Independent variables

Per capita income 7,225 26,633.04 8,760.17 Unemployment rate 7,238 5.74 2.64 Proportion of population ≥ 65 7,239 0.14 0.04 Physicians per 1,000 population 7,239 1.86 1.98 Medicare HMO penetration 7,249 9.41 13.59 Excess capacity 7,239 55.87 36.70 Herfi ndahl-Hirschman Index 6,574 0.85 0.31 Metropolitan vs. nonmetropolitan area Metropolitan area 3,595 (49.59) 3,654 (50.41) Nonmetropolitan area Change in unemployment rate 7,232 0.06 0.24 Hospital beds 7,221 177.00 197.00 Teaching status Teaching 1,594 (21.97) No teaching 5,660 (78.03) Control variables

Occupancy rate 7,155 0.56 0.19 Medicare mix 7,256 0.45 0.20 Medicaid mix 7,256 0.23 0.18 Outpatient mix 7,256 0.43 0.25 System membership Nonmember 4,953 (68.25) Member 2,304 (31.75) Contract Management No contract 6,138 (84.58) Contract 1,119 (15.42) Health network No network 5,465 (75.31) Network 1,792 (24.69)

a Hospital-year is the unit of analysis, N accounts for repeated observations Public Hospitals in Peril: Factors Associated with Financial Distress 23

Figure 2. Results of Random-Effects Logistic Regression with Year and State Fixed-Effectsa

Financial distress (N = 6446)b Odds Ratio Standard Error 95% CIc

Munificence Per capita income 1.01 0.2 0.97 1.06 Unemployment rate 1.07 0.06 0.98 1.16 Proportion of population ≥ 65 1.04 0.04 0.97 1.13 Physicians/1000pop 0.99 0.11 0.80 1.24 Complexity Herfi ndahl-Hirschman Index 0.66 0.16 0.34 1.31 Excess capacity 0.99 0.004 0.99 1.00 Medicare HMO penetration 1.05*** 0.01 1.02 1.07 Metropolitan vs. non-metropolitan area 1.31 0.53 0.59 2.92 Dynamism Change in unemployment rate 0.85 0.34 0.39 1.86 Organizational factors Hospital beds 0.997** 0.001 0.995 1.00 Teaching status 0.82 0.25 0.45 1.51 Control variables Occupancy rate 0.23** 0.12 0.08 0.67 Outpatient mix 0.34* 0.19 0.12 1.01 Medicare mix 0.99 0.53 0.35 2.85 Medicaid mix 0.59 0.33 0.20 1.77 System Membership 3.49*** 0.80 2.23 5.47 Contract management 1.14 0.24 0.75 1.73 Health network 0.71* 0.12 0.49 1.04 Overall Chi-square test 199.89***

*p ≤ 0.1 **p ≤ 0.05 ***p ≤ 0.000 a Year and state dummy variables were included in analysis b N represents hospital-years c Confi dence Interval 24 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Our results also found that some of the that different entities own public hospitals. control variables had statistically signifi cant Unlike private hospitals, public hospitals associations with fi nancial distress. Com- can be owned by a state, a city, or a county; pared to stand-alone hospitals, independent however, the environmental variables are hospitals that became affi liated with multi- measured at the county level. Therefore, hospital systems had 3.49 greater odds of even if the county in which the hospital experiencing fi nancial distress (p ≤ .000) operates is less munifi cent, more dynamic, compared to stand-alone hospitals. Further, or more complex, these situations might not as occupancy rate increased, the odds of directly impact the fi nancial situation of the being in fi nancial distress decreased (OR public hospital if the state or the city that = .22; p ≤ .05). Variables outpatient mix and owns that hospital is still able to fi nance the health network were also negatively associ- hospital. ated with fi nancial distress but these associa- We also found that larger hospitals are tions were marginally signifi cant (OR = .34 less likely to experience fi nancial distress, and .71 respectively; p ≤ .10). which corroborates the results from prior studies suggesting that organizational size Discussion has a positive impact on fi nancial perfor- The purpose of this study was to investi- mance.66 Several factors such as economies gate the environmental and organizational of scale,67 accumulation of slack resources,68 factors associated with public hospitals’ and bargaining power over suppliers69 may fi nancial distress based on resource depend- be among the reasons for the positive associ- ence theory. Our fi ndings suggest that a more ation between organizational size and supe- complex environment in terms of higher rior fi nancial performance. Medicare HMO penetration is the major Our study also found that hospitals that environmental factor associated with pub- are part of multihospital systems are more lic hospital fi nancial distress. Our fi ndings likely to be in fi nancial distress compared are consistent with Kim who reported a sig- to stand-alone hospitals. This fi nding is nifi cant positive association between HMO counterintuitive. From an RDT perspec- penetration and fi nancial distress among not- tive, multihospital systems provide several for-profi t hospitals.65 A greater market share resources to their members such as fi nancial of Medicare benefi ciaries provides Medicare support, technology, and expertise, and they HMOs with stronger bargaining power that may benefi t from economies of scale; how- they can leverage to more aggressively nego- ever, it is possible that multihospital mem- tiate prices. Therefore, hospitals may need bership is endogenous to fi nancial distress. to be highly effi cient when they provide ser- Public hospitals in fi nancial distress might vices for Medicare HMO enrollees. seek affi liation with multihospital systems Except for Medicare HMO penetration, to solve their fi nancial problems.70 On the we did not fi nd any signifi cant association other hand, the acquisition of public hospi- between fi nancial distress and the other tals in fi nancial distress could be a strategy environmental variables that measured for multihospital systems to expand their munifi cence, dynamism, and complexity. market share and strengthen their competi- One reason for this fi nding may be the fact tive position. Public Hospitals in Peril: Factors Associated with Financial Distress 25

Participation in health network is found evaluating public hospitals’ fi nancial condi- to be negatively associated with fi nancial tion; the combination of the most important distress. This fi nding is consistent with prior fi nancial ratios into one discriminant func- studies showing that network membership tion facilitates the evaluation of the organi- was positively associated with hospital zation’s fi nancial condition. Furthermore, fi nancial performance.71 Network participa- the Altman Z-score could be used as a mana- tion is an example of an interorganizational gerial tool for regularly scheduled “stress relationship that health care organizations tests,”76 and consequently intervene before it initiate to have better access to resources.72 is too late. In health care, such resources include fi nan- In addition, our fi nding that participation cial and human resources, legal support, in a health network is negatively associated knowledge, technologies, and the capabil- with fi nancial distress provides evidence that ity to deliver high quality care, among oth- participation in a health network is a strat- ers.73 Additionally, hospitals participating egy to enhance public hospitals’ fi nancial in networks may enhance their image and situation. Health networks might provide competitive positions that they can lever- additional resources that reduce public hos- age in negotiating contracts with managed pitals’ operating costs and prevent fi nancial care organizations. Moreover, health care distress. It also appears that higher Medicare networks increase participants’ bargain- HMO penetration is associated with a higher ing power with suppliers and purchasers; risk of fi nancial distress. Public hospitals and they enhance economies of scale and in such markets may need to focus on effi - scope leading to cost savings and hospi- ciency to ensure competitiveness or should tals’ effi ciency and effectiveness.74 Indeed, actively seek other sources of revenues. It is the strongest motive that leads health care clear that managed care imposes additional organizations to participate in a network burdens on public hospitals and a proactive is the expectation of higher fi nancial per- management style is necessary to avoid its formance and enhanced effi ciency.75 Fill- negative effects. ing more hospital beds also decreases the Our fi nding that larger hospital size, odds of fi nancial distress. It is important for higher occupancy rate, and greater outpa- public hospitals to attract more patients for tient mix are associated with lower odds of increased patient revenue. fi nancial distress offers additional managerial insights. Managers may consider that Managerial and Policy Implications in some markets mergers of public hospi- Hospitals play an important role in tals may be helpful in confronting fi nancial the health care delivery system yet they troubles. In addition, public hospitals might are highly resource intensive. Monitoring consider stronger marketing strategies to the fi nancial situation of hospitals should be attract more patients to fi ll empty beds. Fur- among the major responsibilities of a man- thermore, the fi nding that greater outpatient agement team; however, the Altman Z-score mix was negatively associated with fi nancial has not been widely applied in health care distress makes the provision of outpatient management and health services research. services an attractive strategy to health care The Altman Z-score can serve as a tool in providers.77 26 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

This study also has some policy implica- to the indigent, including information on pri- tions. Since public hospitals serve a larger vately insured and the uninsured could have social purpose rather than pure profi ts, it provided additional insights. Second, the may be important to increase fi nancial sup- operationalization of environmental dyna- port for public hospitals such as the dispro- mism presented another limitation to this portionate share hospital (DSH) program; study. Dynamism, which measures the fl uc- however, PPACA plans to cut DSH payment tuation of the resources in the environment, by $18 billion over a seven-year period from is best measured as the yearly change in the 2014.78 This cut was decided under the argu- measure of an environmental variable such ment that PPACA will reduce the number of as the yearly change in size of the county uninsured to approximately 23 million by population, or the yearly change in the level 2019 and consequently reduce the amount of the county per capita income; however, of uncompensated care.79 Further research the ARF variables do not contain complete is needed to examine the consequences of data that cover all the years in this study. reduction of DSH payments on public hos- That is why dynamism was measured by pitals. An increase in the insured population the yearly change in county unemployment may steer patients away from public hospi- rate. Unemployment rate was the only envi- tals resulting in further deterioration of their ronmental variable with complete data for fi nancial performance. all the years in this study. Finally, the study did not differentiate among public hospitals Limitations by state, county, and city ownership. This Our study has some limitations. First, the classifi cation could be advantageous, given unavailability of data with respect to the pro- that the fi nancing capabilities and policies of portion of privately insured, underinsured, these states, counties, and cities are not the and uninsured patients was a study limita- same. Future research should explore how tion. Because one of the most important the type of public hospital ownership may roles of public hospitals is to provide care infl uence its risk for fi nancial distress.

REFERENCES

1. Gabow, P, “Making a Public Hospital Work,” Characteristics-Report/2009-Public-Hospital- Health Affairs, 20(4):182–187 (2001). Financial-Characteristics-.aspx?FT=.pdf. 2. Fraze, T, Elixhauser, A, Holmquist, L, Johann, 4. Fraze, supra, n.2. California Association of J, Public Hospitals in the United States, 2008, Public Hospitals and Health Systems, “On Statistical Brief # 95, accessed Oct. 30, 2011 the Brink: How the Crisis in California’s Pub- at http://www.hcupus.ahrq.gov/reports/stat- lic Hospitals Threatens Access to Care for Mil- briefs/sb95.jsp. lions”, accessed Dec. 5, 2011 at http://www. 3. National Public Health and Hospital Institute, caph.org/content/upload/AssetMgmt/PDFs/ America’s Public Hospitals and Health Sys- Publications/WhitePaperFINAL.pdf. tems 2009: Results of the Annual NAPH Hos- 5. Bazzoli, JG, Cleverley, WO, “Hospital Bank- pital Characteristics Survey, accessed Nov. 4, ruptcies: An Exploration of Potential Causes 2011, at http://www.naph.org/Main-Menu- and Consequences,” Health Care Manage- Category/Our-Work/Safety-Net-Financing/ ment Review, 19(3):41–51 (1994); Weil, T, Public Hospitals in Peril: Factors Associated with Financial Distress 27

“Privatization of Hospitals: Meeting Divergent an Outcome of Interorganizational Linkages Interests,” Journal of Health Care Finance, and Environmental Conditions,” The Acad- 38(2):1–11 (2011). emy of Management Journal, 30(1):71–90 6. Department of Health and Human Services, (1987). “Health, United States, 2010: With Special 20. Ulrich, D, Barney, JB, “Perspectives in Organ- Feature on Death and Dying,” accessed Nov. izations: Resource Dependence, Effi ciency 10, 2011 at http://www.cdc.gov/nchs/data/ and Population,” The Academy of Manage- hus/hus10.pdf. ment Review, 9(3):(1984). 7. John, K, “Managing Financial Distress and 21. Pfeffer, J, supra, n.15; Ulrich and Barney, Valuing Distressed Securities: A Survey and supra, n.20; Aldrich, HE, Pfeffer J, “Environ- a Research Agenda,” Financial Management, ments of Organizations,” In Inkeles, A. (Ed.) 22(3):60–78 (1993). Annual Review of Sociology, Palo Alto, CA: 8. Id. Annual Reviews (1976). 9. McCue, MJ, “The Use of Cash Flow to Analyze 22. Dess, GG, Beard, DW, “Dimensions of Organ- Financial Distress in California Hospitals,” izational Task Environments,” Administra- Hospital and Health Services Administration, tive Science Quarterly, 29(1): 52–73 (1984); 21(2):223–241 (1991); Bazzoli, JG, Andes, S, Boyd, B, “Corporate Linkages and Organi- “Consequences of Hospital Financial Dis- zational Environment: A Test of the Resource tress,” Hospital and Health Services Admin- Dependence Model,” Strategic Management istration, 40(4):472–95 (1995); Langabeer, J, Journal, 11(6):419–430 (1990). “Predicting Financial Distress in Teaching 23. Dess and Beard, supra, n.22; Alexander, J, Hospitals,” Journal of Health Care Finance, D’Aunno, T, Succi, MJ, “Determinants of 33(2):84–92 (2006); Kim, TH, “Factors Associ- Profound Organizational Change: Choice ated with Financial Distress of Nonprofi t Hos- of Conversion or Closure among Rural Hos- pitals,” The Health Care Manager, 29(1):52 pitals,” Journal of Health and Social Behav- (2010). ior, 37(3):238–251 (1996); Margarethe, FW, 10. McCue, supra, n.9. Bantel, KA, “Top Management Team Turno- 11. Kim, supra, n.9. ver as an Adaptation Mechanism: The Role 12. Bazzoli and Andes, supra, n.9. of the Environment, “ Strategic Management 13. Langabeer, supra, n.9. Journal, 14(7):485–504 (1993); Trinh, HQ, 14. Altman, EI, Corporate Financial Distress and Begun JW, “Strategic Adaptation of US Rural Bankruptcy, New York, NY: John Wiley and Hospitals During an Era of Limited Financial Sons (1993). Resources: A Longitudinal Study, 1983 to 15. Pfeffer, J, “Developing Resource Dependence 1993,” Health Care Management Science, Theory: How Theory Is Affected by its Envi- 2(1):43–52 (1999). ronment. In Smith, K, Hitt, M, (Eds.), Great 24. Brecher, C, Nesbitt, S, “Factors Associ- Minds in Management: The Process of Theory ated with Variation in Financial Condition Development, New York, NY: Oxford Press among Voluntary Hospitals,” Health Services University (2005). Research, 20(3):267–300 (1985); Friedman, 16. Langabeer, supra, n.9. B, Shortell, S, “The Financial Performance of 17. Pfeffer, J, Salancik, G, The External Control Selected Investor-owned and Not-for-profi t of Organizations: A Resource Dependence System Hospitals Before and After Medi- Perspective, New York, NY: Harper and Row care Prospective Payment,” Health Services (1978). Research, 23(2):237–267 (1988). 18. Pennings, J, “The Relevance of Structural 25. Kim, supra, n.9. Contingency Model for Organizational Effec- 26. Begun, JW, Kaissi, A, “Uncertainty in Health tiveness,” Administrative Science Quarterly, Care Environments: Myth or Reality?,” Health 20(3):393–410 (1975). Care Management Review, 29(1):31–19 (2004). 19. Stearns, TM, Hoffman, AN, Heide, JB, “Perfor- 27. Kotha, S, Nair, A, “Strategy and Environment mance of Commercial Television Stations as as Determinants of Performance: Evidence 28 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

from the Japanese Machine Tool Industry,” (2002); Reuter, J, Gaskin, D, “Academic Strategic Management Journal, 16(7):497– Health Centers in Competitive Markets,” 518 (1995); Simerly, RL, Li, M, “Environ- Health Affairs, 16(4):242–252 (1997). mental Dynamism, Capital Structure and 41. Ayanian and Weissman, supra, n.40; Moses Performance: A Theoretical Integration and III, H, Their, S, Matheson, D, “Why Aca- an Empirical Test,” Strategic Management demic Health Centers Survived?,” Jour- Journal, 21(1):31–49 (2000). nal of the American Medical Association, 28. Doherty, RB, “The Certitudes and Uncertain- 293(12):1495–1500 (2005). ties of Health Care Reform,” Annals of Inter- 42. Ayanian and Weissman, supra, n.40. nal Medicine, 152(10):679–682 (2010). 43. Rosko, MD, “Performance of US Teaching 29. Baum, RJ, Wally, S, “Strategic Decision Speed Hospitals: A Panel Analysis of Cost Ineffi - and Firm Performance,” Strategic Manage- ciency,” Health Care Management Science, ment Journal, 24(11):1107–1129 (2003); 7(1):7–16 (2004). Goll, I, Rasheed, A, “The Moderating Effect of 44. Tennyson, DH, Fottler, MD, “Does System Environmental Munifi cence and Dynamism Membership Enhance Financial Performance on the Relationship Between Discretionary in Hospitals?,” Medical Care Research and Social Responsibility and Firm Performance,” Review, 57(1):29–50 (2000). Journal of Business Ethics, 49(1):41–54 45. Younis, MZ, Forgione, DA, “Using Return on (2004). Equity and Total Profi t Margin to Evaluate 30. Dess and Beard, supra, n.22; Begun and Kai- Hospital Performance in the U.S.: A Piece- ssi, supra, n.26. wise Regression Analysis,” Journal of Health 31. Duncan, RB, “Characteristics of Organiza- Care Finance, 31(3):82–88 (2005). tional Environments and Perceived Environ- 46. Pink, GH, Holmes, M, D’Alpe, C, Strunnk, mental Uncertainty,” Administrative Science L, McGee, P, Slifkin, R, “Financial Indica- Quarterly, 17(3):313–327 (1972). tors for Critical Access Hospitals,” accessed 32. Kim, supra, n.9; Brecher and Nesbitt, supra, Nov. 10, 2011 at http://www.fl exmonitoring. n.24; Friedman and Shortell, supra, n.24. org/documents/BriefingPaper7_Financial 33. Brecher and Nesbitt, supra, n.24. Indicators.pdf. 34. Kim, supra, n.9. 47. Langabeer, supra, n.9; Altman, EI, “Finan- 35. Id.; Simerly and Li, supra, n.27; Goll and cial Ratios, Discriminant Analysis and the Rasheed, supra, n.29. Predition of Corporate Bankrupcty,” The 36. Kim, supra, n.9; NORC Walsh Center for Journal of Finance, 23(4):589–609 (1968); Rural Health Analysis, Financially Distressed Altman, EI, “Predicting Financial Distress of Rural Hospitals in Four States, accessed Companies: Revisiting the Z-score and Zeta Jan. 13, 2012 at http://www3.norc.org/NR/ Models,” accessed Nov. 10, 2011 at http:// rdonlyres/99DA930E-4C63-4EBC-B9A7- www.iiiglobal.org/component/jdownloads/ E07148BF4A46/0/WalshCtr2004_Nseries- fi nish/648/5645.html; Almwajeh, O, “Apply- Jan04v2.pdf. ing Altman’s Z-score Model of Bankruptcy for 37. Hall, M, & Weiss, L, “Firm Size and Profi tabil- the Prediction of Financial Distress of Rural ity,” The Review of Economics and Statistics, Hospitals in Western Pennsylvania”, accessed 49(3):319–331 (1967). Jan. 13, 2011 at sunzi.lib.hku.hk/ER/detail/ 38. Porter, M, Competitive Advantage, NY: Free hkul/3517030. Press (1985). 48. Langabeer, supra, n.9. 39. Sharfman, M, Wolf, G, Chase, R, Tansik, D, 49. Altman, EI, “Financial Ratios, Discrimi- “Antecedents of Organizational Slack,” Acad- nant Analysis and the Predition of Corpo- emy of Management Review, 13(4):601–614 rate Bankrupcty,” The Journal of Finance, (1988). 23(4):589–609 (1968); Altman, EI, “Predict- 40. Ayanian, J, Weissman J, “Teaching Hospitals ing Financial Distress of Companies: Revisit- and Quality of Care: A review of the Lit- ing the Z-score and Zeta Models,” accessed erature,” Milbank Quarterly, 80(3):569–593 Nov. 10, 2011 at http://www.iiiglobal.org/ Public Hospitals in Peril: Factors Associated with Financial Distress 29

component/jdownloads/finish/648/5645. Closure: Managerial Incentives and Political html. Costs,” Journal of Public Budgetting, Account- 50. Altman, EI,”Financial Ratios, Discriminant ing and Financial Management, 25(1):31–68 Analysis and the Predition of Corporate Bank- (2011). rupcty,” The Journal of Finance, 23(4):589– 58. American Hospital Association, “Facts on US 609 (1968). Hospitals,” accessed Jul. 27, 2013 at http:// 51. Langabeer, supra, n.9. www.aha.org/research/rc/stat-studies/fast- 52. Lee, S, Alexander, JA, “Managing Hospi- facts.shtml. tals in Turbulent Times: Do Organizational 59. Id. Changes Improve Hospital Survival?,” Health 60. Alexander, JA, Morrisey, MA, “A Resource- Services Research, 34 (4):923 (1999); Sloan, dependence Model of Hospital Contract FA, Ostermann, J, Conover, CJ, “Antecedents Management,” Health Services Research, of Hospital Ownership Conversions, Mergers, 24(2):259–284 (1989). and Closures,” Inquiry-the Journal of Health 61. Wooldridge, JM, Introductory Econometrics: Care Organization Provision and Financing, A Modern Approach, New Delhi: Akash Press 40 (1):39–56 (2003); Weech-Maldonado, R, (2006). Qaseem, A, Mkanta, W, “Operating Environ- 62. Li, B, Lingsma, H, Steyerberg, E, Lesaffre, E, ment and USA Nursing Homes’ Participation “Logistic Random Effects Regression Mod- in the Subacute Care Market: A Longitudi- els: A Comparison of Statistical Packages nal Analysis,” Health Services Management for Binary and Ordinal Outcomes,” BMC Research, 22(1):1–7 (2009); Kim, supra, n.9. Medical Research Methodology, 11: 1–11 53. Kim, supra, n.9. (2011), accessed Jul. 25, 2013 at http://www. 54. Weech-Maldonado, Qaseem, and Mkanta, biomedcentral.com/content/pdf/1471-2288- supra, n.52. 11-77.pdf. 55. Bazzoli, JG, Chan, B, Shortell, S, D’Aunno, 63. Woolridge, supra, n.61. T, “The Financial Performance of Hospitals 64. Field, A, Discovering Statistics Using SPSS, Belonging to Health Networks and Systems,” Thousand Oaks, CA: SAGE Publication, Inc. Inquiry: A Journal of Mecidal Care Organiza- (2009). tion, Provision and Financing, 37(3):234–252 65. Kim, supra, n.9. (2000). 66. Kim, supra, n.9; Kotha and Nair, supra, n.27; 56. Detsky, A, O’Rourke, K, Naylor, D, Stacey, Goll and Rasheed, supra, n.29; NORC, supra, S, Kitchens, J, “Containing Ontario’s hos- n.36. pital costs under universal insurance in the 67. Hall and Weiss, supra, n.37. 1980s: What was the record?,” Canadian 68. Sharfman, et al., supra, n.39; Wan, WP, Yiu, Medical Association Journal, 142(6):565– DW, “From Crisis to Opportunity: Environ- 572 (1990); Vujicic, M, Addai, E, Bosom- mental Jolt, Corporate Acquisitions, and Firm prah, S, “Measuring Health Workforce Performance,” Strategic Management Journal, Productivity: Application of a Simple Meth- 30(7):791–801 (2009). odology in Ghana. Health Nutrition and Pop- 69. Porter, supra, n.38. ulation Discussion Paper” accessed Jan. 24, 70. McCue, MJ, Furst, R, “Financial Character- 2012, at http://siteresources.worldbank.org/ isitcs of Hospitals Purchased by Investor- HEALTHNUTRITIONANDPOPULATION/ owned Chains,” Health Services Research, Resources/281627-1095698140167/Measur- 21(4):515–527 (1986). ingHlthWorkforce.pdf. 71. Bazzoli, et al., supra, n.55; Broyles, RW, 57. McCue, supra, n.9; Kim, supra, n.9; Trussel, J, Brandt Jr., EN, Biard-Holmes, D, “Networks Patrick, P, “An Empirical Analysis of Financial and the Fiscal Performance of Rural Hos- Distress in Pennsylvania Hospitals,” Journal pitals in Oklahoma: Are they Associated?,” of Health Care Finance, 36(1):31–60 (2009); The Journal of Rural Health, 14(4):327–337 Liu, L, Jervis, K, Younis, Mz, Fiorgione, DA, (1998); Nauenberg, E, Brewer, CS, Basu, K, “Hospital Financial Distress, Recovery and Bliss, MK, Osborne, JW, “Network Structure 30 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

and Hospital Financial Performance in 76. Calandro, JJ, “Considering the Utility of Alt- New York State: 1991–1995,” Medical Care man’s Z-score as a Strategic Assessment and Research and Review, 56(4):415–439 (1999). Performance Management Tool,” Strategy & 72. Pfeffer, supra, n.15. Leadership, 35(5):37–43 (2007). 73. Bazzoli, et al., supra, n.55; Nauenberg, et al., 77. Donnelly, JM, “Hospitals Turn to Outpatient supra, n.71; Bazzoli, JG, Shortell, S, Cili- Services to Offset Inpatient Care Cuts”. Bos- berto, F, Kralovec, P, Dubbs, N, “Tracking ton Business Journal, accessed Feb. 25, 2012 the Changing Provider Landscape: Implica- at http://www.bizjournals.com/boston/print- tions for Health Policy and Practice,” Health edition/2011/08/26/hospitals-turn-to-out Affairs, 20(6): 188–196 (2001); Federico, L, patient-services.html. “Strategies for Multihospital Networks: A 78. Andrulis, D, Siddiqui, N, “Health Reform Framework,” Health Services Management Holds Both Risks and Rewards For Safety-Net Research, 18(2):86–99 (2005). Providers and Racially and Ethnically Diverse 74. Bazzoli, et al., supra, n.55; Broyles, et al., Patients,” Health Affairs, 30(10):1830–1836 supra, n.71; Federico, supra, n.73. (2011). 75. Nauenberg, et al., supra, n.71. 79. Id. Using a Social Entrepreneurial Approach to Enhance the Financial and Social Value of Health Care Organizations

Sandra S. Liu, Jui-fen Rachel Lu, and Kristina L. Guo

In this study, a conceptual framework was developed to show that social entrepreneurial practices can be effectively translated to meet the social needs in health care. We used a theory-in-use case study approach that encompasses postulation of a working taxonomy from literature scanning and a deliberation of the taxonomy through triangulation of multilevel data of a case study conducted in a Taiwan-based hospital system. Specifi cally, we demonstrated that a nonprofi t organization can adopt business principles that emphasize both fi nancial and social value. We tested our model and found comprehensive accountability across departments throughout the case hospital system, and this led to sustainable and continual growth of the organization. Through social entrepreneurial practices, we established that both fi nancial value creation and fulfi lling the social mission for the case hospital system can be achieved. Key words: social entrepreneurship; fi nancial value; fi nancial sustainability; social capital; enterprise value; social value

Introduction the health care structure, cultural infl uences, fi nancing, and delivery of services.4 Social The health care sector is in transition glob- entrepreneurship, which is part of a broader ally. Reforms in the United States and emerging economies in Asia and other nations Sandra S. Liu, PhD, is Professor in the Department of around the world are calling for viable mod- Consumer Science, College of Health and Human Sci- els to provide fi nancially sustainable health ences, Purdue University in West Lafayette, Indiana. services in the current turbulent and uncer- Jui-fen Rachel Lu, ScD, is Professor in the Depart- tain environment. There is growing aware- ment of Health Care Management, at Chang Gung ness and concern regarding social demands University (CGU) in Taiwan, where she teaches com- and health care costs. For instance,1 there are parative health systems, health economics, and health ongoing discussions about hospitals’ attempts care fi nancing. to fulfi ll their social mission while maintain- Kristina L. Guo, PhD, MPH, is Chair and Professor of Public Administration and Professor and Director ing fi nancial and operational sustainability. of the Health Care Administration Program at the Many hospitals (mostly nonprofi t) receive University of Hawai’i-West O’ahu (UHWO). credit for services funded by the government. Acknowledgments: The authors express deepest grati- This political factor makes public subsidies tude to Dr. Chao-Shiung Chang, Dr. Delon Wu, Direc- sensitive to changes in federal revenues,1 and tor Wen-Hua Kong, CEO Ming-Long Huang, Executive hence these nonprofi t hospitals are challenged Administrator Shu-Chen Lin, Senior Manager Hsing- to fi nd alternative revenue sources to avoid Yeh Tai, and staff at CGMH for their invaluable inputs. 2 In addition, the grant support of Taiwan National Sci- welfare dependency. Many hospital adminis- ence Council (NSC96-2416-H-182-001-MY3) was trators have advocated the adoption of entre- greatly appreciated. The authors are solely account- preneurship in terms of patient-centered care, able for any error and the contents of the article. market responsiveness, and sustainability in J Health Care Finance 2014; 40(3):31–46 the health sector,3 but barriers remain due to Copyright © 2014 CCH Incorporated 31 32 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

family of entrepreneurship, permits us to ened to include all business practices of an conceptualize systems and processes that are enterprise that assumes fi nancial, personal, designed to achieve social change5 and to and organizational reputation risks when generate surplus to support activities that can- stimulating social changes and progress.10 not generate revenue.6 That is, social entre- Even though businesses focus on eco- preneurship is critical for generating social nomic returns, social enterprises emphasize impact and assuring fi nancial sustainabil- the organization’s mission for generating ity. Entrepreneurial orientation with a social social value and advancing social change11 focus is a tool that allows the leadership team so as to obtain their legitimacy and needed to proactively strategize to anticipate environ- resources.12 As such, a conventional entre- mental changes and to lead the social move- preneurial orientation that encompasses ments that have policy implications. only the entrepreneurial attributes of risk- In this research, the authors aim to develop taking, opportunity-recognition, and inno- a model by which the social entrepreneurial vativeness falls short of addressing issues approach can be effectively translated to meet that contribute to sustaining socially rel- the social needs in health care. We use a theory- evant initiatives. in-use case study approach that encompasses postulation of a working taxonomy from lit- Market Dynamism erature scanning and a deliberation of the As shown in our conceptual model taxonomy through triangulation of multilevel (Figure 1), a taxonomy for building social data of a case study conducted in a Taiwan- value is developed illustrating the process of based hospital system that has an established social entrepreneurial practice. Specifi cally, reputation for effectively using business oper- market dynamism is the driving force of ating principles in managing hospitals.7 In the environment coupled with the organi- this article, we fi rst develop the conceptual zation’s market orientation. Together, these taxonomy that is theoretically underpinned, can effectively impact entrepreneurship. with the plan that it will serve as a framework For instance, when businesses compete in for case delineation and construct building for an especially volatile market, market intel- future research in health care. ligence is critical for those organizations to retain their fi nancial stability and sustaina- Conceptual Framework Development bility. Kohli and Jaworski (1990)13 identifi ed market orientation as an on-going process Entrepreneurial Process with a Social Focus of market information generation and dis- Entrepreneurship involves a process that semination. Market-oriented organizations proactively recognizes opportunities, takes tend to excel in their capabilities for seeking risks, and provides for innovative new and using market information to best gener- approaches for providing customer-centric ate and deliver a superior customer value,14 solutions.8 “Social entrepreneurship” was and their market orientation has been found a term used to refer more narrowly to an to drive entrepreneurship.15 In addition, enterprise’s application of market-oriented environment changes also impact market principles in the nonprofi t or public sector,9 dynamism. For example, global concerns but this defi nition has gradually been broad- of health disparity, social justice, economic Financial and Social Value of Health Care Organizations 33

Figure 1. A Taxonomy for Building Social Value through Social Entrepreneurial Practice

Enterprise Value Market Dynamism

Environmental Social Changes Social Value + Entrepreneurial Financial Value Market Practice Orientation Policy

Shared Social Mission of the Social Capital Top Management Team

recession, and other social-economic envi- entrepreneurs are motivated by recognized ronment changes have led to an increased market failures,19 but they view social value awareness of the sustainable utilization of creation as an organization’s explicit and global resources. More government agen- central mission.20 Hence, this determination cies are using contracts and vouchers to infl uences their specifi c marketing strate- ensure measurable performance in the gies, actions that place a relatively higher delivery of social services by both nonprofi t priority on social value creation while, in and for-profi t organizations. Because these the meantime, borrowing business princi- organizations are competing in the same ples from their commercial counterparts.21 market segment,16 their fi erce competition During the social value creation process, has compelled them to become more inno- risks are inevitable because any failure of vative and outcome-driven.17 Furthermore, the mission may impair the reputation and environmental changes such as recent sustainability of the social enterprise or debates on health care reform in the United even the entrepreneur’s personal reputa- States and other emerging economies have tion.22 The buy-in from the top management created a stronger sense of uncertainty team (TMT)23 and the operation team are regarding the future of health care payer pivotal, yet challenging for the organiza- models and health care delivery systems tion to evolve into a social-entrepreneurial that address the determinants of health.18 As culture. Organizations without the shared a result, market dynamism is the fi rst ele- social vision tend to have multiple and pos- ment affecting social entrepreneurship. sibly divergent directions or views;24 thus, the shared social vision from top manage- Shared Social Visions ment teams is essential to successful entre- of Top Management Teams preneurial practice. In addition to market dynamism, the shared social vision of top management Social and Financial Value Creation teams determines the organization’s entre- For-profi t organizations measure their per- preneurial practice. In particular, the shared formance more obviously by means of their vision is socially driven. Moreover, social profi t generation. For social entrepreneurs, 34 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

understanding and maximizing social value Methods creation is the primary performance focus The global health care industry is becom- and is based on the organization’s social mis- ing increasingly more complex and challeng- sion. Enterprise value is an important part of ing as world nations and economies grow the model because social enterprises are con- ever more closely entangled. The task at hand sidered change agents in society. They are is more than merely leadership development, able to create a sustainable social movement. cost containment, or technology adoption; Social enterprises can adopt an “earned rather, it demands a fundamental develop- income” strategy during the social value ment of a viable business model that will creating process25 to ensure their fi nancial allow hospitals to serve their target segments sustainability and self-suffi ciency.26 Social sustainably. We adopted a theory-in-use case value and economic value creation should study methodology that entails deducing not be dichotomous. As shown in Figure 1, a working taxonomy from literature based social and economic (or fi nancial) values are on scanning and using the taxonomy of an connected. For example, the Roberts Enter- instrumental case study.33 In this epistemo- prise Development Fund (REDF) coined the logical process, the case plays a supportive term “social return on investment” (SROI) as role in facilitating the understanding of the an outcome of the entrepreneurial process.27 concepts within the health care context. The SROI quantifi es the fi nancial outcome and transcripts of the interviews and the content the social value that is generated, and it of the publications and documents related measures the sustainability of the enterprise. to strategies, operation plan, and meeting Kent and Anderson (2002)28 advocated that minutes are reviewed, organized, and ana- social entrepreneurs should be the “bridge lyzed to determine core constructs, the con- builder(s)” working to create communities texts in which these constructs are translated through which social value is to be gener- into activities and actions, and the resulting ated. Social capital, which is broadly defi ned consequences. Based on data analyses, core as an intangible asset embedded in relation- outcomes are measured in terms of the value ships,29 can be leveraged to facilitate action generated in fi nancial and social dimensions. and contribute to the organizations’ levels Both forms of well-defi ned knowledge, of performance by increasing communi- namely constructivism and pragmatism, cation effi ciency,30 decreasing the cost of direct the strategies of inquiry and methods transactions,31 and leading to a synergistic of data collection and analysis in this pro- enhancement of performance.32 Social capi- cess. Patton (2002)34 stressed the importance tal should therefore enhance social and eco- of the study purpose in designing qualitative nomic/fi nancial value creation and result in studies along the theory-action continuum. policy change. As described above, Figure 1 During the preliminary stage of developing depicts the proposed drivers for an organi- a theoretical taxonomy for adopting a social zation’s social entrepreneurial practices and entrepreneurial approach for sustainable the resulting outputs. The case study system- social value creation, the theory-in-use case atically delineates individual constructs and study methodology is both meaningful and discerns their relations using our conceptual appropriate for attaining insights and ideas framework. that aim to increase familiarity with the Financial and Social Value of Health Care Organizations 35

context and issues and to help generate testa- (3) Around the time of the promulgation ble hypotheses.35 The case hospital system in of National Health Insurance, how Taiwan is a private health care system that is was the case system in serving the renowned for its service innovation and effi - target populations? ciency.36 With the proposed taxonomy, we (4) Based on the informant’s experience examined in-depth the relationships between and understanding, what are the sys- the changes in the external environment, the tem’s contributions to and/or infl u- entrepreneur’s market orientation, and the ences on the society and the health organizational development and outputs. Tai- care sector? wan has modeled its health care practices on (5) How has the case hospital system those of the United States after World War II. been adapting to the societal and/or The majority of the key players in its health environmental changes? care sector were US trained. Due to its sin- (6) What are the performance outcomes gle-payer National Health Insurance system, of the case hospital system? which provides universal health coverage to its 23 million people, Taiwan has served A total of 13 informants were interviewed, as a pilot testing laboratory for a number of including the former superintendent (equiva- care models over the years. Hence, the les- lent to the position of hospital director in the sons learned from Taiwan’s health care sys- US), chair of the steering committee, the tem have been widely discussed in various current director of the administrative offi ce, forums in the United States and the Asia- senior managers/administrators from admin- Pacifi c region, in particular, within China.37 istration, clinical departments, the emergency We collected data from focus group studies department, laboratory, and pharmacy, and and one-on-one interviews with the instru- frontline leaders from clinical departments. ment encompassing the following items: A particularly close scrutiny of personal writ- ings from the founder and former director of (1) Please elaborate on the environment the case hospital system’s Offi ce of Adminis- and the impetus for the initial estab- tration provided for a critical understanding lishment of the case hospital system. of the founding mission, vision, and operat- What are the founder’s vision, expec- ing principles. The informants’ experiences tation, strategic orientation, and his with the hospital ranged from 5 to 32 years. perspectives regarding social mission/ They personally witnessed the hospital’s responsibilities? What are the strate- establishment and evolution through their gic directions/planning that you are participation in the strategic development aware of? How much has been accom- and implementation process. All interviews plished so far, from your perspective? were audio-taped and transcribed to provide (2) What are the unique attributes of the accurate data analyses. case organization and its culture? The data analysis process was dynamically What is the evolution of the case intertwined with theory and data, and this pro- hospital system’s social mission, its cess involved (1) discerning recurring catego- organizational culture, and the ser- ries and emerging themes; and (2) employing vices to the community? the constant comparative method to ensure 36 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

the internal and external signifi cance of each that accounted for 17 percent of Taiwan’s theme. The context for data analysis was GDP by 2008. During the formative phase of defi ned within the historical context of the the case hospital, it specifi cally targeted blue- case hospital system and the evolution of its collar workers, operated with a lean process, operations. The corresponding author has and collected timely customer feedback. The maintained close collaborations with the case founder also invested heavily in recruiting the hospital system, through which she has been best talents globally for the leadership team able to continually update the required fi nd- of the case hospital.39 Furthermore, he person- ings and correspond with informants with the ally devoted his time and effort into learning purpose to best address any ongoing issues the business from these experts and directly discovered in the process. from patients and their families through the “Superintendent’s mailbox.” At the manage- Findings ment level, a series of weekly meetings provided a platform for the departments’ business During the 1970s, the health care system in managers to disseminate information and to Taiwan was composed of small private clinics engage in formulating institutional responses and large, primarily public general hospitals. to the information gathered. The case hospital The capacity of these large hospitals in terms system has prided itself on its research, con- of their medical service abilities and quality tinuing education and training of its medical was determined to be insuffi cient for meeting staff, and ownership of the state-of-the-art the market demand. The ratio of the number technology, facility, and equipment. of physicians per 1,000 people was merely Being the fi rst privately owned large-scale 0.4, compared to 1.6 doctors per 1,000 peo- general hospital with a broadened access to ple in 2010,38 and the emergency departments health care for all, the case hospital system (ED) were primarily staffed with resident grew rapidly and, as a result, invited more physicians. This insuffi cient and inappropri- competition into the market place. Accord- ate provision of care at the ED victimized the ing to an interviewee, father of a successful entrepreneur. In memory of his father, the entrepreneur founded the fi rst The government saw us as a strong privately owned general hospital on the island rival and, to stay competitive, they in 1976 with the clearly articulated mission of were willing to approve the budget “promoting social welfare.” The founder man- request(s) from other public hospitals aged to maintain that aspiration for the follow- for their upgrades and expansion in ing three decades when serving as chair of the technology and infrastructure. board of directors for the case hospital system. In addition, its growing economic value Entrepreneur’s Market-Orientation enticed other business conglomerates to In response to the keen competition from enter into Taiwan’s health care sector to the its public hospital counterparts and the lack extent that the government had to establish of support from government, the founder legislation to better monitor the fi nancial adopted the operating principles of his US$69 performance of these foundation-owned billion conglomerate, with an annual revenue private hospitals. Financial and Social Value of Health Care Organizations 37

Shared Mission of the Top Management Team (3) the case hospital system should be The founder formed a leadership team devoted to advanced medical service that was composed of the superintendent, and science through research and the vice superintendent, the chairman of the education. Medical Executive Committee, and directors of the Administrative Offi ces from both Following the formative stage of the case the case hospital system and his business hospital system, the founder had dedicated enterprise. They conducted weekly dinner an ongoing 1 percent of the revenue from meetings on Friday nights to discuss and the daily operations to the system’s charity examine extensively the operational plans fund until he passed away in 2010. This fund and the execution of such plans. These was found to be particularly useful during frequent face-to-face meetings of the man- the pre-National Health Insurance (NHI) agement team were instrumental in both era because the government had imposed cultivating the organizational core value of stringent criteria for qualifying low-income “treating patients as our own family” and families for public medical assistance. establishing a lean and tightly controlled Some informants believed that this practice operational infrastructure at the formative later instigated the government’s policy of phase of the case hospital system. requiring all the foundation-owned hospitals The director from the business enterprise to allocate 10 percent of their surplus to a played a crucial role in building the systems social service fund. and processes by translating business princi- Entrepreneurial Practices Focusing ples into the health care context. The “cost- on Financial and Social Values down” concept became a major concern across the board from areas as divergent as During its formative phase, the case cafeteria logistics to renal dialysis.40 The Fri- hospital system provided transportation to day night dinner meetings later evolved into a patients in rural areas to provide these indi- series of regular daily luncheons, and weekly, viduals with access to proper health care. biweekly, or monthly meetings attended by This system also was the fi rst in Taiwan to various levels of the management teams. In deploy only attending physicians instead of 1985, a formal Medical Steering Committee residents or interns to the emergency depart- was established to govern clinical care, teach- ment (ED) to ensure the proper treatment of ing, and research programs.41 The vision of patients in the ED where the founder’s father the TMT was clearly articulated, according had lost his life as a result of delayed treat- to the informant, as one in which: ment. Moreover, the case hospital system developed a streamlined one-stop payment (1) The case hospital system should system to allow for outpatient and deposit- provide quality care at an affordable free inpatient registration processes. price to the public; Guided by their shared social mission, (2) All leaders should own the not-for- the case hospital system’s top manage- profi t mindset and aim to generate ment team strictly prohibited their physi- social benefi ts instead of personal cians from “moon-lighting” or accepting economic gains; and “gift-money.” The gift-money (more com- 38 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

monly referred to as the “red envelope”) leaders. These executives, functioning like was and is still a prevalent practice in some the think tank of the hospital superintendent, hospitals in Taiwan whereby doctors are routinely rotated across all departments in incentivized to expedite the scheduling order to acquire a fi rst-hand understanding or preferentially treat the patients. Often of the frontline operation. At the operational times, it is given to surgeons before the front, “accountability managers” in turn led operation in accordance with a “market and managed a total of 3,783 (as of the Year price.” Sometimes such red-envelope fees 2011) “accountability centers” across all can be correlated with a signifi cant source clinical departments and business offi ces, of income for more well-known doctors and monitored and audited their respective within the nation. Therefore, to ensure services against 140 existing service and effective implementation of these policies quality indices, for example, wait time and with doctor/patient satisfaction, the case number of visits.43 These key performance hospital system initiated a performance- indices were determined by the Adminis- based physician fee (PF) payment system. tration Center based on the management- The administration determined the PF rate by-objectives (MBO) process. During the based on the physician’s: (1) service vol- monthly operation management meetings, ume, (2) seniority, and (3) overall contribu- individual clinical specialties were reviewed tions to the department in administration, on a periodic basis. Any budgetary gap clinical teaching, and profession. A ceiling called for a closer scrutiny by the responsi- was imposed, and surplus was pooled into ble accountability manager to determine a a hospital fund for sponsoring activities timely and appropriate course of action. such as travel expenses for attending work- This unifi ed information system has been related conferences under the auspices of instrumental in the case hospital system’s the department of medical affairs.42 efforts in bundling patient care for those with The case hospital system emphasized chronic conditions. First, the hospital phar- cost-containment strategies to ensure its macy launched an expedited dispensing ser- healthy fi nancial outlook, and the system vice both to automatically package together established a knowledge management infra- the routine prescription refi lls for frequent structure to better standardize its internal users and to manually dispense new prescrip- operations. The infrastructure integrated the tions. The case managers then coordinated the hospital system’s centralized procurement inter- specialty patient care, which engaged and inventory control system with that of physicians across departments to share and the founder’s conglomerate for procuring evaluate patient information and treatment hospital supplies directly from the com- regimens and also had a greater focus on pany. To manage these complex business assisting in the development and monitoring systems and processes, the case hospital of a comprehensive treatment plan for indi- system staffed its Administration Center at vidual patients. When Taiwan’s NHI launched the senior management level with “profes- the integrated-care initiative for patients with sional executives” who are responsible for multiple chronic conditions in December formulating strategies, supervising hospital 2010 by offering a NT$1,000 (US$34) per operations, and developing departmental case incentive to the participating hospitals, Financial and Social Value of Health Care Organizations 39

the case hospital system was already pre- Social capital. During the 1970s and pared and readily joined the national initia- 1980s, when there were just a few teaching tive, which resulted in an additional revenue hospitals, the case hospital system opened its source for the organization. door and offered internships to medical students from other medical schools.45 Today, Multiple Dimensions of Outputs approximately one-fourth to one-third of all of the practicing physicians in Taiwan are Enterprise value. The case hospital sys- derived from these programs. Many of the tem’s entrepreneurial practice has resulted hospitals that were managed or led by these in the building of a signifi cant enterprise alumni have adopted a similar management value (Tables 1 and 2), which led to its siz- system and the business philosophy. The able social and fi nancial values (Table 2), established strategic alliances and the accu- thus becoming the largest health care sys- mulated social capital evolved into a social tem in Taiwan. The hospital system has force that has continually driven the trans- expanded to maintain more than 10,000 formation of Taiwan’s health care policies beds in eight hospital complexes spreading and systems. As one informant commented: across six different geographical regions; by 2010, it had a total of 342,336 admis- The health care sector in Taiwan went sions and 7,210,016 outpatient visits, from 30 percent privately owned hos- which amounted to a total of 8 percent pitals to 70 percent from the 80s. The of the annual outpatient visits in Taiwan. keener competition continued to drive Its annual revenue of US$1.561 billion in the advancement in hardware (such as 2010 amounted to approximately 10 per- facilities and medical equipment) and cent of the NHI total expenditures rendered software (the managerial and clinical to serving its insured. The business model skills training of the medical profes- borrowed from the founder’s conglomerate sionals), which allowed an earlier insti- has allowed the case hospital system to pro- tution of National Health Insurance vide patient-centered care with a healthy from 2000 to 1995. The availability of surplus margin even with the NHI’s strin- insurance helped particularly those who gent and tightly controlled global budget are under the poverty line and those system since 2002.44 who contract catastrophic ailments.

Table 1. Annual Service Volume and Number of Staff in Case Hospital System, 2008–2010

Annual Service Volume Number of Staff Staff (excluding physicians and Year Outpatient Inpatient Emergency nurses) Physicians Nurses 2008 7,238,168 327,612 493,226 7,908 2,806 7,185 2009 7,332,406 326,342 513,723 8,438 2,906 7,408 2010 7,210,016 342,336 507,962 8,715 2,944 7,579 40 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 2. Financial Performance of the Case Hospital System, 2008–2010

Total revenue/Staff in US$ Year Total revenue in US$a (including Medical staff) Debt/assets 2008 1.486 billions 438,000 3.86% 2009 1.523 billions 423,000 2.77% 2010 1.561 billions 450,000 2.13% aUS$1 is roughly equivalent to NT$30 from 2008–2010.

The case hospital system has been recog- patients for their medical assistance and nized as an important agent of change for required living expenses and assisted living the health care industry in Taiwan; many expenses when needed. Furthermore, the case of those changes have essentially resulted hospital system has sponsored specifi cally in creating social value for patients at the cochlear implants, stem cell transplantations, national level and led to fi nancial success and other medical devices that improve the for the organization. Recognizing the grow- basic quality of life of its socially disadvan- ing global economy, the case hospital sys- taged patients. The case hospital system also tem has begun to build its social capital collaborated with NHI in offering free clinics by training about 160 to 190 interns from to the indigent populations. Table 3 lists the medical schools in the emerging economies investments of the case hospital system when during the years 2008 to 2010. generating these social and fi nancial values Social and economic/fi nancial value with for the underserved populations during the policy implications. With its founding princi- years of 2008 to 2010. ple of “serving the interests of patients fi rst,” Both the fi nancial and social successes all caregivers were required to provide patient- of the case hospital system have drawn centered services with respect and attentive- more attention from the government and the ness. Its Social Service Fund has, in recent existing establishments in the public sector. years, disbursed approximately US$113 mil- A multitude of foundation-owned hospitals lion of medicine to medically underserved fl uxed into the market following the case

Table 3. Social Value Generated for the Underserved Populations, 2008–2010

Medical assistance to the socially Year disadvantaged Free Clinic Persons US$ (M)a Persons Served (M) US$ (K) served 2008 5.900 1.030 3.333 817 2009 10.364 1.590 8.000 2607 2010 18.209 4.020 31.333 10065 a US$1 is roughly equivalent to NT$30 from 2008–2010. Financial and Social Value of Health Care Organizations 41

system’s example in late 1970s to the 1980s. toward becoming more market driven, The government responded to this market whereas those who start with a more mar- force by establishing a new legal entity of ket-driven idea may develop more socially hospitals in 1987 and allowing enterprises conscious practices as a result of the market to establish nonprofi t foundation-owned demands. The fi ndings of this case study hospitals with endowment funds; these hos- show that the volatile environment must be pitals were later regulated under the jurisdic- coupled with market- oriented organizations tion of the Department of Health instead of to lead to entrepreneurial practices in the the Ministry of Internal Affairs. This policy health care context. The positive fi nancial change quickly led to an even more signifi - output of such a business model then attracts cant increase in this segment, and the result- the infl ux of competitors/providers. The ing keener competition compelled providers continuously changing environment, poli- to use more advanced medical technology to cies (such as the tax status and regulations, attract patients. One informant stated: “We insurance and payer systems) and the norm/ believed that health insurance for all the philosophy of operations resulted in further workers on the Island is essential and led the expansions of privately owned general hos- negotiations with the public insurance organ- pitals. Figure 3 illustrates both the steady ization. Once the agreement was reached, growth of the health care sector and the pro- other foundation- and faith-based hospitals vision of health care to the local population. followed suit.” During the period from 1960 to 2010, the Figure 2 delineates the constructs that number of hospital beds in the private sec- were discerned and evidenced in the case tor grew from 30 percent to 66.04 percent, study. The socially focused entrepreneurial a sector in which foundation-owned hospital practices are instigated by the entrepreneur’s systems accounted for 69.14 percent of the market-orientated response to the environ- whole. The provision of health care in terms mental challenges (such as providing care to of beds available improved from 0.7 beds underserved populations and implementing per 1,000 people to 5.8 beds. This phenom- policies to ensure quality of care), and this enon illustrates the importance of the social is facilitated by the management’s shared entrepreneurs and their leadership teams as social mission of generating social benefi ts the driving force for fostering an “ethics- and providing affordable quality of care. The grounded” culture,47 one in which the social multidimensional outputs of social capital entrepreneurial processes resulted in outputs and enterprise value further drive the crea- that encompass both fi nancial and social tion of social and fi nancial value. outcomes as well as having a direct impact on policies.48 The case hospital system con- Discussion stantly operates in a fl uctuating and uncer- Health care as a social service has tain context, and this presented a distinct encountered major challenges globally in opportunity for the organization to instigate an environment with limited resources and a social movement that exerted strong infl u- widening social disparity. According to ences on its public and political environ- Tiku (2008),46 a market’s operational needs ments. Through its entrepreneurial practices may sway the socially driven entrepreneurs of using social and enterprise capital, the 42 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 2. Key Components of the Social Entrepreneurial System of the Case Hospital System

INPUTS PROCESS Market-Oriented Response to the Environment Social Entrepreneurial Practice • Focusing on providing health care to underserved • Patient-centered care payment system populations ο Deposit-free inpatient registration • Initiating innovative forms of private general ο Streamlined one-stop payment system for outpatient hospitals ο Physician Fee system to incentivize physicians for • Implementing policies to ensure quality of health providing fair and compassionate care to patients by services avoiding “moonlighting” and receiving “gift money” • Establishing systems for hearing the voices of from patients the medical experts and the patients • Management strategy ο “Administration center” with job rotations at the Shared Social Mission of the TMT management level • Provision of quality health care at affordable ο “Accountability centers” to manage by objectives, prices to the general public control cost, and monitor performance at the • Emphasis of generating social benefits instead operational level of personal economic gains • Market-orientated knowledge management system • Investment in education and advanced medical ο Standardization of processes with 140 performance technologies indices ο Centralization of procurement and patient care processes

OUTPUTS

Social Capital Creation Enterprise Value Creation • Open residency training programs for • Cost containment and control over interns from other medical schools programs and activities • Transplantation of management • Revenue streams from multiple system and support of clinical customer-oriented service services to member hospitals through innovations strategic alliance

Social and Economic/FinancialValue Creation • Respectful and compassionate care for all • Improved equitable access to quality care ο Free transportation for the rural population ο Social Service Fund for the medically underserved • Financial Value/Viability • Enhanced quality of medical care ο More advanced medical technologies Financial and Social Value of Health Care Organizations 43

Figure 3. Evolution of the Hospital Market in Taiwan, 1960–2010

1960 19701986 19901995 2010

Bed capacity 0.7 2.4 4.0 4.44.7 5.8 Per 1,000 in populations Ownership: 29.7% 39.2% 57.84% 57.28% 60.64% 66.04% Private sector (32.14%) (42.57%) (45.60%) (69.14%) (% of which, foundation- owned) Legislation enacted to regulate NHI implemented CGMH founded in 1995 in 1976 foundation-owned hospitals in 1986

Influx of hospital beds Government Market Impact of • Expand competition NHI payment Scarce medical investment in intensified system on facilities, mainly public with hospital in public sector hospitals increasing market • Open market private beds structure for private investment result led to social value that benefi ts the and processes in the health care context. For general public and fi nancial value that is instance, bundled care models and accounta- advantageous for the organization itself. Fur- bility care organizations in the United States thermore, by building social capital, entre- and privatization of hospitals in China49 are preneurs are enabled to play a strong role examples of entrepreneurial practices. Other in making an impact on the policy/public scholars also recommend the formulation environment that eventually allows a much of networks to infl uence complementary more powerful social value creation through or even competing organizations to mobi- fundamental social change. lize their resources toward activities across The health care industry is in transition organizational and sector boundaries to lead throughout the world. Health care reforms in to maximal social impact.50 the United States and other emerging econo- To achieve these goals, hospital leaders mies, such as China, present an environment should focus their efforts internally to attain that is constantly changing and uncertain. support from their boards while continuing The entrepreneurial orientation of the lead- to engage in dialogues with external stake- ers of health care organizations and policy holders as well as in advancing the social makers is shown to be an important drive for mission of their hospital. The innovative translating business principles into the devel- management system and business model of opment of more sustainable business models the case hospital system have shown to exert 44 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

profound impacts on their counterparts in of them focus only on their social causes the public sector and within the health care regardless of cost.51 However, in this study, market as a form of positive externality. we demonstrated that a nonprofi t organi- Consequently, the government has actively zation can adopt business principles that responded to their evidence-based paradigm emphasize both fi nancial and social value. shifts in practice through on-going policies Using a case hospital system in Taiwan, changes. Nonprofi t health care organiza- we tested our model and found compre- tions usually have a clear social mission, hensive accountability across departments but they are often unable to fully appreci- throughout the hospital system. This led ate the extent to which they may generate to sustainable and continual growth of the social values for the public at large. This organization. Consequently through social case study provides an element of the ini- entrepreneurial practices, we have shown tial evidence of real impact that such steps that both fi nancial value creation and ful- taken in the nonprofi t health care market fi lling the social mission for the case hos- may hold. The fi nancial viability of the case pital system can be achieved. In future hospital system has motivated other health studies, we intend to test our model derived care institutions in Taiwan, China, and Asia from the taxonomy presented in Figure 1 at large to emulate its business model. with the purposes of demonstrating that applying entrepreneurial practices to the Conclusion health care industry can produce positive fi nancial and social impacts as well as Socially driven entrepreneurial organi- guide the direction of organizational and zations are built to achieve change but most government policy changes.

REFERENCES

1. Tuckman, H, Chang, C, “A methodology for 4. Guo, KL, Buss, TF, “Entrepreneurship in health measuring the fi nancial vulnerability of chari- and human services organizations: A sympo- table nonprofi t organizations,” Nonprofi t sium,” Journal of Health and Human Sciences and Voluntary Sector Quarterly, 20:445–460 Administration, 28:468–471(2006). (1991). 5. Austin, JE, Stevenson, H, Wei-Skillern, J, 2. Fischer, RL, Wilsker, A, Young, DR, “Explor- “Social and commercial entrepreneurship: ing the revenue mix of nonprofi t organiza- Same, different, or both?” Entrepreneur- tions: Does it relate to publicness,” Nonprofi t ship: Theory & Practice, 30:1–22 (2006); and Voluntary Sector Quarterly, 40:662– Dees, GJ, Anderson, BB, “For-profi t social 681(2011); Sullivan Mort, G, Weerawardena, ventures,” International Journal of Entrepre- J, Carnegie, K, “Social entrepreneurship: neurship Education, 2:1–26 (2001); Weera- Towards conceptualization,” International wardena, J, Sullivan Mort, G, “Investigating Journal of Nonprofi t and Voluntary Sector social entrepreneurship: A multidimensional Marketing, 8:76–88 (2003). model,” Journal of World Business, 41:21– 3. Griffi th, JR, “Finding the frontier of hospital 35 (2006). management,” Journal of Health Care Man- 6. Campbell, S, “Social entrepreneurship: How agement, 54(1):57–73 (2009). to develop new social-purpose business Financial and Social Value of Health Care Organizations 45

venture,” Health Care Strategic Management, of explanation, prediction, and delight,” Jour- 16:17–18 (1998). nal of World Business, 41:36–44 (2006). 7. Lu, JR, Chiang, T, “Evolution of Taiwan’s health 21. Austin, et al., supra, n.5. care system. Health Economics,” Policy and 22. Grenier, supra, n.10. Law, 6(1):85–107 (2011). 23. Hambrick, DC, Mason, PA, “Upper ech- 8. Covin, JG, Slevin, DP, “A conceptual model of elons: The organization as a refl ection of its entrepreneurship as fi rm behavior,” Entrepre- top managers,” Academy of Management neurship Theory and Practice, 1:7–25 (1991). Review, 9:193–206 (1984). 9. Weerawardena and Sullivan Mort, supra, n.5. 24. Sinkula, JM, Baker, WE, Noordewier, T, “A 10. Grenier, P, “The function of social entrepre- framework for market-based organizational neurship in the U.K,” available at: http:// learning: Linking values, knowledge, and www.istr.org/conferences/capetown/ behavior,” Journal of Academy of Marketing volume/grenier.pdf (2002). Science, 25:305–318 (1997). 11. Dees and Anderson, supra, n.5; Austin, et al., 25. Mair and Mart, supra, n.20. supra, n.5. 26. Venkataraman, S, “The distinctive domain of 12. Euske, NA, Euske, KJ, “Institutional theory: entrepreneurship research: An editor’s per- Employing the other side of rationality in non- spective,” In J. Katz and R. Brockhaus (Eds), profi t organizations,” British Journal of Man- Advances in Entrepreneurship, Firm Emergence, agement, 2:81–88 (1991). and Growth, CT: JAI Press: Greenwich (1997). 13. Kohli, AK, Jaworski, BJ, “Market orientation: 27. Brooks, AC, Social entrepreneurship: A mod- The construct, research propositions, and ern approach to social value creation, NJ: managerial implications,” The Journal of Mar- Pearson Prentice Hall (2009). keting, 54:1–18 (1990). 28. Kent, CA, Anderson, LP, “Social capital, social 14. Hurley, RF, Hult, GTM, “Innovation, market entrepreneurship and entrepreneurship edu- orientation, and organizational learning: An cation,” International Journal of Entrepreneur- integration and empirical examination,” The ship Education, 2:1–18 (2002). Journal of Marketing, 62:42–54 (1998). 29. Leana, CR, van Buren, HJ, “Organizational 15. Liu, SS, Luo, X, Shi, Y, “Integrating customer social capital and employment performance,” orientation, corporate entrepreneurship, Academy of Management Review, 24:538– and learning orientation in organizations- 555 (1999). in- transition: An empirical study,” Journal of 30. Burt, RS, Structural holes, Cambridge, MA: Research in Marketing, 19(4):367–383 (2002). Harvard University Press (1992). 16. Eikenberry, AM, Kluver, JD, “The marketiza- 31. Putnam, RD, “Tuning in, tuning out: The tion of the nonprofi t sector: Civil society at strange disappearance of social capital in risk?” Public Administration Review, 64:132– America. PS,” Political Science and Politics, 141 (2004). 28:664–683 (1995). 17. Weerawardena and Sullivan Mort,supra, n.5; 32. Luo, X, Griffi th, DA, Liu, SS, Shi, YZ, “The Grenier, supra, n.10. effects of customer relationships and social 18. Guo, KL, “Entrepreneurship management in capital on fi rm performance: A Chinese busi- health services: An integrative model,” Journal ness illustration,” Journal of International Mar- of Health & Human Services Administration, keting, 12(4):25–45 (2004). 28:504–530 (2006); McCleary, KJ, Rivers, 33. Stake, RE, Case studies in Denzin, NK, and PA, Schneller, ES, “A diagnostic approach Lincoln, YS, (Eds.), Handbook of Qualitative to understanding entrepreneurship in health Research, CA: Thousand Oaks: Sage (1994). care,” Journal of Health & Human Services 34. Patton, MQ, Qualitative Evaluation and Administration, 28:550–577 (2006). Research Methods (3rd ed.), CA: Newbury 19. Dees and Anderson, supra, n.5; Austin, et al., Park: Sage (2002). supra, n.5. 35. Kohli, AK, Jaworski, BJ, Kumar, A, “MARKOR: 20. Dees and Anderson, supra, n.5; Mair, J, Mart, I, A measure of market orientation. Journal of “Social entrepreneurship research: A source Marketing Research, 30(4):467–477 (1993). 46 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

36. Lu and Chiang, supra, n.7. 45. Lu, et al., supra, n.41. 37. Lu, JR, Hsiao, WC, “Does universal health insur- 46. Tiku, N, Do-gooder ﬁ nance: How a new crop ance make health care unaffordable? Lessons of investors is helping social entrepreneurs. from Taiwan,” Health Affairs, 22:77–88 (2003). INC. Magazine, February, 29–30 (2008). 38. Lu and Chiang, supra, n.7. 47. Nelson, WA, Donnellan, J, “An executive- 39. Wu, D, An ideal state—Dr. Delon Wu’s mem- driven ethical culture,” Health care Executive, oir. Taipei: Family Art Collection Co., Ltd: (in Nov/Dec:44–46 (2009). Chinese) (2004). 48. Brooks, supra, n.27. 40. Id. 49. Haglund, R, “Bundled care model at Spectrum 41. Lu, JR, Wu, S, Chuang, Y, Chang, Gung, Health may be roadmap for reduced health care “Memorial Hospital: The 43.55 percent puz- costs,” available at: http://www.crainsdetroit. zle for the performance and costing of the com/apps/pbcs.dll/article?AID=/20100915/ Department of Ophthalmology,” Chang Gung C03/100919947/0&template=prin University College of Management teaching tart&template=printart&template= case collection (2004). printart (2010). 42. Id. 50. Brooks, supra, n.27; Wei-Skillern, J, “Net- 43. Chuang, Y, Hwang, C, Health service manage- works as a type of social entrepreneurship ment series – management system in health to advance population health,” Preventive care organization, Taipei: Farseeing Publish- Chronic Disease, 7:1–5 (2010). ing Co., Ltd (2000). 51. Light, PC, “Social entrepreneurship revisited,” 44. Kong, W, “2011 Annual Report of Chang Stanford Social Innovation Review, Summer Gung Memorial Hospital,” January (2011). Issue:21–22 (2009). Health Care Reform and the Development of Health Insurance Plans: The Case of the Emirate of Abu Dhabi, UAE

Samer Hamidi, Dr. PH, Sami Shaban, PhD, Ashraf A. Mahate, PhD, and Mustafa Z. Younis

Introduction: The Emirate of Abu Dhabi has taken concrete steps to reform health insurance by improving the access to health providers as well as freedom of choice. The health system in Abu Dhabi is self- fi nanced through universal mandatory health insurance for all residents that has three main sources of fi nancing: employers or sponsors, the government, and individuals. Objectives: This article explores and discusses the policy inherent in private health care schemes that have been implemented by the Emirate of Abu Dhabi. Methods: Data was collected in early 2013 on health care plans in Abu Dhabi. The authors provide descriptive statistics to examine the progress and outcome of the introduction of the health insurance scheme to Abu-Dhabi. Results: The Abu Dhabi model has private sector involvement but the government sets prices and benefi ts. The Abu Dhabi model adequately deals with the problem of adverse selection through making insurance coverage a mandatory requirement. There are issues with moral hazards, which are a combination of individual and medical practitioner behavior that might affect the effi ciency of the system. Over time there is a general increase in the usage of medical services, which may be refl ective of greater awareness of the policy and its benefi ts as well as lifestyle change. Conclusion: Although the current health care system level of usage is adequate for the current population, as the level of usage increases, the government and other market participants may face an increase in their fi nancial burden in premiums, deductibles, and the amount of co-pays. The Emirate might need to explore the best model to sustain affordable, quality, health care services in either imposing a single- payer system, or government/public healthcare system. The other policy option is to introduce competition to the health care market, which is not easy to implement due to the nature of the health care sector in comparison to other sectors of the economy. Additional policies should be considered, such as adopting educational programs to encourage the population to conduct a healthy lifestyle and encourage precautionary actions. Key words: health system, health fi nance, health insurance, health plan, Abu Dhabi UAE

Ashraf A. Mahate, PhD, Dubai Export, Dubai Samer Hamidi, Dr.PH, is Associate Professor, Act- Department of Economic Development, Dubai, ing Dean at the School of Health and Environmen- United Arab Emirates. tal Studies, Hamadan Bin Mohammad e-University, Mustafa Z. Younis, is Tenured Professor of Health- Dubai, United Arab Emirates. care Finance & Financial Management, Jackson State University, Jackson, Mississippi. Sami Shaban, PhD, is Associate Professor of Health Informatics, United Arab Emirates University, Alain, J Health Care Finance 2014; 40(3):47–66 United Arab Emirates. Copyright © 2014 CCH Incorporated

47 48 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

1. Introduction tended to employ a safety net such as the US Patient Protection and Affordable Care Act, Effi cient and effective health care is the enacted in 2010, which seeks to increase cornerstone in the development of any coun- access by making insurance available to the try and most western governments during the 50 million Americans that are excluded from post-war period sought to provide a state-run public or private coverage.4 Market failure system. Under the state run health care sys- in the health care sector is also an impor- tem all individuals receive access to health tant area for developing countries especially care services. However, the last 60 years have those that are resource rich and have a rela- shown that health care is an area fraught with tively small population. One such example is many challenges. Health care by its nature is that of the UAE emirate of Abu Dhabi, which impacted by the increasing cost of care, infra- sought to ensure that all residents whether structure, and medicines.1 In most countries national or expatriate have medical insurance health care infl ation tends to be well above the by making it compulsory for employers to general infl ation level. As a direct result health provide a health care plan for their employ- care expenditure as a proportion of GDP has ees and families. The UAE is a small country been increasing in most developed countries.2 yet abundant with resources that allows it to At the same time, medical advances have easily afford to not only extend its provision allowed for an increase in the life expectancy of medical services but also allows for a state- of individuals, which has meant that govern- run health care system. However, two policy ments now need to spend more money on long- issues arise. First, the emirate of Abu Dhabi term care. The growing cost of health care and has to decide which government expendi- the impact of the global fi nancial crisis have tures to prioritize and which can be covered meant that many countries are no longer able by the private sector. Second, the emirate of to solely bear the cost. As a result many coun- Abu Dhabi like the rest of the UAE has a tries have sought to overhaul their health care particular problem in that the nationals are system and introduced some market-based a minority while the majority are expatriate system to share the burden of provision with workers and their families.5 In such a situa- consumers, the private sector, and employers.3 tion a government-funded program will ben- The difference between the private and efi t nonnationals more than nationals. public sector is that the former requires a Within the UAE, the emirate of Abu Dhabi fi nancial return on its investment while the was the fi rst to introduce a compulsory health latter tends to be operated as a nonprofi t insurance scheme for all residents of the organization. Therefore, the involvement of emirate in 2007. This insurance is fi nanced the private sector in the provision of health by premiums shared between employers and care plans invariably implies that there will employees, copayments, and coinsurances. be coverage but at a cost. The usual scenario The system was fl exible so as to allow resi- is that low-risk individuals or those who can dents of the other six emirates within the UAE afford to pay tend to receive medical cover- to join on a voluntary basis. The compulsory age while those who are at high risk or cannot insurance scheme was extended in 2008 when afford to pay are excluded from the system. the Health Authority of Abu Dhabi (HAAD) Due to this market failure governments have established the Thiqa program for UAE Health Care Reform and the Development of Health Insurance Plans 49

nationals managed by the National Health comprehensive health services. Through Insurance Company (DAMAN). The unusual implementing appropriate government poli- aspect of this health care plan was that it was cies or intervention strategies such as incen- mandatory for all UAE nationals to subscribe tives and regulations they can “conscript to but was provided free of charge. Further- private insurance to serve the public goal of more, this particular health care plan had an equitable access.”7 As an exploratory article extensive range of benefi ts and coverage. it does not have any prior assumption or In the post-global fi nancial crisis environ- hypothesis but simply seeks to understand ment, private health care insurance appears how the policy issues inherent in private to be a favored route for governments health care schemes have been dealt with by to promote. For policy makers, private the emirate of Abu Dhabi. At the same time, health care insurance allows them to pro- the article seeks to understand whether the vide fi nancial protection for their citizens development of the compulsory health care through prepayment while at the same time insurance has impacted the design and creating a risk-pooling mechanism that per- availability of policies. mits providers to make a return.6 However, This article consists of six sections of for providers to make a return they need to which the fi rst introduces the issues that will ensure that they can implement cost-control be explored in this study. In the next sec- mechanisms, which may include thinner tion we describe the health care system in benefi ts, increased copayments, and higher the emirate of Abu Dhabi. The third section deductibles. From a consumer perspective explains the role of private sector insurance these reductions in benefi ts and increased and the issues that typically arise that gov- copayments may not be aligned with their ernment policy needs to address. The fourth wants. The experience of many countries section discusses the aspects of the avail- shows that over time consumers tend to able health care plans in the emirate of Abu demand a greater variety of options. As a Dhabi. The fi fth section discusses the policy result, health care plans may start off being issues in the context of Abu Dhabi’s experi- the same but due to multiple benefi t con- ence. Finally, the sixth section provides the fi gurations and choice of provider networks conclusions and recommendations. they can develop into a large portfolio of schemes. Insurance markets by their nature are 2. The Health Care System in Abu Dhabi subject to various market failures, which open up government policy issues. The The Abu Dhabi health care system con- aim of this article is to examine the gov- sists of two main components: the Ministry ernment policy implications that arise from of Health (MOH) and the Abu Dhabi Health the move towards a compulsory health care Authority (HAAD); the former is a federal coverage system. In some countries these entity and the latter is only for the emirate policy issues have been dealt with through of Abu Dhabi. HAAD is fi nanced through direct government intervention so as to compulsory insurance contributions in promote the public health aims and objec- which costs are shared between employers tives of equity, affordability and access to and employees. A number of health reforms 50 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

have occurred recently to move Abu Dhabi revenue they generate from insurance, they towards achieving its vision for 2020. Some get direct funds from the Abu Dhabi Depart- of the key changes that have taken place are: ment of Finance both for ongoing operational costs as well as for capital projects. Most • Mandatory health insurance was intro- of the larger SEHA hospitals have interna- duced starting January 1, 2007. tional management partners (Johns Hopkins • The General Authority for Health International, Cleveland Clinic, and VAMed, Services (GAHS) was split into the for example). The dominant private provid- Health Authority Abu Dhabi (HAAD), ers operate hospitals that also serve as large which is the regulatory authority, and polyclinics for outpatient traffi c, the largest SEHA/Health Services Sector, which of which are Al Noor and NMC Hospital. manages public providers. This clear Mubadala-owned facilities, such as the Impe- separation of roles is intended to rial College London Diabetes Center and the enhance transparency and provide the Molecular Imaging Center, are in between basis for an open system. public and private. On the one hand they • The largest public hospitals now have receive such public provider privileges as not management contracts with interna- being subject to a copayment for pharmaceu- tional providers: ticals under the Thiqa Program; on the other • Sheikh Khalifa Medical City hand they independently negotiate prices sep- (SKMC) has a contract with the arately from SEHA as private facilities. Cleveland Clinic from the US. • Mafraq Hospital has a contract 2.2 Payers with Bumrungrad from Thailand. Medical insurance can be bought directly • Al Ain Hospital has a contract with from medical insurance companies or indi- VAmed from Austria. rectly through medical insurance brokers. • Tawam Hospital has a contract In 2013, there were 39 authorized insurance with Johns Hopkins from the US. providers, 50 authorized brokers, and 13 authorized third-party administrators (TPA) The health system fi nances itself through competing for this segment. A TPA is an mandatory health insurance for all Abu organization that processes insurance claims Dhabi Emirate residents. Providers are inde- or certain aspects of employee benefi t plans pendent and predominantly private. The for a separate entity. This can be viewed private sector receives its fi nancial support as “outsourcing” the administration of the from contributions by nationals and expa- claims processing, as the TPA tends to carry triates seeking their health services through out tasks traditionally managed by the com- mainly out-of-pocket payment. pany providing the insurance coverage. The payer market consists of 39 companies and is 2.1 Providers dominated by DAMAN, which almost exclu- Public providers are bundled under the sively administers the Basic Health Insurance SEHA umbrella and control a market share of Plan and the Thiqa Plan. In addition, it holds roughly one-third of the outpatient and two- the largest market share, about 30 percent of thirds of the inpatient sector. In addition to the the market. Other large players are regional Health Care Reform and the Development of Health Insurance Plans 51

companies such as Oman Insurance, ADNIC, mandatory minimum coverage offered at the and Green Crescent, or international groups government-subsidized price of AED 600 such as AXA. Many of the smaller players (US$163) per year, a premium that is deter- outsource their administration to TPAs. The mined by an executive decision from HAAD larger of these companies are NAS and Next- and administered by DAMAN and the other Care (part of Allianz Group). Although the 11 insurance companies. HAAD sets the health care insurance market in Abu Dhabi reimbursement rate (that is, the standard tar- has increased to more than one million mem- iff), and approves the price list for services bers it tends to be highly concentrated. Almost covered in the basic plan. Employers are 60 percent of the market is controlled by three required to provide health insurance to all payers: DAMAN (28.6 percent), Oman Insur- expatriate employees, their spouses, and up ance (16.4 percent), and Abu Dhabi National to three children (under age 18). Gulf Coop- Insurance Company (ADNIC) (14.5 percent). eration Council (GCC) nationals are exempt from this requirement. Expatriate employ- 2.3 Health Insurance Plans in Abu Dhabi ees’ health insurance must provide basic There are four broad categories of health coverage including hospitalization, medi- insurance plans in Abu Dhabi: Thiqa, Basic, cal exams, treatment, primary care, tests, Enhanced, and Visitor. In addition to the X-rays, dental care (excluding orthodon- Thiqa program, which is specifi cally for tics and dentures), prescription drugs, and nationals, laws and regulations established accommodation fees for family members or three types of private health insurance plans other caregivers. that can be sold to nonnationals: The maximum limit for basic health ser- 1. The Basic product, for individuals vices per insured person (that is, the ceiling) with limited income and the depend- is AED 250,000 (US$68,120) annually. The ents of nonnationals who are not eligi- plan levies a deductible of AED 20 (US$5.50) ble to be covered by the nonnational’s per outpatient visit to the general practitioner, employment-based insurance. and AED 10 (US$2.25) for a specialist when 2. The Enhanced product, for individuals referred by an approved general practitioner. above the income threshold set by the The plan also levies a copayment of AED regulations for the Basic product and 10 (US$2.25) per laboratory test or radiology available to all nonnationals. diagnostic services (including MRI & CT). In 3. The Visitor product, for visitors to the the case of pharmaceuticals the patients need Emirate holding certain types of visas. to pay 30percent (the coinsurance amount), which is capped at AED 1,500 (US$411) per year after which the insurer must pay the 2.3.1 Basic Health Insurance Plan full charge of the prescription. The policy is The basic plan is for expatriates with a total limited to the emirate of Abu Dhabi although monthly salary package of under or equal emergency cases are covered in the whole of to AED 4,000 (US$1,090) with housing the UAE. Furthermore, the policy provides allowance, or AED 5,000 (US$1,363) when maternity coverage at full cost within the housing is not provided by the employer. network with an AED 500 (US$137) deduct- Nonnationals must enroll in the basic plan as ible for each delivery. 52 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

2.3.2 Enhanced Health Insurance Plan 2.3.3 Thiqa Health Insurance Plan Expatriates who are not eligible for Thiqa is administered by the national the basic plan can obtain medical cover- health insurance company, DAMAN, age through purchasing the enhanced plan. and is regulated by HAAD. Thiqa Health Under the regulations, insurance provid- Insurance Program is provided to all UAE ers that offer the enhanced policy cannot nationals working and residing in the emir- exclude any of the benefi ts that are available ate of Abu Dhabi with comprehensive, free under the basic policy. In addition to what health care coverage at all public and pri- is offered in the basic policy, insurance pro- vate hospitals registered within DAMAN’s viders are required to make available a mini- network. All Thiqa members can use health mum of two signifi cant enhancements, such care services from more than 1,600 medical as increasing the upper limit coverage, geo- providers including health care facilities in graphical area, inpatient services, outpatient the emirate of Abu Dhabi and facilities out- services, and so on. For outpatient services, side the UAE for emergency cases. Thiqa one of the following two options may be members are eligible to receive outpatient selected: Option (1) remove the deductible coverage, day treatment, and inpatient cov- amount or coinsurance on all the elements of erage with a private room with one bed. the outpatient health care services and keep About 400,000 citizens are covered under the coverage for pharmaceuticals at AED this scheme as of 2013. 1,500 per year; or Option (2) increase the DAMAN, a government-owned, special- coverage for pharmaceuticals to AED 3,000 ized health insurance company that was set per year with a coinsurance of 15 percent. up in 2005 was assigned to manage the Thiqa Also, providers cannot restrict the upper age plan. The program is provided by the gov- limit for insurance coverage. Finally, the ernment of Abu Dhabi through the Depart- amount of the deductible cannot exceed a ment of Finance. UAE nationals through maximum of AED 50 for outpatient services Thiqa Health Insurance Program have the other than medicines. following benefi ts: The premium for the enhanced policy is to be determined by market rates and is • Services provided outside the emirate based on many factors such as age, gender, of Abu Dhabi are subject to 10 percent maternity and dental coverage, geographical coinsurance. coverage, and medical conditions or preex- • Nationals pay a coinsurance of 50 isting diseases. Members older than 55 years percent for dental treatment and or children younger than 7 months need to pharmaceuticals in all private-sector submit a medical report describing their facilities. health status. DAMAN covers dental treat- • Pharmaceuticals are free if prescribed ment for groups who are part of enhanced in a private facility but obtained in a plans such as UAE Regional, International, public facility. and Global Plans. For individuals, dental • The maximum limit for health services services can be added only to International per insured person (that is, the ceiling) and Global plans. is AED 500,000 (US$137,000) per year. Health Care Reform and the Development of Health Insurance Plans 53

• The territorial limit of the policy is through their corresponding sponsors. All restricted to the UAE. policies with the exception of the emergency • Emergency cases are covered in a speci- policy are valid for at least one year and fi ed list of countries. should be renewed accordingly. The insured • Maternity is fully covered within is not entitled to claim back the premium. network. • Dental benefi ts are fully covered 3. The Concept of Private Insurance within network with no coverage outside network. Governments that have sought to bring • Second opinion service is provided by in medical insurance schemes have done so Europe Assistance GCS. on the premise that it provides citizens with • Annual Breast Cancer Screening (appli- access to medical care while limiting their cable for females >35 years in desig- fi nancial exposure to the insurance premi- nated network) is covered. ums and the copayments that may need to • Annual Prostate Cancer Screening be paid. The three functional components are (applicable for males >45 years in des- the collection of premiums, pooling of funds, ignated network) is covered. and purchasing of medical services through UAE nationals working and living in Abu the establishment of a supplier network. In Dhabi are required to have a Thiqa card and a post-paid system whereby the insurance should not purchase other health insurance company pays the customer upon receipt of plans. Employers seeking to provide their the invoices there is no real need to establish a UAE national employees with more health network but the essential activity of purchas- insurance benefi ts can still purchase a “Thiqa ing services stills exists. Technically, this per- Top-Up” that adds to the existing benefi ts of tains to whether the insurance is provided by Thiqa. International Patient Care (IPC) is a a private or publicly owned scheme although scheme for treatment abroad also covered in the latter the premiums can be collected by the government. In addition, the govern- through tax. Tax is not the only differentiating ment funds defi ned mandates for health care factors between a private and publicly owned services and programs that serve the public scheme and Sekhri, Savedoff and Thripathi good and that are not covered by the Health (2005) have developed three parameters that Insurance Scheme (that is, funded mandates). seek to understand the difference between the two as fi nancial innovation and the need to 2.3.4 Visitor Plans transfer costs from governments to citizens has tended to blur the distinction between For those on a visit visa, the visitor plan the two types of medical coverage.8 Sekhri, provides an aggregate limit on covered care Savedoff and Thripathi (2005) have catego- of no less than AED 100,000 (US$27,400). rized the three aspects as follows:9 The premium is to be determined on the basis of the duration of stay and market rates. Cur- • Enrolment: whether the insurance is rently, 13 companies provide visitor prod- mandatory or voluntary. ucts. The visitor plan is provided for visitors • Underwriting/pricing: whether con- or for those with visit visas and purchased tributions are risk-rated (minimal risk 54 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 1. Private and Publicly Funded Insurance Schemes

Factor Privately Funded ← → Publicly Funded Enrollment VVVVMVMVMMandatory Pricing RRCCCICI IIncome Rated Management NPC NPC NPT G NPC NP NPC G NPC Public V=voluntary, M=Mandatory, R=Risk Rated, C=Community Rated, I=Income Rated, P=For Profi t, N=Nonprofi t, NPC=Nonprofi t Commercial, NPT=Nonprofi t Community, G=Public

Source: Sekhri, Savedoff and Thripathi, 2005.10

transfer), community-rated (transfers operate in the case of Abu Dhabi.11 What between healthy and sick), or income is interesting about the Abu Dhabi health- based (transfers between higher- and care model is that it is a mixture of a state lower-income individuals). benevolent scheme for UAE nationals under • Organizational structure: whether man- the Thiqa scheme whereby the premiums are agement of the scheme is commercial paid by the government. DAMAN, which for-profi t, private non-profi t, or public/ is the company that manages the Thiqa quasi-public. scheme, has exclusive access to the public providers in Abu Dhabi. Although DAMAN Based on these three aspects they have is owned by the government, it is operated as developed a spectrum of different possibilities a commercial enterprise. At the same time, that can arise, which are illustrated in Figure 1. expatriates must subscribe with one of the Under this framework the two polar 39 insurance providers all of whom, except extremes are the privately and publicly funded for DAMAN, are privately owned. Some are programs. Within these two extremes one has nonprofi t, commercial enterprises such as schemes in which enrollment is either volun- the UK company Bupa while others are pri- tary or mandatory. As the programs become vate, for-profi t companies such as ADNIC or publicly funded the enrollment tends to be AXA. In this respect, Figure 2 summarizes mandatory, and, in contrast, voluntary for the various health care plans available in privately funded ones. The publicly funded Abu Dhabi. programs tend to have an income-rated pric- The unique nature of the Abu Dhabi ing mechanism while private ones are risk- scheme makes not only an interesting study rated. This is the key difference between the but allows us to extend the Sekhri, Savedoff private and publicly funded programs in that and Thripathi (2005) framework. As men- the latter seek to be inclusive while the for- tioned above, Abu Dhabi has a small pop- mer have the room for pricing low-income ulation of UAE nationals and a high level individuals out of the market. of income from its investments as well as Sekhri, Savedoff and Thripathi (2005) hydrocarbons. As a result Abu Dhabi is in describe a convenient framework that can a privileged position to fully fund medical help us understand the different ways in care for its people. In addition to receiving which private and public insurance schemes free in-country care, UAE nationals are also Health Care Reform and the Development of Health Insurance Plans 55

Figure 2. Summary of Health Insurance Plans

Product Beneficiaries Payers Premium Paid by Reach Who sets prices Basic Low income DAMAN 600 AED Employers, cover- Abu Dhabi HAAD: basic expats plus 11 ing their employ- (plus price list other ees and families UAE for companies (1 spouse and 3 emergency children under 18) Enhanced Higher 39 Risk Employers, cover- Varying Negotiation income insurers adjusted ing their employ- between payer expats (largely ees and families and provider. group (1 spouse and 3 HAAD sets fl oor business) children under 18) and ceiling Thiqa Nationals Depart- Free DAMAN, which is UAE plus HAAD (Currently ment of the TPA worldwide set to be the Finance in emergency same as DAMAN Abu Dhabi Network 1) entitled to apply for out-of-country treat- is that of “un-insurable risk”. That is, the risk ment. The out-of-country treatment is pro- that an insurance company is not willing to vided when such facilities and resources do cover due to either information asymmetry, not exist within the UAE. This ensures that inability to control the insured individual, UAE nationals receive the best treatment low profi t, and so on,13 argues that the con- whether it is within or outside the country cept of insurable risk requires that any losses at no cost to them or their employer. As far arising cannot be controlled by the insured as expatriates are concerned, the government and are purely accidental as well as unpre- has established an income-based approach dictable. Furthermore, the risk for a particu- whereby the lowest paid are subsidized and a lar individual may be unpredictable, but for a cost ceiling is applied for the higher income population it needs to be predictable. Under individuals. This dual paternalistic approach these circumstances an insurance provider is is unique and sets it apart from all other able to provide affordable coverage. In the examples globally. Figure 3 places the Abu case of health care insurance, the risks to a Dhabi health care model within the Sekhri, certain extent are in the control of the indi- Savedoff and Thripathi (2005) framework vidual through lifestyle and behavior, such for comparison purposes.12 as obesity and smoking-related cancer, for example. This implies that the current situa- 3.1 The need for policy intervention tion may not be the case in the future, leading in the health insurance markets to unpredictability in assessing the insurance Insurance markets tend to suffer from provider’s risk exposure. As a result of this, market failure, which provides a strong argu- health care insurance markets fi nd it diffi cult ment for intervention by the government to provide coverage for long-term health either directly or through a regulator. The risks, leading to a gap in the market provi- fundamental problem with private insurance sion and a need for government intervention. 56 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 3. The Abu Dhabi Health Care Model

Funding Privately Funded ←→ Publicly Funded Nationality Expatriates UAE Nationals Factor Low Income Higher Income Basic With additional benefi ts Enrollment Mandatory Mandatory Mandatory Mandatory Pricing Government set and Risk rated but government Government Individuals pay extra for subsidized sets a fl oor and ceiling set and paid the additional benefi ts Management Both private and Both private and publicly Nonprofi t Nonprofi t public publicly owned entities owned entities public

Typically insurers tend to have less infor- ance coverage they tend to use medical mation regarding an individual’s health than services more often than if they were not the health care provider. In order to limit their insured.14 Such an inability to control the own fi nancial exposure health care insurance actions of individuals after purchasing insur- providers tend to set the premiums higher than ance is called “moral hazard.” In health care they need to be. When medical insurance is the problem of moral hazard is made worse voluntary, then only those who need the cov- by the fact that the medical practitioner can erage (that is, the sick) will purchase it even at overprescribe medications or conduct unnec- the higher price, leading to a pool of high-risk essary tests and investigations. Health care individuals. The greater the risk, the higher providers can take action to reduce the mis- the cost of the medical coverage, which will use through copayments for customers and exacerbate the problem, leading to greater case rates for medical providers. However, adverse selection. Adverse selection not only the latter may also lead to poor quality care as increases the payouts for the insurance com- health care providers seek to maximize their pany but also threatens its survival. Therefore, profi ts. Therefore, government intervention health care providers tend to discourage sick is required to ensure that under-provision of customers from purchasing insurance or fi nd care does not take place. At the same time ways to insure only lower-risk individuals. government intervention needs to ensure that This leaves a pool of sick individuals that medical practitioners do not abuse the insur- have no private insurance company willing to ance coverage through conducting needless insure them regardless of their willingness to tests resulting in higher future premiums. pay for the coverage. From a public concern, Another reason for government inter- government intervention is required so that vention in the provision of health care all citizens receive adequate medical cover- plans deals with negative externalities. For age. One of the possible interventions is to instance, from a government perspective make insurance mandatory and force insurers greater access to health services is consid- to accept all applicants or to subsidize the cost ered a merit. However, the consumption of of care for the high-risk individuals. health care services by one individual may Health care insurance providers have actually have a negative impact on another. found that when individuals purchase insur- Therefore, government intervention seeks Health Care Reform and the Development of Health Insurance Plans 57

to readdress this imbalance so that the good our opinion, we believe that HAAD should aspects can be promoted. Eventually, lack of encourage more hospitals to open in the emir- access to health care affects other areas of ate so that the disciplinary effect is strong the economy such as productivity in work- and the benefi ts to the patients are realized. places. Government intervention in this area From the viewpoint of patient coverage can take place in the form of requiring health we found that the Abu Dhabi system has care insurance providers to include a pack- successfully managed to create a health care age of health services that are considered system that does not exclude any single indi- to be in the public interest. This interven- vidual regardless of age or health risk. At the tion can also include subsidizing the cost of same time the ceilings ensure that the lower health care for those on a low income. income households are not disadvantaged through heavy deductibles and coinsurance 4. Discussion payments. Interestingly, the Abu Dhabi system also takes into consideration the income Market Failure levels of the bulk of its residents and has set One of the interesting aspects of the hospi- a reasonable limit on pharmaceuticals. The tal sector in Abu Dhabi is that it is highly con- pharmaceutical limit is currently at a point at centrated in terms of the number of providers which coinsurance is about a month’s salary as well as the owners. More importantly, only of an average laborer. This implies that those one insurance company (DAMAN) is able to who need pharmaceuticals can have access to contract with government- owned hospitals. them while ensuring that their usage is carried This has two implications. First, the level of out in a reasonable manner through codeduct- provision offered is less than what may be ibles at the lower level. As far as specialized possible under a fully competitive system. treatment is concerned, this is available to all Second, from a patient viewpoint, not being nationals outside the country and includes the able to use all the hospitals implies that the cost of travel of one additional family mem- access to the level of treatment is not equal. ber for the duration of the overseas trip. Thus, there is no real competition between the hospitals and each has some level of Adverse Selection monopolistic power. As a result, hospitals One of the key issues in any insurance plan are able to charge a higher fee than they nor- is the problem of adverse selection or what is mally would with open competition. One sometimes referred to as negative selection. way of increasing the level of competition The basis of adverse selection is that there is in the hospital sector is simply to extend the asymmetric information between the provid- ability of all insurance providers to contract ers of insurance and the buyers so that an indi- with government hospitals. Even if there is vidual’s demand is positively correlated with little competition currently in the market, it his or her risk of loss. As a result one fi nds is important for the present set of hospitals that those with the greatest risk of loss tend to to know that the market is open and there be those that purchase insurance leaving the could be new entrants in the near future. This insurance provider with a pool of individuals type of action is referred to by Vickers and that are a bad risk. In some cases the insur- Yarrow (1988) as the disciplinary effect.15 In ance provider may be allowed to correlate the 58 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

price with the risk due to government regula- our analysis shows that two companies— tions. This situation is referred to as regulatory DAMAN and ADNIC—cover almost half adverse selection.16 We found that regulatory the expatriate population while the rest have adverse selection does exist in the case of the market shares of typically 2 percent. This Abu Dhabi model in that insurance providers implies that the larger two companies have cannot set their own rates based on the indi- portfolios that are well diversifi ed while the vidual risk profi le. Also, insurance providers smaller providers may be exposed to greater cannot refuse a particular individual from risk. We believe that in time some of the obtaining medical coverage based on age, smaller providers may fi nd it unprofi table or other factors. Interestingly, market access to be in the market and may exit. Although and adverse selection are confl icting issues concentration in a regulated insurance market in that the former seeks not to exclude any may not be undesirable it should nevertheless individual from obtaining medical coverage be managed, and procedures should be put in while the latter favors a risk-based pricing place so that individuals are covered in case model in which some can be excluded on this of a company going out of business or being basis. In the case of the Abu Dhabi model the taken over. issue of adverse selection has been reduced by creating large enough pools of individuals Moral Hazard so that the overall portfolio risk is reduced. In Another important issue in insurance is the case of nationals they can purchase only that of moral hazard whereby an insurance the Thiqa policy through a single supplier. In provider cannot control the actions of the this manner, we found that the average risk is insured once the policy has been taken out. reduced. As shown in Figure 4 the member- For example, individuals with insurance may ship of the Thiqa policy now covers almost have a tendency to take risks that they would the entire national population of Abu Dhabi, not normally take because any incurred costs and as such the portfolio risk is lower than it will not be borne by them but rather by the would be had the government not mandated insurance provider.18 The essence of moral a single company to provide such coverage. hazard lies in the insureds not taking respon- In the case of expatriates we found that as it sibility for their actions leading to a fi nancial is mandatory to have medical insurance there loss by the insurance provider. This is a spe- is equal probability that those with higher risk cial case of information asymmetry because are spread across all the companies. However, the insured has more information of the post-

Figure 4. Number of Members in Each Plan

Type of Health Insurance Plan 2008 2009 2010 2011 Basic 944,344 936,207 1,204,418 1,322,804 Enhanced 932,610 981,743 1,044,734 1,053,893 Thiqa 383,795 394,618 422,239 442,261 Total 2,260,749 2,312,568 2,671,391 2,818,958 Source: HAAD Health Statistics 201117 Health Care Reform and the Development of Health Insurance Plans 59

insurance behavior than the insurance pro- illnesses. We feel that there are two reasons vider. Two issues arise from the problem of that may be able to explain the possible dif- a moral hazard in that it is a behavioral issue ferences in usage and corresponding cost. and not that every individual is likely to act The basic plan has deductibles that, relative in an irresponsible manner. The reason for to the income levels of individuals, are high. this is that essentially the underlying aspect In the case of the Thiqa the deductibles are of acting irresponsibly comes down to ethics, negligible compared to average income lev- and this differs from individual to individual. els for that group. Therefore, we believe that In a large portfolio there will be an averaging the higher propensity to use medical services factor, which implies that it may suffer from arises from the lack of a controlling factor. individuals that act in an irresponsible behav- Second, medical practitioners may overen- ior but also individuals that do not. The sec- courage individuals on the Thiqa policy to ond problem is that medical providers may revisit them for checkups, etc. The difference encourage individuals to undergo tests and in the usage of medical services is an area of treatment that are unnecessary so as to drive further research that needs to be addressed. up their own revenue. Here, the irresponsi- The second striking feature in Figure 5 is ble behavior is derived from the actions of that the usage of outpatient medical treat- the medical provider and not the individuals ment has increased for all policy types in the themselves. In this case the size and diversity period 2009 to 2011 while inpatient treat- of the portfolio cannot by itself reduce the ment has remained fairly constant. We found problem of moral hazard but requires con- that basic policy holders have increased their trols placed on the medical providers. usage of outpatient treatment from about two Figure 5 illustrates the inpatient and outpa- visits a year to three while usage for those on tient usage of medical services by the policy the enhanced and Thiqa policy has increased type. The most striking feature is that indi- by about 25 percent over the period 2009 to viduals on the Thiqa policy make more than 2011. The increased usage especially for the four times the number of inpatient claims as basic policy holders we believe is largely do those on the basic and enhanced policy. due to a greater awareness of the policy and In the case of outpatient claims it is four and the services available. This may have altered half as much as the basic and three times that the individual’s behavior; previously they of the enhanced policy holders. Individuals may have taken an over-the-counter medi- on the basic policy visit a medical doctor cine to deal with an ailment but now choose for outpatient treatment about once every to consult a medical practitioner. We believe quarter, which increases to bimonthly for that with greater awareness the usage levels enhanced policy holders and more than once will increase and this may impose a fi nancial a month for those on the Thiqa policy. Such burden as far as the subsidy is concerned. We a striking difference between patient groups believe that a long-term solution is to edu- is unusual especially as there is an inverse cate individuals with regard to a healthier correlation with income in that lower income lifestyle and taking precautionary actions. groups use fewer medical services. This is We found that the average cost of treat- not typical as lower income households tend ment for those on the basic and enhanced to suffer from a greater level of poverty-based policies is roughly similar. However, when 60 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 5. Usage and Claims by Policy Type

Average # of Cost/Claim Claims # of Claims per member (AED) (1,000) 2011 2011 2009 2010 2011 Basic 9,006 38 0.03 0.03 0.03 Enhanced 9,380 35 0.05 0.05 0.03 Inpatient Thiqa 12,939 59 0.18 0.18 0.13 Average 10,861 131 0.06 0.06 0.05 Basic 4,138 154 2.25 2.43 3.13 Enhanced 4,860 342 4.00 3.97 4.61 Outpatient Thiqa 6,199 364 11.2 13.8 14.0 Average 15,197 300 4.52 4.84 5.39 Both Average 15,328 390 4.58 4.90 5.44

Source: HAAD Health Statistics 2011.19 compared to the Thiqa policy there appears Thiqa policy covering all the nationals in the to be an increase of about 50 percent. As emirate. As such, nationals essentially do not far as inpatient treatment is concerned the have the privilege of consumer choice. In the higher cost can be explained in terms of a case of the expatriate population, both basic single room compared with two or more and enhanced policy holders have a choice patients sharing the same room. But in the of 16 and 39 companies respectively. As case of outpatient treatment such differences the cost of the policy is government deter- should not exist as the treatment should be mined, companies compete on the basis of fairly similar. We believe that there may be nonprice factors giving consumers consider- evidence to support the claim that in the case able choice. This is more so the case with of Thiqa policyholders there may be medical enhanced policies where 39 companies offer practitioner-induced moral hazard. In other 9,450 different policies. However, we found words Thiqa policyholders may be encour- that two companies control 46 percent of the aged to conduct additional tests or treatment enhanced policy market leaving a further that may not be essential. 17 companies to control 51 percent of the market that is an average of 3 percent per Competition company. The remaining 20 companies have The appendix lists all 39 providers of about 1 percent of the market that is an aver- medical coverage in Abu Dhabi along with age of 0.05 percent per company. It is rather the types of policies offered. With such unusual for two companies to dominate a large number of providers one would the market especially when the policy ben- assume that there is perfect competition in efi ts are largely dictated by the government. the market; however, this is not the case. Interestingly, we do not fi nd any correlation We found that only one company offers the between market dominance and the number Health Care Reform and the Development of Health Insurance Plans 61

of plans offered or the level of coverage. However, we believe that as the level of We believe that there are additional factors usage increases and the current population such as “fi rst mover advantage” or even non- ages this may pose a fi nancial burden on the market-based factors, such as customer ser- government. Although we do not propose a vice and ability to deal with claims, that may risk-based pricing model due to the fact that explain why these two companies are able to it may price out individuals, we neverthe- capture such a large share of the market. less believe that the government needs to place safeguards in order to limit its expo- 6. Summary of Results, Conclusion, sure. These safeguards may be transferring and Recommendations the risk of the additional cost to the employers. In addition, we believe that the market This exploratory study was designed to for medical treatment needs to be made analyze the provision of medical coverage in more competitive so as to ensure that they the emirate of Abu Dhabi within the frame- are not behaving in a monopolistic manner. work of the Sekhri, Savedoff and Thripathi Also, we believe that through greater com- (2005) model.20 We found that the Abu petition there will be a lower likelihood of Dhabi model is unique with a small local price infl ation and increased probability of population and the need to provide the best effi ciency enhancements. medical treatment. As a result, the govern- We found that the Abu Dhabi model ade- ment has established a publicly owned but quately deals with the problem of adverse privately managed company to provide med- selection through making insurance cover- ical coverage to all nationals with predefi ned age a mandatory requirement. As a result benefi ts at no cost and additional features to we found that the portfolio of any insur- be paid by the insured. In addition, nationals ance provider to cover those who are good receive out of country treatment when such and bad risks and thus reducing the average a facility or expertise does not exist in the exposure of the company. This is more so UAE. In the case of expatriates, two poli- the case with nationals because only a single cies are available provided by 39 companies company is able to provide the Thiqa policy. based on income levels and subsidized by We do not fi nd that such a system negatively the government of Abu Dhabi. As such, the impacts consumer sovereignty or the abil- Abu Dhabi model has private sector involve- ity of the individual to decide which comment but government-set prices and lists of pany they wish to be insured under. Also, benefi ts. we believe that additional features may be We found that the Abu Dhabi model seems restrictive as there is no competition. We to have dealt with the key issue of market believe that additional providers offering the access whereby insurance providers cannot Thiqa policy may enhance effi ciency as well refuse coverage for a particular individual as the level of consumer sovereignty. and there exists renewal guarantee. We also This study fi nds that there are issues with found that the limits on pharmaceuticals moral hazard in the Abu Dhabi health care imply that lower-income individuals are pro- model that are most pronounced with the tected and can continue with their treatment Thiqa policy. We believe that the issue of without deductibles beyond a certain point. moral hazard is a combination of individual 62 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

as well as medical practitioner-induced to make individuals aware of a healthier life- behavior. We found that on average Thiqa style and encourage precautionary actions. policy holders use outpatient medical ser- Our view is that the Abu Dhabi model has vices a little more than once a month com- made considerable inroads in blending public pared with once a trimester and quarter for as well as private provision of medical ser- basic and enhanced policy holders respec- vices and insurance coverage. However, we tively. We also found that Thiqa policy- feel that the system needs to be refi ned so that holders use four times the level of inpatient it is more effi cient while ensuring regulatory services compared to those on the basic and control. In this study we have outlined some enhanced plans. We feel that this is an unu- of the ways in which effi ciency gains can be sual abnormality that needs further research made as well as limiting the fi nancial exposure to understand the underlying factors as to of the government to future costs. Regarding why Thiqa policyholders have a far greater the issues of regulation, we believe that some tendency to use medical services com- oversight needs to take place to ensure that pared to the rest of the population. Finally, any medical practitioner-induced moral haz- we found that over time there is a general ard is limited. We also believe that the medi- increase in the usage of medical services and cal insurance market requires a specialized this may be refl ective of greater awareness of regulator to oversee the safe and responsible the policy and its benefi ts as well as a change conduct of medical insurance providers. We in lifestyle. We believe that with time this feel that a market dominated by two players will impose a fi nancial burden to the govern- and 20 companies sharing one percent of the ment and the Abu Dhabi government needs market poses risks of corporate failure. Health Care Reform and the Development of Health Insurance Plans 63

Appendix A: Details on Health Care Coverage

Table A1: Number of Providers and Plans

Insurance Company Basic Enhanced Visitor Thiqa Grand Total Abu Dhabi National Insurance Company (ADNIC) 1,543 1,543 Abu Dhabi National Takaful Co. P.S.C 1 103 104 Al Ain Ahlia Insurance Co. 113 113 Al Buhaira National Insurance Co. - Abu Dhabi 1 369 370 Al Dhafra Insurance Co. 1 544 4 549 AL Fujairah National Insurance Company 35 35 AL HILAL TAKAFUL - PSC 1 227 228 Al Khazna Insurance Company 1 298 299 AL Sagr National Insurance Co. - Abu Dhabi 1 252 1 254 Al Wathba National Insurance Co. (P.J.S.C) 292 292 Al-Ittihad Alwatani General Insurance Company 63 63 Alliance Insurance - P.S.C - Abu Dhabi 27 27 American Life Insurance Co. 567 567 Arab Orient Insurance Co. - P.S.C - Abu Dhabi 430 430 Arabia Insurance Company S.A.L 1 40 41 Arabian Scandinavian Insurance Co. P.L.C. 1 11 12 AXA INSURANCE - GULF - BSC- C- ABU DHABI 443 443 DAMAN (National Health Insurance Company) 1 453 9 15 478 Dubai Insurance Company - Abu Dhabi 1 43 44 Dubai Islamic Insurance & Reinsurance - AMAN 1 97 98 Dubai National Insurance and Reinsurance Co. 31 31 Emirates Insurance Co. 375 375 Green Crescent Insurance Company 848 848 INSURANCE HOUSE-PSC 1 64 65 Islamic Arab Insurance Company - SALAMA 44 44 Methaq Takaful Insurance P S C 120 120 National General Insurance Co. ( P.S.C ) 202 202 NOOR TAKAFUL FAMILY - PJSC - Abu Dhabi 72 72 Oman Insurance Company Limited – Abu Dhabi 817 817 Qatar Insurance Co. – Abu Dhabi 214 214 Ras Al Khaimah National Insurance Co. - Abu Dhabi 98 98 Royal & Sun Alliance Insurance (Middle East) 203 203 Saudi Arabian Insurance Company Ltd 118 118

Continued ... 64 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Insurance Company Basic Enhanced Visitor Thiqa Grand Total Sharjah Insurance Co. - Abu Dhabi Branch 1 10 11 Takaful Emarat Insurance P.S.C - Abu Dhabi 1 126 127 Union Insurance Company - Abu Dhabi 1 28 29 United Insurance Company 1 75 76 Watania (National Takaful Company) 37 37 Zurich Insurance Middle East S A L-Abu Dhabi 18 18 Grand Total 16 9,450 14 15 9,495

Source: HAAD Health Insurance Search Web site, https://bpmweb.haad.ae/Insurance/SearchInsurance/ SearchInsurance.aspx, accessed July 1, 2013.

Table A2: Average Annual Coverage in the UAE (AED) (Conversion: US$1 = AED 3.65)

Insurance Company Basic Enhanced Visitor Thiqa Grand Total Abu Dhabi National Insurance Company 732,404 732,404 Abu Dhabi National Takaful Co. P.S.C 250,000 1,126,881 1,118,450 Al Ain Ahlia Insurance Co. 2,743,615 2,743,615 Al Buhaira National Insurance Co. - Abu Dhabi 250,000 331,911 331,689 Al Dhafra Insurance Co. 250,000 352,513 100,000 350,486 AL Fujairah National Insurance Company 262,857 262,857 AL HILAL TAKAFUL - PSC 250,000 476,872 475,877 Al Khazna Insurance Company 250,000 357,383 357,023 AL Sagr National Insurance Co. - Abu Dhabi 250,000 317,460 100,000 316,339 Al Wathba National Insurance Co. (P.J.S.C) 710,959 710,959 Al-Ittihad Alwatani General Insurance Company 314,619 314,619 Alliance Insurance - P.S.C - Abu Dhabi 464,815 464,815 American Life Insurance Co. 582,606 582,606 Arab Orient Insurance Co. - P.S.C - Abu Dhabi 6,109,784 6,109,784 Arabia Insurance Company S.A.L 250,000 509,800 503,463 Arabian Scandinavian Insurance Co. P.L.C. 250,000 309,091 304,167 AXA INSURANCE - GULF - BSC- C- ABU DHABI 1,239,436 1,239,436 DAMAN (National Health Insurance Company) 250,000 10,440,033 122,222 500,000 10,154,903 Dubai Insurance Company - Abu Dhabi 250,000 3,450,837 3,378,091 Dubai Islamic Insurance & Reinsurance - AMAN 250,000 406,701 405,102

Continued ... Health Care Reform and the Development of Health Insurance Plans 65

Dubai National Insurance and Reinsurance Co. 300,000 300,000 Emirates Insurance Co. 1,147,240 1,147,240 Green Crescent Insurance Company 770,578 770,578 INSURANCE HOUSE-PSC 250,000 371,875 370,000 Islamic Arab Insurance Company - SALAMA 452,273 452,273 Methaq Takaful Insurance P S C 348,550 348,550 National General Insurance Co. ( P.S.C ) 441,807 441,807 NOOR TAKAFUL FAMILY - PJSC - Abu Dhabi 634,028 634,028 Oman Insurance Company Limited – Abu Dhabi 531,832 531,832 Qatar Insurance Co. – Abu Dhabi 2,126,246 2,126,246 Ras Al Khaimah National Insurance Co. - Abu Dhabi 516,423 516,423 Royal & Sun Alliance Insurance (Middle East) 4,779,906 4,779,906 Saudi Arabian Insurance Company Ltd 17,167,894 17,167,894 Sharjah Insurance Co. - Abu Dhabi Branch 250,000 300,000 295,455 Takaful Emarat Insurance P.S.C - Abu Dhabi 250,000 370,635 369,685 Union Insurance Company - Abu Dhabi 250,000 316,071 313,793 United Insurance Company 250,000 268,667 268,421 Watania (National Takaful Company) 624,324 624,324 Zurich Insurance Middle East S A L-Abu Dhabi 344,444 344,444 Grand Total 250,000 1,716,359 114,286 500,000 1,711,135

Source: HAAD Health Insurance Search Web site, https://bpmweb.haad.ae/Insurance/SearchInsurance/ SearchInsurance.aspx, accessed July 1, 2013.

REFERENCES

1. Everybody’s Business: Strengthening Health 3. Smith, E, Brugha, R, Zwi, A, “Working with Systems To Improve Health Outcomes, WHO’s Private Sector Providers Better Health Care: Framework For Action (2007), http://www. An Introductory Guide,” World Health who.int/healthsystems/strategy/ everybodys_ Organization, 2001, http://www.who.int/ business.pdf, accessed 30/7/2013. management/partnerships/private/private- 2. Erixon, F, van der Marel, E, “What Is Driv- sectorguide.pdf, accessed 30/7/2013. ing The Rise In Health Care Expenditures? 4. Gruber, J, “The Impacts of the Affordable An Inquiry into the Nature and Causes of Care Act: How Reasonable Are the Projec- the Cost Disease,” ECIPE Working Paper tions?” NBER Working Paper No. 17168, No. 5, 2011, http://www.ecipe.org/media/ 2011, http://www.nber.org/papers/w17168, publication_pdfs/what-is-driving-the-rise-in- accessed 30/7/2013. health-care-expenditures-an-inquiry-into-the- 5. Abu Dhabi’s Economic Performance in the nature-and-causes-of-the-cost-disease_1.pdf, Last 10 Years: Chart book (2001-2010), Abu accessed 30/7/2013. Dhabi Council for Economic Development 66 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

(2012), http://www.adced.ae/uploads/Chart- 14. Dong, Y, “How Health Insurance Affects book.pdf, accessed 30/7/2013. Health Care Demand. A Structural Analysis of 6. Sekhri, N, Savedoff, W, Thripathi, S, “Regu- Behavioral Moral Hazard and Adverse Selec- lation Private health insurance to Serve the tion,” Economic Inquiry, 2011, 51(2):1324– Public Interest Policy Issues for Develop- 1344. ing Countries,” World Health Organization, 15. Vickers, J, Yarrow, G, “Privatisation: An Eco- 2005, http://www.who.int/health_ﬁ nancing/ nomic Analysis,” Massachusetts: MIT Press, regulation_private_health_insurance_ 1988. dp_05_3.pdf, accessed 30/7/2013. 16. Polborn, M K, Hoy, M, Sadanand, A, “Advan- 7. Jost, T, “Private or Public Approaches to Insur- tageous Effects of Regulatory Adverse Selec- ing the Uninsured: Lessons from International tion in the Life Insurance Market,” Economic Experience with Private Insurance,” New York Journal, 2006, 116 (508):327–354. University Law Review, 2001, 76(2):419–492. 17 Health Statistics (2011), Health Authority 8. Sekhri, et al., supra, n.6. Abu Dhabi (HAAD), http://www.haad.ae/ 9. Id. HAAD/LinkClick.aspx?ﬁ leticket=JY0sMXQXr 10. Id. OU%3d&tabid=349, accessed 30/7/2013. 11. Id. 18. Krugman, P, “The Return of Depression Eco- 12. Id. nomics and the Crisis of 2008,” New York: 13. Outreville, J F, “Theory and Practice of Insur- W. W. Norton & Company, 2009. ance,” Kluwer Academic Publishers: Boston, 19. HAAD, supra, n.17. 1998. 20. Sekhri, et al., supra, n.6. The Impact of Star Physicians on Diffusion of a Medical Technology: The Case of Laparoscopic Gastric Bypass Surgery

Laura Shinn

Using data on all bariatric surgeries performed in the state of Pennsylvania from 1995 through 2007, this paper uses logistic and OLS regressions to measure the effect of star physicians and star hospitals on the diffusion of an innovation in bariatric surgery called laparoscopic gastric bypass surgery (LGBS). This paper tests for effects at both the hospital and physician level. Compared to hospitals with no star physicians (11 percent adoption rate), those with star physicians on staff show a much higher adoption rate (89 percent). Compared to hospitals that are not classifi ed as star hospitals (13 percent diffusion rate), hospitals with star status show a much higher diffusion rate (87 percent from fi rst quarter 2000 to fourth quarter 2001); being a star hospital raises the likelihood of that hospital diffusing LGBS from 13 percent to 87 percent. At the physician level, the empirical results indicate that star physicians exert positive asymmetric infl uence on the adoption and utilization rates of nonstars at the same hospital. Stars are those who: (1) graduated from a Top 30 medical school, (2) completed residency at a Top 30 hospital, or (3) are included in a Castle Connolly Top Doctors® publication. The results of this paper support earlier work on the role of key individuals in technology diffusion. It extends research on medical technology diffusion by testing a new data set for a chronic disease treatment. JEL classifi cations: D2, I10, I11, L2, O33. D2 production and organizations; L2 fi rm objectives, organization and behavior; I10 health general; I11 Analysis of health care markets; O33 technological change: choices and consequences; diffusion processes. Key words: production and organizations; fi rm objectives, organization and behavior; health general; analysis of health care markets; technological change; diffusion processes; star physicians

Introduction viduals have been recognized as opinion leaders1 who bear infl uential social char- Economists and sociologists have acteristics, such as technological status,2 debated the role of individual, institutional, which make them change agents3 within and market factors in the diffusion of tech- their peer groups, hospitals, and hospital nology. In medical technology, key indi- markets.

diffusion of a medical technology, and incentives in Dr. Shinn received her BS in Economics (concen- suicide intervention. tration in fi nance and accounting) from the Whar- Upon graduating from UPenn, Dr. Shinn worked as ton School of the University of Pennsylvania. She a fi nancial analyst for an independent power devel- received her MA and PhD in economics from Temple oper, overseeing analysis of more than $500 million University (2011). in municipal bond and private equity investments in Her research interests include applied microeconom- alternative energy projects. She became a manage- ics, health economics, and economic education. She ment shareholder and pursued various business pro- completed her postdoctoral research fellowship at jects before returning to graduate school. the Center for Health Equity Research & Promotion. Her current research includes end-of-life planning for J Health Care Finance 2014; 40(3):67–85 end stage renal care patients, racial disparities in the Copyright © 2014 CCH Incorporated

67 68 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Recently, Burke, Fournier and Prasad includes peer and patient input, a Castle Con- (hereafter BFP),4 have found diffusion of a nolly Top Doctors® publication.9 A limited medical innovation is positively related to qualitative analysis of one market, conducted the presence of physicians they call stars. by interviewing physicians and administrators They used 148,174 observations of angio- (representing about half of one market) lends plasty patients to trace the diffusion of coro- some qualitative support to the star defi nition. nary stents through all nonfederal Florida This study is conducted in two parts. In the hospitals over the period 1995–2001, and fi rst part, hospital level diffusion is studied. found that star physicians infl uence nonstar Controlling for hospital and market effects, physicians through social interactions within a hospital with a star physician has a much local peer groups. Nonstars either imitate higher likelihood of offering laparoscopic gas- star physicians or learn directly from them. tric bypass surgery (89 percent) than a hospital Huesch5 questions BFP’s conceptual that does not have a star physician (11 percent). model. He cites Van den Bulte and Lilien’s6 In the second part of the study, physician level analysis of Coleman, Katz and Menzel’s7 diffusion is studied. Nonstars increase adop- seminal work in medical technology diffu- tion and use of laparoscopic gastric bypass sion. Van den Bulte and Lilien found no signif- surgery when they operate at the same hospital icance to physician peer effects on diffusion as stars. Nonstars increase their utilization by of a drug, after they controlled for the effects 4.9 percent for every co-located star. of marketers. Huesch questions (1) BFP’s defi nition of star physician, (2) whether BFP Medical Background have suffi ciently controlled for hospital and Bariatric, or weight loss, surgery can be hospital market factors; (3) BFP’s focus on characterized as malabsorptive, restrictive, or the operating physician, as opposed to the a combination of both. Malabsorptive surger- attending physician, and (4) inclusion of ies reduce the body’s absorption of nutrients. nonadopting stars in the model. BFP address Restrictive surgeries reduce the body’s ability these questions and support their fi ndings to take in food. The fi rst surgeries, performed with additional tests on their data.8 in the 1950s were malabsorptive. The process This paper answers the call of both Huesch changed very little until the 1980s. In 1991 and BFP for further testing of the asymmet- the National Institutes of Health endorsed ric infl uence of star physicians on a medical bariatric surgery as an effective treatment technology diffusion. This paper presents of morbid obesity.10 Since the late 1990s, data and analysis that provides support for the most prevalent type of bariatric surgery BFP’s fi nding of asymmetric infl uence of performed is Roux-en-Y (RYGB) gastric star physicians in a different setting, the dif- bypass surgery.11 Gastric bypass surgery is fusion of laparoscopic gastric bypass surgery a combination of malabsorptive and restric- among physicians in 166 nonfederal hospitals tive surgery. It chemically and mechanically in Pennsylvania. BFP’s defi nition of star is alters the digestive system by reducing the expanded to include not only residency at a size of the stomach and redirecting a piece of top hospital, but also those physicians who the digestive tract so that absorption of cer- graduated from top medical schools or are tain high-fat foods is signifi cantly reduced. designated Top Doctors in a publication that Substitutes for surgery include dieting and The Impact of Star Physicians on Diffusion of a Medical Technology 69

drug treatments, both of which have low up to eight Diagnostically Related Group long-term rates of success for weight loss. (DRG) procedure codes. Procedure data Surgical treatments for obesity are clinically for each surgery are classifi ed according to proven to be the most effective long-term the International Classifi cation of Diseases, method for resolving morbid obesity.12 9th Revision, Clinical Modifi cation (ICD- Laparoscopic Roux-en-Y (LGBS) gastric 9-CM)16 procedure codes: 44.31, the code bypass is an innovation in surgical technol- for open gastric bypass surgery, and 44.38, ogy that provides an alternative to open Roux- the code for laparoscopic gastric bypass. en-Y gastric bypass, which was the most widely performed bariatric procedure until Bariatric Surgery in the US and in 2004. Nationwide, bariatric surgery shifted Pennsylvania to predominantly laparoscopic technique Annual rates of increase in the number of between 2004 and 2006.13 Laparoscopic, surgeries in Pennsylvania follow a pattern or minimally invasive, surgery allows sur- similar to that observed at the national level. geons to operate using laparoscopes, instru- For the US, there are dramatic increases in ments inserted through small openings in the number of surgeries from 1996 through the abdominal cavity, while observing their 2004, followed by slower growth thereaf- actions on video screens. For abdominal sur- ter. In Pennsylvania, the number of surger- gery, fi ve or six small, 0.5 cm to 1.0 cm inci- ies increased more than ten-fold from 1996 sions are made instead of one 10-cm-long through 2004. In Pennsylvania, the high- incision typical of traditional open surgery. est number of surgeries occurred in 2004, Analysis of bariatric surgeries at academic perhaps refl ecting a policy change by the medical centers show the shift from open to Department of Health and Human Services laparoscopic bariatric surgery is associated in July of that year that rescinded an ear- with lower post-surgery morbidity rates.14 lier Medicare policy statement and formally Minimally invasive surgery is also attractive recognized obesity as a disease for the fi rst because, on average, it decreases hospital time. Later in the year, Medicare cover- stays by three days, decreases patient recov- age of bariatric surgery began. In 2005 and ery time, and decreases the risk of hospital 2006, a slight decrease in bariatric surger- borne infections.15 ies may be suggestive of the entry into the market for outpatient bariatric surgery. It Data and Empirical Specifi cation may also be due to a 2006 Medicare rul- Patient Data. Proprietary data from the ing that bariatric surgery reimbursed under Pennsylvania Health Care Cost Containment Medicare must be performed at a CMS- Council (PHC4) include 39,918 patient- designated Center of Excellence. Moreo- level observations from each nonfederal, ver, the nationwide recession may have short-term, acute-care hospital in the state decreased the demand for surgery if patients of Pennsylvania from fourth quarter 1995 found copayments and the lost time from through second quarter 2007. Patient infor- work to be obstacles to surgery. In 2007, the mation includes patient diagnoses, hospital level of surgeries rose again. Table 1 shows length of stay, age, race, gender, county, and the number of bariatric surgeries in the US ZIP code of residence. Diagnoses include and Pennsylvania. 70 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 1. US and Pennsylvania of open and laparoscopic gastric bypass sur- Bariatric Surgeries geries in Pennsylvania from fourth quarter 1995 to second quarter 2007. Year US Pennsylvania Hospital Data. For years 1995–2007, the Financial Analysis, 1996 – 70 PHC4’s public report, Volume 1, General Acute Care Hospitals,17 198 1997 12,203* is used to track individual and aggregate 1998 – 362 hospital characteristics. The reports include 1999 – 674 hospital size (measured in beds), hospital clo- 2000 – 1315 sures, openings, mergers and name changes. Outpatient data are excluded. Each hospital’s 2001 – 2684 Council of Teaching Hospital (COTH) status 2002 71,733 4128 is obtained from the American Association 2003 101,144 6215 of Medical Colleges’ Council of Teaching 18 2004 121,055 7119 Hospitals and Health Systems. 2005 140,640 6724 To control for hospital market-level characteristics, a market for each hospital is 5746 2006 127,335 defi ned using a variable radius technique. 2007 186,000 6471 The 19-mile market radius used in this study 2008 220,000 – is consistent with the work of Gresenz, 19 Note: Data are from the Pennsylvania Health Care Rogowski, and Escarce and Phibbs and Cost Containment Council and The Annual Health Robinson.20 Hospital characteristics that Grades Bariatric Trends in American Hospitals are statistically signifi cant (p<=.05) include Study (2006, 2007, 2008, 2009 and 2010).31 *The US total from 1990-1997 is 12,203. Annual the following: population density, num- US totals are unavailable for 1998, 1999, 2000, ber of hospitals within 15 miles, average and 2001. patient length of stay, log of total charges, and hospital teaching status. Table 2 shows a comparison of the variables used in this Over the time period studied, gastric study and the variables used by Phibbs and bypass surgery accounted for more than Robinson21 and Gresenz, Rogowski, and 95 perccent of the bariatric surgeries in Escarce.22 Ordinary Least Squares Regres- Pennsylvania. Emergency surgeries and sions are used to calculate the 90 percent repair surgeries are excluded from the data. patient fl ows. The radii for bariatric surgery The number of LGBS more than doubled markets in Pennsylvania compare favora- between 2000 and 2001, which follows the bly with the previous work. Nineteen and pattern of expansion of open gastric bypass one-fi fth miles is the mean patient distance surgery. Between 2004 and 2005, the num- to hospital, which captures 90 percent of ber of LGBS nearly tripled, while the num- patients. The regression is then used to pre- ber of open gastric bypass surgeries declined dict market radii. The radius predicted is by more than half. Surgeries changed over 18.7 miles. This study uses 19.0 miles to from predominantly open to laparoscopic by calculate the radius of a hospital’s market the end of 2005. Figure 1 shows the number for each hospital in the sample. The Impact of Star Physicians on Diffusion of a Medical Technology 71

Figure 1. Open and Laparoscopic Gastric Bypass Surgery in Pennsylvania

Open and Laparoscopic Gastric Bypass Surgery in Pennsylvania 8000

7000

6000

5000

4000

3000

2000

1000

0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Laparoscopic 0 0 1 4 34 114 269 484 820 1377 4309 4071 2242 Open 14 70 197 358 653 1190 2385 3867 6143 6177 2453 1399 538 Note. 2007 contains only one-half year of data. Data are from the Pennsylvania Health Care Cost Containment Council.

Table 2. Comparison of Market Radii with Earlier Studies: 90 Percent of Patient Discharges

Standard Mean Median deviation Minimum Maximum Actual radius (miles) Phibbs & Robinson 17.8 14.0 12.7 0.6 124.4 Gresenz, Rogowski, & Escarce 21.5 15.7 19.7 0.4 179.0 Shinn 19.2 13.4 18.5 0.1 97.8 Predicted radius (miles) Phibbs & Robinson 17.8 17.2 7.9 4.7 42.6 Gresenz, Rogowski, & Escarce 22.8 21.0 12.2 0.2 105.1 Shinn 18.7 17.6 8.1 0.0 45.6

Note: Phibbs and Robinson,32 1983, California, 355 hospitals. Gresenz, Rogowski, and Escarce, 1997 data,33 9 states, 1,246 hospitals. Shinn,34 1995 through 2007, Pennsylvania, 166 hospitals. Gresenz, Rogowski, and Escarce calculate slightly large radii for rural areas. Urban area data are shown above. 72 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Physician Data. The PHC4 proprietary scopic bariatric surgery technology? Do star data include the Pennsylvania medical license hospitals diffuse laparoscopic gastric bypass number of the operating physician, attending surgery more quickly than nonstar hospitals? physician, and referring physician for each Let = the log odds of hospital i diffusing entry. Each license number is matched with laparoscopic surgery. the physician’s name and address from the Let Xi = a vector of physician, hospital,

Pennsylvania Department of State database. and market variables. Xi = { X1, X2, …., X9 }. There are 297 operating physicians. Resi- Then, the probability that a given hospital dency, special training, medical school grad- has diffused laparoscopic gastric bypass at uation date, and school are matched from time t is given by the publicly available American Medical P Association physician database. Data from log i = Z = α + β X + e [Equation 1] ® i i i America’s Top Doctors publications were 1–Pi graciously provided by offi cers from Castle Table 4 shows Model 1 variables and Connolly Medical Limited.23 Castle Con- descriptions. Table 5 shows variable means nolly produce annual publications that rank and standard deviations. Table 6 presents the top 1 percent of physicians in the US estimation results from Model 1 logistic according to specialty.24 Physician informa- regressions. tion is also obtained from individual hospital Web sites and from phone calls and emails to individual physicians. Table 3 shows the Discussion of Model 1 Results number of physicians adopting laparoscopic Compared with a hospital that has no star gastric bypass surgery and their utilization physician (11 percent probability), a hospital rates, conditional on adoption. that has a star physician has an 89 percent probability of diffusing laparoscopic gastric Empirical Specifi cation bypass. The result seems reasonable if the Two models are presented. Model 1 pre- variable doc_star is capturing the qualities sents diffusion at the hospital level, control- of infl uential individuals. From the inter- ling for facility and market characteristics. views, it appears that the criteria used in this The presence of a star physician at a hospital study are a reasonable proxy for star power or the status of a hospital as a star hospital (See Appendix 2, Description of Interviews). are shown as positively related to diffusion of A shortcoming of the data is that the data laparoscopic gastric bypass surgery. Model include only physicians who ultimately per- 2 presents diffusion at the physician level. form bariatric surgery over the time period. Model 2 considers behavior of nonstar and Some physicians who perform general sur- star physicians and considers the diffusion gery learn and specialize in bariatric surgery process in two parts: adoption and utilization. during the study time period. Many general surgeons with the characteristics choose not MODEL 1 to learn or specialize in bariatric surgery. Some physicians, especially by the end of Does the presence of star physician(s) the study time period, are specifi cally trained infl uence a hospital’s diffusion rate of laparo- in bariatric surgery and make the decision The Impact of Star Physicians on Diffusion of a Medical Technology 73

Table 3. Number of Physicians and Laparoscopic Surgeries by Quarter

Cumulative Leave Number Lap Utilization Qtr Physicians Adopters (%) Sample Enter Sample Surgeries Rate 1999:1 12 2 0 11 3 2.70 1999:2 19 4 0 9 6 3.90 1999:3 26 4 0 5 6 3.70 1999:4 27 5 1 3 15 8.02 2000:1 29 7 1 3 21 8.61 2000:2 37 8 1 9 18 6.19 2000:3 37 10 1 2 24 7.79 2000:4 44 12 1 5 47 11.81 2001:1 50 16 0 6 54 11.09 2001:2 58 19 2 8 44 6.88 2001:3 66 21 1 5 75 11.13 2001:4 70 23 2 3 87 12.17 2002:1 70 25 2 1 101 12.29 2002:2 70 29 3 4 104 10.26 2002:3 76 28 3 4 146 12.37 2002:4 79 32 1 7 142 11.41 2003:1 80 37 3 4 200 14.30 2003:2 87 41 1 7 207 11.95 2003:3 94 48 1 8 227 12.30 2003:4 98 52 2 3 243 12.43 2004:1 99 54 4 4 211 9.94 2004:2 98 54 6 4 185 10.16 2004:3 97 52 4 2 309 15.62 2004:4 85 69 0 3 756 45.35 2005:1 93 72 5 2 1002 59.64 2005:2 93 69 5 1 1060 61.95 2005:3 87 74 5 2 1056 64.35 2005:4 85 74 4 1 1017 67.08 2006:1 84 75 4 2 910 69.89 2006:2 80 79 3 0 921 72.01 2006:3 82 76 3 2 865 66.74 2006:4 80 74 3 2 864 73.03 2007:1 78 75 1 1 945 75.60 2007:2 78 76 0 0 912 76.51

Note: Utilization rate is calculated as the ratio of the number of laparoscopic gastric bypass surgeries to the total number of laparoscopic and open gastric bypass surgeries. 74 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 4. Model 1 Variables and Descriptions

xn Variable Name Description Diffused laparoscopic surgery. dif_lap = 0 if hospital has not diffused laparoscopic gastric bypass; dif_lap dif_lap = 1 if hospital has diffused laparoscopic gastric bypass. Diffusion is assumed if a hospital performs average of 6 surgeries per quarter for four consecutive quarters (Tian 2006). Physician star characteristic.

X1 doc_star doc_star = 0 if physician is not star; doc_star = 1 if physician is a star. Hospital star characteristic.

X2 coth coth = 0 if hospital is not COTH member; coth = 1 if hospital is a COTH member. Hospital size variable.

X3 beds_201 beds_201 = 0 if hospital has <= 200 beds; beds_201 = 1 if hospital has > 200 beds. Diffused open surgery.

X4 dif_op dif_op = 0 if hospital has not diffused open gastric bypass; dif_op = 1 if hospital has diffused open gastric bypass. Market competition measure.

X5 lag_ sat Continuous variable of each hospital’s market saturation rate, lagged one quarter. Market competition measure. hhi_type = 1 if HHI < 1500, market is competitive; X hhi_type 6 hhi_type = 2 if 1500 < HHI ≤ 2500, market is moderately concentrated; hhi_type = 3 if HHI > 2500, market is highly concentrated. Hospital profi t status

X7 for_profi t for_profi t = 0 if hospital is not for profi t institution; for_profi t = 1 if hospital is a for profi t institution. Market competition measure.

X8 Beds_per_1000 Continuous variable equals number of hospital beds per thousand population in the hospital’s market

X9 d_op_star d_op_star, interaction term = doc_star * dif_op Note: N = 5984. Mean HHI is 3,753, standard deviation 2556, minimum 462 and maximum 10,000. Mean number of beds is 219, standard deviation 176, minimum 25, and maximum 814. to adopt bariatric surgery before they join The empirical analysis supports only a a hospital. In this study, general surgeons qualifi ed statistical relationship between who do not adopt bariatric surgery are not star hospitals and diffusion of laparoscopic included. A study that includes all physi- surgery. This statistical outcome appears cians who perform general or abdominal counterfactual, since COTH hospitals have surgery during the time period, but who do been shown to account for a disproportion- not adopt bariatric surgery would likely pro- ate share of technology-intensive hospital duce less striking results. services. It may be that the data include a The Impact of Star Physicians on Diffusion of a Medical Technology 75

Table 5. Model 1 Means and Standard Deviations

xn Variable Mean Minimum Maximum 0.102 dif_lap 0 1.00 (.303) 0.059 X doc_star 0 1.00 1 (.235) 0.187 X coth 0 1.00 2 (.390) 0.751 X beds_201 0 1.00 3 (.432) 0.211 X dif_op 0 1.00 4 (.408) 8.934 X lag_sat 0100 5 (16.426) 2.48 X hhi_type 1 3.00 6 (.778) 0.054 X for_profi t 0 1.00 7 (.227) 3.578 X beds_per_1000 2.380 4.46 8 (.730) 3.65e-11 X d_op_star −.917 .083 9 (.067)

Note: N = 5984. Mean HHI is 3,753, standard deviation 2556, minimum 462 and maximum 10,000. Mean number of beds is 219, standard deviation 176, minimum 25, and maximum 814. long enough time frame, that by the end of scopic gastric bypass surgery, the strongest the time period, laparoscopic gastric bypass statistical impact from teaching hospital is no longer considered a high technology status occurs in the earliest phases of use or technology-intensive service. Then, other of a new technology. That is not surpris- factors, such as whether a hospital has star ing. Presumably, teaching hospitals would physicians, might become more important be expected to be leaders in most new pro- factors in hospital’s decision to adopt and cedures. Model 1 controls for hospital size diffuse a technology. The result is statisti- and hospital profi t status, as well as charac- cally signifi cant if a shorter time frame is teristics of the local market, such as market considered. If the statistical analysis is lim- concentration (measured by HHI), market ited to the years 2000 through 2002, then saturation rate, and a measure of market being designated a Council of Teaching capacity. Large, for-profi t hospitals are Hospital (COTH) increases the likelihood more likely to diffusion laparoscopic gastric of a hospital diffusing laparoscopic gastric bypass surgery. A limitation of Model 1 is bypass surgery from 13 percent to 87 per- that the number of stars and the interaction cent. It appears that, in the case of laparo- among peers is not measured in the model. 76 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 6. Model 1: Estimation Results of Logistic Regression for Hospital Level Diffusion

Coefficient

xn Variable Name (Robust Std Err) z P>|z| Odds Ratio 2.051 x doc_star 1 (.488) 4.20 0.000* 7.776 0.753 x coth 2 (.642) 1.17 0.241 2.120 3.225 x beds_201 3 (.792) 4.07 0.000* 25.156 3.351 x dif_op 4 (.941) 3.56 0.000* 28.545 0.123 x lag_sat 5 (.015) 8.23 0.000* 1.131 -0.256 x hhi_ type 6 (.305) −.84 0.402 0.774 1.693 x for_profi t 7 (.599) 2.82 0.005* 5.433 -0.620 x beds_per1000 8 (.343) −1.81 0.070** 0.537 -1.510 x d_op_star 9 (.913) −1.65 0.098** 0.221

Note: N = 5984. McFadden’s R2 = .724. Wald chi(9) =103.13. Hosmer-Lemeshow chi2(7) = 8.34. Prob > chi2 = .3039. *p =.05, **p=.10

Model 2 addresses physician-level adoption equations are used to test the interactions: (1) and utilization and accounts for the interac- on adoption and (2) on utilization. tion among stars and nonstars. We seek an equation describing adoption. Let = the log odds of physician i adopting MODEL 2 laparoscopic surgery at time t. Let = physician’s lagged number of cumu- Empirical Model 2 brings the analysis to lative bariatric procedures. the level of the individual physician behav- Let,, and = dummy variables for each ior, concentrating on the signifi cance of the physician, calendar quarter, and hospital, interactions between star and nonstar physi- respectively. cians at co-located hospitals. Model 2 fol- Let x1,rt = mean patient age lows the methodology of Burke, Fournier and Let x2,rt = patient gender 25 Prasad by adding interaction terms to cap- Let St = number of star peers at time t. ture any impact on diffusion attributable to Let NSt = number of non-star peers at time t. the infl uence from star to nonstar physicians Let g1 = member of group including star and from nonstar to star physicians. Separate physicians. The Impact of Star Physicians on Diffusion of a Medical Technology 77

Let g2 = member of group including non- the utilization equation estimate for each star physicians physician. Equation 2 estimates the log odds of the u = p + c + α + β x + β x + γg2 probability of all physicians adopting lapa- i,t i t h 1 1,rt 2 2,rt * S + μg1* NS + δV + e roscopic gastric bypass at time t. rt rt r,t–1 i,t

α β β γ [Equation 3] zi = pi + ct + h + 1x1,rt + 2x2,rt + g2 * S + μg1* NS + δV + e rt rt r,t–1 i,t Table 8 shows the means, standard deviations, and expected signs for the terms in [Equation 2] Model 2: Utilization. Tables 9 and 10 present The interaction terms capture peer effects: the results of the regressions for Model 2. the term captures the effect of nonstars on stars; the term captures the effect of stars on Discussion of Model 2 Results nonstars. Table 7 shows the means, standard For the eight calendar quarters 2000:1 to deviations, and expected signs for the terms 2001:4, the logistic regression shows a posi- in the Model 2: Adoption. tive, asymmetric infl uence of star physicians We seek an equation describing utilization on nonstar physicians. The probability of rates among physicians. adoption increases from 4 percent for non- Let = the ratio of number of laparoscopic stars with no co-located stars to 96 percent gastric bypass surgeries performed by a for nonstar physicians co-located with star physician to the total number of gastric physicians. The utilization linear regression, bypass surgeries, open and laparoscopic, conditional on a physician having adopted performed by that same physician. Replac- laparoscopic surgery, shows that utilization ing with in Equation 2 yields Equation 3, rates for nonstar physicians increase by 4.88

Table 7. Model 2 Variable Means, Standard Deviations: Adoption

Variable Name Mean (Std Dev) Minimum Maximum Adoption model (n = 101 observations)

Vr,t-1 Lagged physician volume 7.137 (10.912) 0 46 Patient gender Mean patient age 42.010 (1.636) 35.857 45.444

Srt number of stars encountered at time t 1.543 (1.769) 0 5

NSrt number of non-stars encountered at time t 1.348 (1.133) 0 3 grp1 Physicians in grp1, (star) .380 (.488) 0 1 grp2

grp1nst grp1*NSrt .848 (1.266) 0 3 .533 (1.338) grp2st grp2*Srt 05 Note: N =101. 78 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 8. Model 2 Variable Means, Standard Deviations: Utilization

Variable Name Mean (Std Dev) Minimum Maximum Utilization model (n = 334 observations) Utilization rate .286 (.371) 0 1

Vr,t-1 Lagged physician volume 9.86 (13.64) 0 100 Physicians in grp1, (“star”) .26 (.44) 0 1 Mean patient age 43.14 (2.77) 20 61

Srt number of stars encountered .57 (1.09) 0 6

NSrt number of non-stars encountered 1.05 (1.25) 0 9 grp1nst grp1*nst .39 (1.00) 0 9 grp2st grp2*st .31 (.81) 0 6

Note: N = 334 observations.

Table 9. Model 2 Results of Logistic Regression: Adoption

Variable Coefficient z P>|z| Odds Ratio grp2st grp2*st 3.106 (1.752) 1.77 .076 22.323 grp1nst grp1*nst −3.277 (2.407) −1.36 .173 .038

Vr,t-1 Lagged number of open surgeries −.229 (.115) −1.99 .047 .795 Dummy variables

cˆt For each quarter α ˆ t For each hospital ρ ˆt For each physician Note: N = 101.

Table 10. Model 2 Results of Linear Regression: Utilization

Variable Coefficient t P>|t| grp2st grp2*st .0488 (.023) 2.14 .033 grp1nst grp1*nst .007 (.014) .48 .629 Lagged number of open surgeries .002 (.001) −2.27 .024

Dummy variables

cˆt For each quarter α ˆ t For each hospital ρ ˆt For each physician Note: N = 334. The Impact of Star Physicians on Diffusion of a Medical Technology 79

percent for every one co-located star physi- to nonstar physicians. This study supports cian above the average number of co-located Burke, Fournier, and Prasad’s fi ndings by stars. As expected, the coeffi cient refl ecting testing their theory on a set of observations the infl uence of nonstars on stars is not sta- for a different technology. Though this study tistically signifi cant. Stars (group 1) are not uses a smaller data set, it provides statistical statistically infl uenced by the presence of support of Burke, Fournier, and Prasad’s fi nd- nonstars. The control variables for physi- ing in cardiac surgery with data from another cian, hospital, calendar quarter, patient age state and for a different technology. and gender are not statistically signifi cant. There are several limitations to this study. This study does not control for time-varying Conclusions hospital effects; however, it controls for local This paper examines the role of star power market and hospital fi xed effects. This study in the diffusion of a medical technology. Cer- does not consider measures of cost, quality, tain physicians and hospitals are considered morbidity or mortality.27 Evidence from the to be stars, that is, highly attractive to patients. personal interviews suggested that there are In the case of physicians, stars are defi ned a minimum number of surgeries per physi- as those who have graduated from a Top 30 cian to achieve best outcomes. Several cited Medical School or completed a residency at the 100 surgery level,28 which is the mini- a Top 30 Hospital, or are included in Castle mum per year required by the American Col- Connolly’s Top Doctors® publications. In the lege of Surgeons (ACS) and the American case of hospitals, stars are defi ned as mem- Society for Metabolic & Bariatric Surgery bers of the American Association of Medi- (ASMBS) to achieve Center of Excellence cal College’s Council of Teaching Hospitals. (COE) designation. Both those agencies The suitability of the proxy measures is sub- track in-hospital morbidity, mortality and stantiated by interviews with physicians and 30-day readmission rates. administrators. The empirical results indicate This study suggests areas for policy that hospitals with star physicians and hospi- intervention. Lack of star physicians may tals that are stars are more likely to diffuse mean welfare losses occur. Efforts to draw bariatric surgery than hospitals without such star physicians to hospitals without stars characteristics. The empirical results indicate could improve welfare in two ways: (1) the that star physicians exert positive asymmet- star physicians can diffuse gastric bypass ric infl uence on the adoption and utilization surgery, and (2) nonstars may learn from rates of nonstars at the same hospital. or may be inspired by co-located stars and Two areas in which this study supports and increase diffusion rates. This could be wel- extends earlier research in technology diffu- fare enhancing because it could improve sion: (1) fi nding positive evidence for the role access to bariatric surgery, and in particular, of key individuals, “stars,” and (2) fi nding evi- to an apparently superior technology. dence for the role of market and institutional This study fi nds that diffusion of a medical factors in technology diffusion. Recent work technology is more likely when certain hos- by Burke, Fournier, and Prasad26 emphasize pital and market features are present: a large, the role of social learning by testing the role urban hospital with a star physician is more of star physicians in diffusion of a technology likely to diffuse technology than a smaller, 80 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

rural hospital or a hospital with no star physi- It may be that the tradeoff between quality cians. It may be that economies of scale are and distance traveled is preferred by patients present that make it more effi cient for larger and providers for this technology. Further hospitals to offer laparoscopic gastric bypass studies, especially of chronic disease treat- surgery; however, there may be questions of ments, might help defi ne whether differences access and equity. Like previously mentioned in observations have to do with patient pref- research, this study also raises the issue that erences or the nature of the disease treated. laggards in technology diffusion impose wel- Because chronic disease affects almost one fare costs on society. Twenty-seven percent out every two adults in the US, the tradeoffs of hospitals in this study are either catego- between patient quality of care and concen- rized as stars or have diffused gastric bypass trations of technology for particular treat- surgery by the end of the sample period. All ments become interesting areas for further else being equal, is it effi cient or equitable study. Further studies of technology for acute for only some Pennsylvania hospitals to have disease treatments might reveal a different diffused laparoscopic gastric bypass surgery set of preference tradeoffs and might indicate by the end of the sample period? Should all that concentrations of the technology in rela- hospitals have laparoscopic gastric bypass tively few centers is socially wasteful because surgery? If so, this suggests an area for policy it limits access to vulnerable populations.30 intervention. Efforts to draw star physicians to small, rural hospitals could improve wel- Appendix 1: Description of the Data fare in two ways: (1) the star physicians can Proprietary data from the Pennsylvania diffuse gastric bypass surgery and (2) non- Health Care Cost Containment Council stars may learn from or may be inspired by (PHC4) include 39,918 patient level obser- co-located stars and increase diffusion rates vations from each nonfederal, short-term, in rural areas. This could be welfare enhanc- acute-care hospital in the state of Pennsylva- ing because it could improve access to bari- nia from fourth quarter 1995 through second atric surgery; however, if economies of scale quarter 2007. Patient information includes are present and rural markets are very small, patient diagnoses, hospital length of stay, effi ciency losses must be carefully weighed age, race, gender, county, and ZIP code of against equity gains. Alternately, it may be residence. Diagnoses include up to eight helpful to send physicians from rural hos- Diagnostically Related Group (DRG) proce- pitals to learn from star physicians at larger, dure codes. Procedure data for each surgery urban hospitals. are classifi ed according to the International This study suggests some areas for further Classifi cation of Diseases, 9th Revision, investigation. Burke, Fournier, and Prasad29 Clinical Modifi cation (ICD-9-CM)35 proce- studied a treatment that is generally for an dure codes: 44.31, the code for open gastric acute condition. This study considered a bypass surgery, and 44.38, the code for lapa- treatment for a chronic condition, obesity. roscopic gastric bypass. Perhaps concentration of surgeries in rela- For years 1995–2007, the PHC4’s public tively few hospitals may improve skill levels, report, Financial Analysis, Volume 1, Gen- reduce mortality, and be an overall supe- eral Acute Care Hospitals,36 is used. The rior allocation of society’s scarce resources. reports include hospital size (measured in The Impact of Star Physicians on Diffusion of a Medical Technology 81

beds), hospital closures, openings, mergers, from individual hospital Web sites and and name changes. Each hospital’s Coun- from phone calls and emails to individual cil of Teaching Hospital (COTH) status is physicians. obtained from the American Association of I conducted a series of informal inter- Medical Colleges’ Council of Teaching Hos- views with a small sample of physicians pitals and Health Systems.37 and administrators. This group represented The PHC4 proprietary data include the 12 of 24 hospitals in the Philadelphia met- Pennsylvania medical license number of ropolitan market. The interviews provide a the operating physician, attending physi- limited qualitative analysis of the decision- cian, and referring physician for each entry. making process of hospitals and physicians Each license number is matched with the in adoption and diffusion of bariatric sur- physician’s name and address from the gery. The key fi nding regards the decision- Pennsylvania Department of State data- making process. Two discernible patterns base. Residency, special training, medi- emerge. In some cases, diffusion is initiated cal school graduation date, and school are by a physician. In other cases, diffusion is matched from the publicly available Amer- led by hospital administration. The fi nd- ican Medical Association physician data- ings may not be applicable to all hospitals, base. Data from America’s Top Doctors® but they suggest certain explanations for publications38 were graciously provided observed diffusion patterns. by offi cers from Castle Connolly Medical The interviews were approved by Temple Limited. Castle Connolly produce annual University IRB #13184, dated August 23, publications that rank the top 1 percent of 2010. The introductory letter and question- physicians in the US according to specialty. naire follows. The questionnaire was used as Physician information is also obtained a starting point for discussion.39 82 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Introduction Letter TEMPLE UNIVERSITY®

Dear Sir/Madam,

A research team at Temple University is conducting a study on how star physicians impact the diffusion of medical technology. We are requesting hospital administrators, board members and physicians to comment on the following survey questions in an interview. This survey has a total of 12 questions.

Your participation in this study is voluntary. You may choose not to participate and/or to with- draw from the study at any time. This survey is anonymous.

The overall results of the study may be published, but the research will be conducted with an assurance of confi dentiality for you and your organization. Neither your name nor your individual answers will be known.

Participation in this interview will be considered your consent to participate.

If you have any questions concerning this interview, please contact Laura Shinn at (xxx) xxx-xxx, or by email, [email protected].

______Primary Investigator’s Signature Date The Impact of Star Physicians on Diffusion of a Medical Technology 83

Questionnaire

Agree Disagree Strongly Neutral Strongly 1. A single physician or group of physicians origi- 1234 5 nates bariatric surgery at this hospital. 2. Administration or the board of directors originates 1234 5 bariatric surgery at this hospital. 3. One physician in the hospital is recognized 1234 5 by the other doctors as a Top Doc, or Star, or opinion leader in bariatric surgery. 4. A star brings status to other surgeons in the 1234 5 hospital. 5. Stars may generate a spill over effect to increase 1234 5 demand for the entire spectrum of services offered by the hospital. 6. Bariatric surgery is profi table for the hospital. 1 2 3 4 5 7. Bariatric surgery is highly profi table for the 1234 5 physician, that is, compared with other abdominal surgeries. 8. Laparoscopic surgery is a substitute for 1234 5 open surgery. 9. Patients prefer laparoscopic surgery over 1234 5 open surgery. 10. Laparoscopic restrictive devices are substitutes 1234 5 for gastric bypass surgery. Other questions: Which hospitals do you see as your main competitors in the market for bariatric surger y? Is my assumption of at least two surgeries per month for four consecutive quarters, that is, 24 surgeries per year, is indicative of a hospital having diffused bariatric surgery. What do you estimate as a minimum number of bariatric surgeries a hospital would need to do to achieve efficiency? 84 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

REFERENCES

1. Rogers, E, Diffusion of Innovations (5th ed.), 12. Sjöström, L, Narbro, C, Sjöström, CD, The Free Press, New York, 2003; Coleman, J, Karason, K, Larsson, B, Wedel, H, et al., Katz, E, and Menzel, H, “The Diffusion of an “Effects Of Bariatric Surgery On Mortality In Innovation Among Physicians,” Sociometry, Swedish Obese Subjects,” New England Jour- December 20:253–270 (1957). nal of Medicine, 357(8):741–52 (2007). 2. Huckman, RS, “The Utilization Of Competing 13. Health Grades Inc., The Third Annual Health Technologies Within The Firm: Evidence From Grades Bariatric Surgery Trends in American Cardiac Procedures,” Management Science, Hospitals Study, Golden, Colorado: Health 49(5):599–617 (2003). Grades, Inc. (2008). 3. Berwick, DM, “Disseminating Innovations In 14. Jones, K, Afram, B, Benotti, P, Capella, R, Health Care,” Journal of the American Medi- et al., “Open versus Laparoscopic Roux-en-Y cal Association, 289:1959–75 (2003). Gastric Bypass: A Comparative Study of Over 4. Burke, M, Fournier, G, and Prasad, K, “The 25,000 Open Cases and the Major Laparo- Diffusion Of A Medical Innovation:Is Success scopic Bariatric Reported Series,” Obesity In The Stars?” Southern Economic Journal, Surgery, 16: 721–727 (2006). 73(3):588–603 (2007); Burke, M, Fournier, G, 15. Schauer, P,, Ikramuddin, S, Gourash, W, Ram- and Prasad, K, “The Diffusion Of A Medical anathan, R, Luketich, J, “Outcomes After Innovation: Is Success In The Stars? Further Laparoscopic Rous-en-Y Gastric Bypass for Evidence,” Southern Economic Journal, 75(4): Morbid Obesity,” Annals of Surgery, 232: 124–78 (2009). 515–529 (2000). 5. Huesch, MD, Comment On “The Diffusion 16. National Center for Health Statistics, “Interna- Of A Medical Innovation: Is Success In The tional Classifi cation of Diseases, Ninth Revi- Stars?” Southern Economic Journal, 75(4): sion, Clinical Modifi cation,” retrieved from 1270–1273 (2009). http://www.cdc.gov/nchs/icd/icd9cm.htm 6. VanDenBulte, C, Lilien, G, “Contagion versus (2012). Marketing Effort,” American Journal of Sociol- 17. Pennsylvania Health Care Cost Containment ogy, 106(5):1409–1435 (2001). Council, PHC4 Reports County Profi les, 7. Coleman, supra, n.1. retrieved April 9, 2010, from PHC4: http:// 8. Burke, et al., (2009) supra, n.4 www.phc4.org/countyprofi les/ (2010). 9. Castle Connolly maintains a database with 18. AAMC, Council of Teaching Hospitals and regional and national Top Doctors. Surgeons Health Systems (COTH), Washington, DC: may be named Top Doctors in national or American Association of Medical Colleges, regional print publications and/or online retrieved from https://www.aamc.org/mem- directories. Connolly, J, Morgan, J, America’s bers/coth/ (2011). Top Doctors, 12th ed., Castle Connolly, Ltd, 19. Gresenz, C, Rogowski, J, and Escarce, J, New York, 2012. Top Doctors: Arizona 2010 . “Updated Variable-Radius Measures of Hos- 10. National Institutes of Health, 1991, Gastroin- pital Competition,” Health Services Research, testinal surgery for severe obesity, NIH Con- 39(2): 417–430 (2004). sensus Statement Online, 9(1):1–20. 20. Phibbs, CS, and Robinson, J, “A Variable- 11. Health Grades Inc., The Second Annual Radius Measure of Local Hospital Market Health Grades Bariatric Surgery Trends in Structure,” Health Services Research, 28(3): American Hospitals Study, Golden, Colorado: 313–324 (1993). Health Grades Inc. (2007); Thomson Reuters, 21. Phibbs, supra, n.20. “Healthcare Trend Tracker: Growth In Bariatric 22. Gresenz, supra, n.19. Surgery Focuses On Changes In Technique,” 23. Castle, supra, n.9. retrieved from http://home.thomsonhealth- 24. Castle, supra, n.9. care.com/News/view/?id=1241 (2007). 25. Burke, supra, n.4. The Impact of Star Physicians on Diffusion of a Medical Technology 85

26. Burke, supra, n.4. Hospitals Study, Golden, Colorado: Health 27. See Picot, J, Jones, J, Colquitt, JL, Gospodarevs- Grades Inc. (2006); Health Grades Inc., kaya, E, et al., “The Clinical Effectiveness And supra, n.11; Health Grades Inc., supra, n.13; Cost-Effectiveness Of Bariatric (Weight Loss) Health Grades Inc., The Fourth Annual Health Surgery For Obesity: A Systematic Review Grades Bariatric Surgery Trends in American And Economic Evaluation,” Health Technol- Hospitals Study, Golden, Colorado: Health ogy Assessment, 13(41):1–190 (2009). Grades Inc. (2009); Health Grades Inc., The 28. Berwick, DM, “Disseminating Innovations Fifth Annual Health Grades Bariatric Surgery In Health Care.” Journal of the American Trends in American Hospitals Study, Golden, Medical Association, 289 (April 16), 1959–75 Colorado: Health Grades Inc. (2010). (2003); Picot, J, supra, n.27. 32. Phibbs, supra, n.20. 29. Burke, supra, n.4. 33. Gresenz, supra, n.19. 30. Martin, M, Beekley, A, Kjorstad, R, and 34. Shinn, L, Ph.D. dissertation, The Impact of Sebesta, J, “Socioeconomic disparities in Star Physicians on Diffusion of a Medical eligibility and access to bariatric surgery: a Technology: the Case of Laparoscopic Gastric national population-based analysis,” Surgery Bypass Surgery, Temple University, United for Obesity and Related Diseases, 6(1): 8–15 States–Pennsylvania (2011). (2010); Wallace, AE, Young-Xu, Y, Hartley, D, 35. National Center for Health Statistics, supra, and Weeks, WB, “Racial, Socioeconomic, n.16. and Rural-Urban Disparities in Obesity- 36. Pennsylvania Health Care Cost Containment Related Bariatric Surgery,” Obesity Surgery, Council, supra, n.17. 20(10):1354–60 (2010). 37. AAMC, supra, n.18. 31. Health Grades Inc., The First Annual Health 38. Castle, supra, n.9. Grades Bariatric Surgery Trends in American 39. Shinn, L, supra, n.34. Health Policies and Intervention Strategies

A Description of Current Issues and Approaches to Care of the Public Health and Health Care System in the United Arab Emirates

Aliye T. Mosaad and Mustafa Z. Younis

The demographic factors of the United Arab Emirates (UAE) have changed drastically within one generation. This is evident in how quickly it has moved from a developing nation, where ﬁ shing was once the main source of income, to a country that is quite developed, competing on a global level. From one perspective, socio-economic progress has brought many beneﬁ ts to the population. These include improved education, better access to health care, and safe drinking water. However, on the other hand, economic development has been the cause for changes in lifestyles, eating habits, and traditional societal and family structures.1 Over time, these changes have added up, creating an unprecedented impact on the population’s health. This impact has crept up onto the society until suddenly a notable epidemic has become recognized in the country. According to the UAE Ministry of Health, 19.5 percent of the UAE population has diabetes, making it the second highest rate in the world.2 The structure and responsibilities of the current UAE health care systems along with other cultural factors were investigated in order to determine their impact on the growing epidemic. Key words: United Arab Emirates, demographics, diabetes, health systems, cultural barriers, health awareness, food marketing, health initiatives

Introduction Aliye T. Mosaad has worked as a research assistant for the University of Virginia Psychology Department, Demographical Information and later at a neuro-spinal facility in Florida, as a neu- The United Arab Emirates (UAE), rologist assistant. Ms. Mosaad also served as the fi rst founded in 1971, is a country located in the woman hired by the Saudi Arabian government in the United Arab Emirates (UAE). Her post was at the Royal Middle East, bordering the Gulf of Oman Consulate General of Saudi Arabia in Dubai, UAE. and the Persian (Arabian) Gulf, between She also served as an attaché under the Saudi Arabian Oman and Saudi Arabia. It consists of Mission to the United Nations in New York City. She seven emirates or “states” that have united received her bachelor’s degree from the University of to form the UAE. These emirates include: Virginia, and Masters Degree in Public Health from Abu Dhabi, Dubai, Sharjah, Ajman, Umm the University of South Florida. Ms. Mosaad is currently taking biology-based courses in the Washington Al Quwain, Fujairah, and Ras Al-Khaimah. DC metropolitan area in preparation for applying to The two richest and most widely known medical schools in the United States. emirates are Abu Dhabi (the capital of the Mustafa Z. Younis is tenured professor at Jackson country) and Dubai. The United Arab Emir- State University. His areas of interest are Healthcare ates put itself on the map by discovering oil Accounting, Finance, Healthcare Finance & Health more than 30 years ago, opening the door for Economics. Dr. Younis served as a consultant and the development of an open economy and expert for universities and programs internationally. Additionally he held a visiting professor appointment allowing the country to compete on an inter- at Jena University School of Medicine, Jena, Germany. national scale. The country is now a central Acknowledgements: The authors would like to thank hub in the eastern hemisphere for business Dr. William B. Ward for his valuable comments and development, engineering feats, and a place feedback on this manuscript. where technological advances are fi rst being J Health Care Finance 2014; 40(3):86–100 introduced. The UAE is the most liberal of Copyright © 2014 CCH Incorporated

86 Health Policies and Intervention Strategies 87

the Gulf States, and the most modern Mid- the majority of the population will consist dle Eastern country.3 These factors have of the elderly. This shift—better known also given the country an economic boost as demographic aging—will bring about allowing it to play an important role in the higher levels of chronic diseases such as region. diabetes, if something is not done (see Fig- The UAE has the characteristics of a ure 1).4 developing country, including: high fertil- The country is one of a desert climate with ity rates, low literacy among the traditional extremely hot and humid summers. Regard- population, strong traditional cultural val- ing religious practices, 96 percent of the pop- AU: Delete ues, and many people living in multigenera- ulation follows the practice of Islam while “or” or is tional homes. It also has the characteristics the other 4 percent are Christian or. What something missing here? of a developed country in terms of a high makes the UAE unique is that less than 19 economic status, with most families hav- percent of the total population of the country ing servants, an urbanized population, and is actually originally “Emirati,” also known a high growth rate of the elderly (those 65 as the UAE nationals (see Figure 1). Most years of age and older). As the young cur- of the country is composed of expatriates rently dominate the population percentage- (see Table 1). Because the actual citizens of wise, implications are that in the near future the country are quite low in number, the gov-

Figure 1. UAE National versus. Immigrants (Expatriates) Population and Diabetes Prevalence

Share of Immigrants in the Comparative Diabetes Prevalence in the UAE in the Age UAE Population (2010) Group of 20–79 years

100% 90% Overall 18% 80% 70% Male 18% 60% 7.32 Female 18% 50% 40% Overall 21% 30% 20% Male 22% 10% 0.95 0% Female 20% UAE Population National Immigrants 0% 5% 10% 15% 20% 25%

Source: Middle East Health Magazine Jan 2013. Data collected from Booz Allen Hamilton analysis of data from Boston Analytics; International Monetary Fund World Outlook Database; International Diabetes Federation. Diabetes Atlas, 5th Ed. 2009; World Bank Atlas 2012; World Bank Databank 2009.5 88 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Table 1. Statistical comparison of the Emirati Population to the Expatriate Population.

Population by Nationality (National and Nonnational) and Sex (2010 mid-year estimates) Nationality Male Female Total National 479,109 468,888 947,997 Nonnational 5,682,711 1,633,362 7,316,073 Total 6,161,820 2,102,250 8,264,070

Source: United Arab Emirates National Bureau of Statistics. ernment is trying to increase their citizens’ ditional lifestyles.9 Previous research has population by providing incentives for citi- been done on some countries and indigenous zens to marry one another, such as monetary populations with the highest rates of diabe- gifts upon marriage.6 tes in the world. This epidemic has had an exponential growth rate within recent years. The Gulf Cooperation Council A recent study done in the United Arab Emirates indicates that high rates of indus- The Gulf Cooperation Council (GCC) trialization and modernization are correlated countries consist of Bahrain, Kuwait, Qatar, with a decrease in fertility rates and infec- Oman, Saudi Arabia, and the United Arab tious disease. On the fl ip side, an increase in Emirates. These countries are composed chronic disease and an increase in life expec- of populations that are of similar cultural tancy are also seen.10 This trend is part of backgrounds, religion, language, lifestyles, the epidemiologic transition and has started and income; thus they are comparable and to impact the health status and policy deci- all face analogous health issues. Many coun- sions of the UAE. tries in the Gulf region of the Middle East The GCC countries have undergone a are ranked in the top 10 countries with the substantial amount of modernization within highest prevalence of diabetes (see Figure 1). recent decades. The UAE infant mortality rate is ranked 138th in the world with 12.7 Modernization’s Impact on deaths per 1,000 live births, and the aver- Society and its Health age life expectancy of individuals is 76.11 years.11 Health care spending in the region An explosive epidemic of diabetes has is expected to increase from $12 billion begun to plague the developing countries of today, to approximately $60 billion by the the world that are undergoing modernization. year 2025. Diseases of affl uence and their Modernization in society can be defi ned as risk factors are to blame for such a spike the process of economic and socio-cultural in health care spending. Governments and change, by which traditional cultures indus- their health institutions are not currently trialize and develop a capitalist economy, equipped or prepared for the main risk fac- characterized by division of labor, reduced tors needing their attention, which include: importance of kinship, and changes in tra- population growth, an aging population, and Health Policies and Intervention Strategies 89

Figure 1. Estimated Prevalence of Diabetes Worldwide in 2007

ESTIMATED PREVALENCE OF DIABETES IN 2007

Diabetes in population % 14 – 20% 10 – 14% 8 – 10% 6 – 8% 4 – 6% Less than 4% Map shows type 1 and type 2 diabetes Obesity and type 2 diabetes are causally linked

AU:Endnote says 2007. Source: The Diabetes Atlas 2006.8 please confi rm. health risk factors. A change in the health tional values, close family ties, the univer- policy and regulations of the UAE Minis- sal practice of religion, and high economic tries of Health is essential in order to cope resources, all of which are attributed to with the growth of chronic ailments such the elderly maintaining a higher level of as diabetes, and to care for its growing and health into their later years.12 It is rare for aging population. the elderly to be seen living in institutional- ized care facilities unless they are severely Aging Population disabled. It was also found that the annual One study involved participants chosen UAE growth rate for those over the age of from the UAE community to answer survey 65 is around 10.3 percent, which is the high- questions based on their health status and est predicted in the world between the years diabetes. It was noted that in comparing of 1999 and 2025. This indicates the urgent the US and UAE, both countries had simi- need for planning the care of the elderly and lar rates of functional independence, but the the huge number of individuals with chronic US had higher rates of chronic problems in illnesses, such as diabetes, that will need the elderly, with diabetes as the exception. treatment in the near future.13 A point of discussion in these results was Life expectancy in the GCC rose from that the UAE has a higher regard for tradi- 60.5 years in 1978 to 73 years in 2004. The 90 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

population above the age of 60 years in the study done on childhood obesity and dia- GCC is expected to increase to 15 million betes in the Emirates, important factors by 2050.14 Dubai’s population demographic noted include: the relatively high percent- currently has 88 percent of its population age of consanguine relationships, the ele- below the age of 45 (see Figure 2). The vated genetic risk for diabetes compared Dubai Health Authority expects that as its with many other populations of the world, population ages, the rates of diabetics will and the widespread use of traditional cloth- double from 350,000 in 2000 to 700,000 in ing. What is unique about the clothing in 2030. This is a major concern as more life- the Gulf region is that men and women style diseases mean higher per capita health wear loose-fi tting robe-type outfi ts known care costs are to be expected. in Arabic as the white “dish dash” for men, and the black “abaya” for women. Unlike Obesity and Diabetes what men and women wear in western- More than 70 percent of the UAE pop- ized countries, such as jeans and fi tted ulation is overweight.16 According to a shirts where one notices their weight and

Figure 2. Distribution of Various Age Groups in the UAE population in 2011.

85+ 80–84 75–79 Female Male National 70–74 Expatriate 65–69 60–64 55–59 50–54 45–49 40–44 35–39 30–34 25–29 20–24 15–19 10–14 05–09 0–4 1,000 10,000 90,000 190,000 290,000

Source: Statistics Center of Abu Dhabi population estimates for Nationals, additional Health Authority of Abu Dhabi assumptions and analysis based on raw insurance data, Health Statistics 2011 from the Health Authority of Abu Dhabi.15 Health Policies and Intervention Strategies 91

waistline, the loose clothing conceals familial aggregation of diabetes. An odds the body shapes of individuals in the ratio of 6:2 for familial aggregation sug- UAE. The climate also plays into the obe- gests a causal association with diabetes. sity problem. With such hot climates most More studies need to be done to confi rm people sit indoors or go from place to place the accuracy of such fi ndings.18 by car. It is not common to see an Emirati walking or exercising outdoors. One might The Health System of the be considered from a lower socioeconomic United Arab Emirates status if she or he did not go from place The World Health Organization (WHO) to place by car. Lastly, another obstacle in defi nes health systems as “all organiza- the fi ght against obesity, which is similar to tions, institutions, and resources that are that of many other Mediterranean and Afri- devoted to producing health actions.19 can countries, involves the idea that being Health system strengthening (HSS) overweight shows good health, wealth, and involves strategizing in order to improve success.17 the functions of the system with the goal of improving health through improved Marriage Trends access, quality, coverage, or effi ciency in Intra-familial marriage is common health care.20 “The World Health Report in the UAE and has been thought to be 2000 identifi es four key functions of the related to the genetic predisposition for health system, which include: steward- diabetes. Because the Emirati population ship (often referred to as governance or is so low in comparison to the expatriates oversight), fi nancing, human and physical of the country, the government is push- resources, and organization and manage- ing for marriage between Emiratis. This ment of service delivery.”21 Stewardship is being done with the government setting in health is composed of factors that are up incentives, such as a marriage fund, for not directly related to the health care sys- Emiratis who marry other Emiratis. Intra- tem, yet have an impact on the health of familial marriage is a common practice a population, some of which include: the that has taken place over hundreds of years environment, levels of education, and pov- in Bedouin tribes and continues to be com- erty. The World Health Organization has mon today. There have been some studies recommended the development of health done to test for possible links to diabetes. sector policy, as the primary role of a gov- One study was conducted in Saudi Arabia, ernment’s Ministry of Health, with the aim where intermarriage, especially between of improving the performance of the health fi rst cousins, a common form of mar- system and the health of the population.22 riage, was analyzed to see whether there The health care laws of the UAE are reg- was a correlation between consanguin- ulated at the Federal level and the Emirate ity and diabetes prevalence. The study level. Federal regulation dates back to the involved standardized questionnaires that 1970s and 1980s.23 Public health care ser- were administered detailing general demo- vices are controlled and regulated by four graphics, dietary patterns, anthropometric main authorities: The Ministry of Health, the values, family history, consanguinity, and Health Authority of Abu Dhabi (HAAD), the 92 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Dubai Health Authority (DHA), and Emir- Financing Health Care between ates Health Authority (EHA). the Public and Private Sector The main UAE Ministry of Health, Currently the region uses and consumes established in 1972, is in charge of licens- about 25 percent of health care spending in ing companies and individuals who provide private care centers. In recent years, it has health care, manage health institutions, and become common for many patients to go regulate various aspects of the fi eld. The to public government hospitals for actual Dubai Health Authority is authorized to health care treatments, and only use private regulate all the health facilities in Dubai.24 care facilities for diagnoses. With public In 2009 the Dubai Health Authority imple- government hospital occupancy rates aver- mented a new Policy and Strategy Sector. aging 80 percent, private hospital occu- Its mission is to oversee the development pancy is signifi cantly lower. The reason of policies and strategic planning under- for this pattern is that insurance companies taken by the Dubai Strategic Plan. This do not cover care undertaken at private department works on intense policy anal- facilities.28 ysis, strategy execution, and partnership A study done by McKinsey and Com- with consumers, providers, regulators, and pany analyzing the health ailments of the funders of health and health care in Dubai GCC countries indicates various aspects and the UAE.25 of the current health policy situation.29 The Health Authority of Abu Dhabi The company created a propriety model (HAAD) is the entity responsible for regu- of health care demand by focusing on fi ve lating the health sector within Abu Dhabi. change factors, which include: popula- It works by setting up evidence-based pro- tion growth, the demographic profi le, the tocols, and developing structured programs development of risk factors, treatment aimed at patient and physician compliance, patterns, and medical infl ation.30 Treat- and geared towards measurable results.26 ment demand is expected to increase in the Abu Dhabi established a unique division region by 240 percent by the year 2025. called the Abu Dhabi Health Services Hospital bed demand is predicted to more Company (SEHA) in 2007. The Abu Dhabi than double in the UAE and Saudi Arabia. government owns this public joint stock Currently the highest utilization of health company and it manages and owns the care in the GCC is by publicly run, govern- public health facilities. This entity is cur- ment health care facilities. Research found rently collaborating with renowned health that many individuals are not satisfi ed by care institutions such as Johns Hopkins the quality of care in public facilities, and Medical and The Cleveland Clinic for the complain about long wait times, uncom- management of certain hospitals. Lastly, fortable facilities, and limited availability the northern Emirates, Ras Al Khaimah, of appointments.31 Ajman, Umm Al Quwain, Sharjah, and The privatization of health care delivery Fujairah, as of 2009 are under the control is focused on three areas: health care deliv- of the Emirates Health Authority, which ery, health insurance, and support services. has similar regulations as the HAAD and Government entities in the GCC region, the DHA.27 such as the UAE Health Ministries, now Health Policies and Intervention Strategies 93

want to lessen their involvement in health lic and private providers, which would both care delivery, and focus such efforts on pri- be covered by insurance. Qatar and Oman vate health facilities, as well as have private are in the planning process of implementing entities manage public facilities. With the mandatory health insurance for all of their construction sector boom in the UAE and citizens.33 neighboring GCC countries, many expatri- Enforcing such a policy change would ate workers fl ocked to the region for work. completely revamp the health care system Expatriates now comprise 25 percent of the of the country. It is also their hope that the GCC population and 80 percent of the UAE private sector would assist in improving the population (4.8 million out of 6 million).32 quality of other areas, such as implementing The fi rst two ministries of health to pass an electronic medical record system. There laws requiring employers to purchase private are many issues that must be taken into con- insurance for their workforce of expatriates sideration while trying to privatize the health are Saudi Arabia and Abu Dhabi (UAE). The care industry. A patient threshold volume goal of policymakers is aimed at creating a needs to be maintained in order to establish mixed public-private payer method to get a private model of care based on general patients to have a choice between using pub- practice. A referral network is lacking, and

Figure 3. Current Model of Problems with the Health Care System in Abu Dhabi

Model of care How health services are currently used, what’s wrong? There are no Open Outpatient Inpatient systems in place 1 to support patient access self-cae and Primary care management of center/clinic chronic disease.

Ambulance Hospital ER

Hospital Specialist admission

Hospital Screening Specialist programs are not (yet) fully aligned to prevent and treat chronic Screening conditions.

Patient access to services is not streamed leading Laboratory and Diagnostic services to over-servicing over- radiology are not optimally supply and inappropriate integrated into service use. Diagnostics treatment paths.

Source: Strategy Analysis, Health Statistics 2011, Health Authority of Abu Dhabi.34 Note: Access to SEHA hospital specialists is only via referral from SEHA Centers/Clinics and ER departments. Some SEHA Hospital ER departments also direct nonemergency patients to adjacent “Urgent care” centers. 94 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

needs to be established, as it is an important This initiative came as part of the Abu Dhabi portion of a health care system. Lastly, clini- Cardiovascular Disease program, aimed at cal staff and support needs to increase in the the number one disease killer in the coun- region. try. Individuals ranging in age from 18 to 75 years seeking the UAE government’s new, Health Care Awareness and Quality free, comprehensive health insurance plan, In the UAE and the rest of the GCC made available to all UAE nationals residing region there is a lack of effective regulation in Abu Dhabi, are now required to undergo by the government, and most patients do not screening for cardiovascular risk factors. understand what good quality in health care As a result, UAE nationals receive Thiqa entails. Patients tend to make their health health insurance cards giving them access to care decisions based on word of mouth, the private and public sector health care.37 This external physical appearance of the facility, screening program was developed in accord- and advertising. Better local treatment needs ance with recent recommendations from to be made available as so much money the Institute of Medicine, WHO, the United from the government is spent on sending Nations, and the World Economic Forum for citizens abroad for treatment. In 2007, the addressing this chronic disease at the popu- Dubai government sent approximately 1,000 lation level. Results of this screening pro- patients abroad for medical treatment cost- gram found many cardiovascular risk factors ing $48 million. An estimate of $2 billion is overlapping with diabetic patients. Among spent annually in sending citizens abroad for the patients, 54 percent were obese and 82 medical treatment.35 For those who under- percent were centrally obese (excessive stand health care quality, they do not trust abdominal fat), 73 percent had dyslipidemia, the quality of care of either the facilities or and 59 percent were hypertensive. It was the physicians in the region, and decide to found that by the ages of 18 to 20 years old, go abroad for treatment. Many times there 49 percent of men and 34 percent of women are also cases in which specialists in treat- were already overweight or obese.38 As part ing certain conditions cannot even be found of the Thiqa insurance renewal, female in the country, leaving individuals no other adults between the ages of 40 and 69 are option than to go abroad. required to be screened for breast cancer.39 According to the Dubai Health Author- Whole population screening programs are ity, awareness is also lacking regarding the a step in creating a prevention type mindset importance of primary health care and pre- in the population. The Weqaya program tar- vention. Primary care service utilization by geted interventions in well-stratifi ed groups UAE nationals is at or below benchmark using new technology. levels. Given the high levels of diseases To work on quality issues, the HAAD such as diabetes, primary care utilization, is starting the “Pay for Quality” initiative such as screenings and yearly physical “where health facilities will receive incentive exams, should be much higher.36 A strategic payments for compliance with HAAD stand- move came in April 2008 when Abu Dhabi ards of care.” In addition, another portion of decided to implement a program called The the initiative will tackle the volume incentive Weqaya Program (Arabic for prevention). needed by physicians of maintaining a certain Health Policies and Intervention Strategies 95

number of patients in the facilities. HAAD The aim is to lower the prevalence rates in has begun creating databases for vital sta- the country by 1 to 2 percent per year.41 tistics and health data with the hope of inte- Cultural Attitudes and grating business models for such databases, releasing its fi rst collection of data at the end Barriers towards Health of 2011 (see Figure 4). The goal of both ini- In the nearby country of Qatar, ranked as tiatives is to complement the work of such the richest nation in the world, one of the few providers as a step in creating a higher qual- traditions that Qatari nationals have been able ity system of health care for the other emir- to maintain is intermarriage between family ates to follow in the country.40 In 2009, the members. Combining a lifestyle of wealth Ministry of Health also began an initiative to and interfamilial marriage, citizens are complete the fi rst Emirati diabetic database. becoming more obese and passing on many The objective in creating the database is to genetic disorders. This country ranks as the provide the Emiratis of the country with the sixth highest in the world concerning diabe- required information about diabetes, includ- tes prevalence. There are now many public ing methods of prevention and treatment. awareness campaigns intended to educate The Minister of Health, Humaid Mohammad the public about healthy eating, exercise, fi t- Obaid Al-Qutami, launched the Web site as ness, and sport habits.43 The problem is that the fi rst UAE National Registry of Diabetics. although most people have now heard of

Figure 4. Abu Dhabi Diabetes Prevalence 2011 released by HAAD

Prevalence of Diabetes Mellitus by age group

70% 90,000 % At risk of diabetes Average 80,000 60% Population 70,000 50% 60,000

40% 50,000

30% 40,000

30,000 20% 20,000 10% 10,000

0% 0 20–29 30–39 40–49 50–59 60–69 70–79 80+

Source: Preliminary analysis of Weqaya sample of 112,301 UAE Nationals in the Emirate screened in 2008–9, Health Statistics 2011 from Health Authority of Abu Dhabi.42 96 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

diabetes and its health implications no one is on health awareness, do not believe in mod- really taking steps to avoid the disease until it ern medicine. is too late. This passive attitude is a common A physician interviewed in Dubai for this problem in the culture of the countries of the review noted that many families tie stigma to GCC. Although the Ministry of Health may having a disease, so they go into denial and have policies and programs aimed at meth- do not address the problem. “Some patients ods of prevention, the mindset and lifestyles will be offended if you even point out that of the people are not yet cooperative enough they have a problem.” Another critical issue to care. The government is trying to con- is that “people of the UAE do not grasp the struct parks, sidewalks, pedestrian crossings, concept of caring for themselves; they need residential neighborhoods, outdoor exercise someone to do everything for them. Even if equipment, and automated bicycle rental I teach someone how to inject him/herself kiosks. These are all great projects to get the with insulin, most still refuse to do it them- public physically active; however, the climate selves.” A big problem in the region is also of Qatar along with UAE and the other GCC the food that people are consuming. Not countries, has people spending at least four only are the western foods that have entered to fi ve months of the year mainly indoors the region bad for your health, but also many as they are unable to tolerate the summer Arab foods are quite fattening. “The culture heat, which can reach approximately 50°C of the society now has people sitting around (~120°F). In addition, traditional women of for hours socializing over food into very late any GCC country will not be seen exercising hours of the night, which is bad for one’s outdoors, or in front of men, as it is not cul- health.”45 turally acceptable. There are now many health care work- In order to tackle the many chronic con- shops for doctors including primary care ditions plaguing the countries, educating physicians. Currently expatriates comprise the public as a preventive strategy is essen- 82 percent of physicians in the GCC region. tial. Preventative care is still at a primitive An issue the region has faced with expatri- level in the Middle East. The current health ate physicians is that they have shown a high policy model of care by the government of turnover rate. Also with many expatriate AU: What Abu Dhabi does not focus on prevention of physicians having to collaborate their work, do u mean by “col- health ailments. More awareness needs to there is a collection of different cultures hav- laborate” be implemented in order for patients to feel ing to work together from different training here? empowered, more educated, and actively backgrounds and work ethics, which can Please clarify. seek out screening.44 Many Arabs are pas- sometimes make things diffi cult. Cultural sive and they wait until something happens understanding is essential between the phy- before they go to the doctor. This fatalistic sician and the patient. When this is not pre- behavior is rooted in the Arab/Islamic cul- sent, the physician will have a hard time in ture’s surrender to “God’s will.” This atti- dealing with the mindset of the patient. This tude offsets the recognition of necessary poses the need for more GCC national phy- preventative measures. Many patients, espe- sicians and nurses in the health care indus- cially those of Bedouin origins, that hold try, with quality training and standards of tightly onto traditions and are not educated performance.46 Health Policies and Intervention Strategies 97

Food Marketing Techniques regulation has been implemented in New York, California, and Oregon. In the United An interview conducted with a market- Arab Emirates, the fast-food industry has ing executive at a regional canned foods been growing at a 25 percent annual rate.49 company located in Dubai indicated that the Restaurants lack healthy food options, as branding image of Middle Eastern women it is not considered part of the culture. The is the primary target of all marketers in the government needs to take initiative in this region’s food industry. “Our target audience, regard, mandating calorie counts and health women of the Middle East, either newly food options as a preventative measure, and married or those from the age of 30 to 40 to help those already plagued with a chronic gain a form of self-validation from feed- disease. Understanding food labeling and ing their loved ones. They perceive them- nutritional content is an important part of selves as good mothers when they are able understanding what is good and healthy for to prepare a good meal for their families. In one’s body. Many people in the region do their defi nition, a good meal is a delicious not understand how to read food labels. The meal that is well-presented.” It was further government should require that such infor- explained that if their message is not com- municated in this way, many women would mation be taught in school health classes. fi nd they are unable to relate to the brand. Further, during the interview, the market- Health Initiatives ing executive contrasted Middle Eastern Initiatives have begun in schools aimed branding and US branding by stating, “In at educating students on the importance of the United States, food branding is more a good diet, after an alarming 70 percent focused on health-consciousness. However, obesity rate has been noted in the country. In in the Middle East, many target consumers the emirate of Ras Al-Khaimah, a six-month have still not caught onto the importance of awareness campaign has been started in the healthy products. Even in the supermarket local schools. The goal of the program is to aisle, you can easily see there are very few get teachers to motivate the children to stay healthy products, which make them a niche away from junk food through various school part of the market”.47 events.50 In order to tackle the diabetes epidemic, Educational awareness campaigns are also the government needs to start by tackling beginning to appear on public displays. The one of the sources of the problem, the food Walk for Life campaign hosted annually for and restaurant industry. According to Al the past fi ve years showed the greatest turn Haddad of The National newspaper of Abu up of 16,000 individuals on November 25, Dhabi, many customers in Dubai and Abu 2011, at the Yas Marina Circuit, as part of Dhabi have begun to voice that they would the “award-winning public health awareness AU: please like to have nutritional information on their 51 confi rm campaign ‘Diabetes-Knowledge-Action.’ ” location of 48 menus. In the United States, a federal law This event was held under the patronage of close quote. is being implemented that requires all big Her Highness Sheikha Fatima Bint Mubarak, restaurant chains to put calorie information organized by the Imperial College of London, on menus and drive-through signs. This sponsored by the Abu Dhabi government 98 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

holding company Mubadala, and supported resources needed due to its mountainous by The National Health Insurance Company, terrain.54 The United Arab Emirates has Daman. The goal of the event was to bring undertaken steps in reforming its health more public awareness to diabetes and to care services sector. A growing and aging push individuals to take charge of their life- population with many chronic ailments will style by embracing a healthy diet along with strain the health sector if things are not done regular physical activity. Since the start of to revamp the system, including chang- the campaign, it has reached out to more than ing current policies. Progress is currently 125,000 individuals in the UAE.52 According underway in the country with the govern- to the National Newspaper of Abu Dhabi, a ment taking initiative in improving quality, similar campaign in 2013 called the “Kobe awareness initiatives, screening programs, Health and Fitness Weekend” involves the and a new emphasis on the private sector UAE government hosting the celebrity bas- getting more involved by increasing health ketball player Kobe Bryant with the goal of care facility options. Although education raising more diabetes awareness.53 and awareness are now increasing concerning diabetes, the challenge remains of Conclusion changing people’s attitudes and behaviors towards other health problems. Educating The path to health care reform is more the youth as a preventative strategy, and achievable in the UAE compared to lower emphasizing the importance of primary care income countries in Middle East such as must increase in the UAE and throughout Yemen, which lacks access to the health the region.

REFERENCES

1. Abduelkarem, A, “Childhood obesity and 5. McLaughlin, G, Kemp, J, and Stoltz, K, Dia- type 2 diabetes: a growing public health chal- betes in the UAE portrait of a country at risk lenge in UAE,” Diabetes Voice, 50(3):9–11 (Jan. 22, 2013). Retrieved from http://www. (2005). middleeasthealthmag.com/cgi-bin/index. 2. Glass, A, “Diabetes costs the UAE mil- cgi?http://www.middleeasthealthmag.com/ lions in 2007” (October 8, 2008). Retrieved jan2013/feature11.htm October 15, 2011, from Healthcare-Arabian- 6. Margolis, S, et al., supra, n.4. Business.com, http://www.arabianbusiness. 7. United Arab Emirates National Bureau of Sta- com/533531-diabetes-costs-the-uae-millions- tistics. Population Estimates 2006–2010. in-2007. 8. International Diabetes Federation, Diabetes 3. Edwards, HS, “The Richest, Fattest Nation Atlas 2007. on Earth (It’s not the United States)” (2011). 9. Ely, J, Zavaskis, T, and Wilson, S, “Diabetes Retrieved November 20, 2011, from The and stress: an anthropological review for Atlantic, http://www.theatlantic.com/life/ study of modernizing populations in the US- archive/2011/11/the-richest-fattest-nation- Mexico border region,” Journal of Rural and on-earth-its-not-the-united-states/248366/. Remote Health Research, Education, Practice, 4. Margolis, S, Carter, T, Dunn, E, and Reed, R, and Policy, 1–17 (2011). “The health status of community based elderly 10. Margolis, S, et al., supra, n.4. in the United Arab Emirates,” Archive of Ger- 11. United Arab Emirates. (n.d.). Retrieved Octo- ontology and Geriatrics, 37: 1–12 (2002). ber 22, 2011, from CIA-The World Factbook: Health Policies and Intervention Strategies 99

https://www.cia.gov/library/publications/ the UAE, http://www.lw.com/upload/pub- the-world-factbook/geos/ae.html. Content/_pdf/pub3951_1.pdf. 12. Ward, W. and Younis, M, Steps Toward a 24. Id. Planning framework for Elder Care in the Arab 25. Health Policy and Strategy (September 13, World, Springer, NY, USA (2013); Younis, M, 2010). Retrieved November 5, 2011, from Al-Hajeri, M, Celik, Y, Kisa, A, Richard, P, Dubai Health Authority, http://www.dha.gov. and Parkash, J, “Healthcare for Aging Popula- ae/EN/SectorsDirectorates/Sectors/Health- tion in Kuwait,” Ageing International, August Policy/Pages/default.aspx. 2012. 26. Weqaya Disease Management Programme— 13. Margolis, S, et al., supra, n.4. Expression of interest, Abu Dhabi: Health 14. Why invest in Healthcare in Dubai (June 12, Authority—Abu Dhabi (2011). 2010). Retrieved November 2, 2011, from 27. Latham and Watkins, supra, n.23. Dubai Health Authority, http://www.dha. 28. WHO (World Health Organization), World gov.ae/EN/SectorsDirectorates/Directorates/ Health Report 2006: Working Together for HealthRegulation/HealthCareInvestment/ Health. Geneva: WHO 2006, retrieved Pages/WhyinvestinHealthcareinDubai.aspx. November 15, 2011, from http://www.who. 15. Health Authority of Abu Dhabi, Health Statis- int/whr/2006/en/index.html. tics 2011 (October 12, 2012). Retrieved from 29. Mourshed, M, Hediger, V, and Lambert, T, http://www.haad.ae/HAAD/LinkClick.aspx?f Gulf Cooperation Council Health Care: Chal- ileticket=JY0sMXQXrOU=&tabid=1243. lenges and Opportunities. McKinsey & Com- 16. Health Ministry Increases Diabetes Fight, pany (2009). (April 13, 2009). Retrieved March 25, 2010, 30. Id. from Ministry of Health United Arab Emir- 31. Id. ates, http://www.moh.gov.ae/en/NewsDe- 32. Dubai Health Authority, supra, n.14. tails461.aspx. 33. Id. 17. Abduelkarem, A, “Childhood obesity and 34. Health Authority of Abu Dhabi, supra, n.15 type 2 diabetes: a growing public health chal- 35. Dubai Health Authority, supra, n.14. lenge in UAE,” Diabetes Voice, 50(3):9–11 36. Id. (2005). 37. Key Health Statistics for Abu Dhabi. Abu 18. Anokute, C, “Suspected Synergism Between Dhabi, United Arab Emirates: Health Author- Consanguinity and Familial Aggregation in ity of Abu Dhabi (2010). Type 2 Diabetes Mellitus in Saudi Arabia,” 38. Hajat, C, Harrison, O, and Al Siksek, Z, The Journal of the Royal Society for Promo- “Weqaya: A Population-Wide Cardiovascu- tion of Health, 111(4):167–169 (1992). lar Screening Program in Abu Dhabi, United 19. Islam, M, A Health System Assessment Arab Emirates,” American Journal of Public Approach: A how-to manual, (2007). Health, 1–6 (2011). Submitted to the US Agency for International 39. Health Authority of Abu Dhabi, supra, n.37. Development in collaboration with Health 40. Health Authority of Abu Dhabi, supra, n.26. Systems 20/20, Partners for Health Reform 41. Health Ministry Increases Diabetes Fight. (2009, plus, Quality Assurance Project, and Rational April 13). Retrieved March 25, 2010, from Min- Pharmaceutical Management Plus, Arlington, istry of Health United Arab Emirates: http:// VA: Management Sciences for Health. www.moh.gov.ae/en/NewsDetails461.aspx. 20. Id. 42. Health Authority of Abu Dhabi, supra, n.26. 21. Partners for Health Reform, Health systems 43. Edwards, HS, supra, n.3 strengthening: An introduction, (2005), 44. Health Authority of Abu Dhabi, supra, n.37. http://www.healthsystems2020.org/ 45. Dubai Physician—Interview (October 22, ﬁ les/571_ﬁ le_01_Chapter_1.pdf. 2011) (A. Mosaad, Interviewer). 22. Islam, M, supra, n.19. 46. Mourshed, M, et al., supra, n.29. 23. Latham and Watkins (2011), retrieved Novem- 47. Marketing Executive—Interview (November ber 20, 2011, from Healthcare Regulation in 11, 2011) (A. Mosaad, Interviewer). 100 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

48. Al Haddad, A, Appetite to put calorie count 51. Walk for Life (November 25, 2011). Retrieved on the menu (September 3, 2011). Retrieved November 26, 2011 from Diabetes UAE November 22, 2011, from The National, 2011, http://www.diabetesuae.ae/walk-life-1. http://www.thenational.ae/news/uae-news/ 52. Id. health/appetite-to-put-calorie-counts-on-the- 53. NBA star Kobe Bryant to visit UAE to raise dia- menu. betes awa reness (June 19, 2013). Retrieved 49. Id. from http://www.thenational.ae/news/uae- 50. RAK kicks off campaign to promote healthy news/health/nba-star-kobe-bryant-to-visit- eating habits among students (March 18, uae-to-raise-diabetes-awareness. 2010). Retrieved April 20, 2010, from The 54. Younis, M, “Comparative effectiveness research Free Library, http://www.thefreelibrary.com/ and challenges to healthcare reform in Middle RAK+kicks+off+campaign+to+promote+hea East and USA,” Journal of Comparative Effective- lthy+eating+habits+among...-a0221385336. ness Research,2(3): 223–225 (May 23, 2013). Determinants of Differentials in Pneumonia Mortality in the UK and France

Rizwan ul Haq, Patrick Rivers, and Muhammad Umar

Pneumonia is one of the major causes of death in the world. Age-adjusted mortality from pneumonia in the United Kingdom was three times higher than it was in France in 2004. The purpose of this article is to ﬁ nd the underlying determinants of pneumonia mortality differences between these two countries. The main research question is “what are the determinants of pneumonia mortality in the UK and France?” Reviewing the underlying determinants of health inequalities, we expected that behavioral factors, environmental factors, and the health care system would account for the differences, but they do not actually account for much of the differences in Pneumonia mortality between the UK and France. The main difference is due to data quality problems particularly relating to diagnosis and certiﬁ cation in both countries. Key words: pneumonia, mortality differentials

1. Introduction tions include pleural effusion (an accumulation of fl uid around the lungs), breathing 1.1. Background diffi culties, septicemia (a spread of the infection to the patient’s blood), and the Pneumonia is an infl ammatory lung disease most dangerous one, blood poisoning.4 normally caused by bacteria, a virus or chem- The main determinants of pneumonia ical irritants. Moreover any lung injury or ill- mortality include environmental, demo- ness such as lung cancer or alcohol abuse may graphic, and behavioral factors, as well as also cause the disease. Pneumonia belongs to the health care system of a country. Being a respiratory disease group that ranks among in poor health, old age, smoking, alcohol the top 10 causes of death in the world. Viruses abuse, heart disease, lung diseases, and low cause half the reported pneumonia cases and immunity to infection (AIDS) are among the are believed to cause less severe illness than major predisposing factors for pneumonia. A bacteria-caused pneumonia.1 pneumococcal vaccine is available for pre- Pneumonia is more common in the venting pneumonia caused by Streptococcus United Kingdom (UK) than in France. Age- standardized mortality from respiratory diseases was about 67 per 100,000 people in the Rizwan ul Haq, is research demographer at Pakistan UK and 27 per 100,000 people in France in Institute of Development Economics, Islamabad, 2010.2 Pakistan. Pneumonia is considered as one of the fatal Patrick Rivers, is a Professor of Healthcare Finance, Management & Strategy in the the College of Applied respiratory diseases. Children and elderly Sciences & Arts and a Professor (Joint) of Manage- people most commonly get pneumonia, and ment in the College of Business at Southern Illinois the autumn or winter are the two seasons University Carbondale. in which most of the pneumonia cases are Muhammad Umar works at COMSATS University of reported.3 Science and Technology, Abbotabad, Pakistan.

The symptoms depend on the severity and J Health Care Finance 2014; 40(3):101–110 type of infection. Most common complica- Copyright © 2014 CCH Incorporated 101 102 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

pneumonia infection. Eradication of bad pneumococcal pneumonia. Other microor- habits including smoking and drinking may ganisms that can cause other forms of pneu- also reduce the susceptibility of pneumonia. monia are classed as “atypical pneumonia” and include Legionnaire’s disease (caused 1.2. Research Question and Objective by a bacterium) and SARS (severe acute respiratory syndrome, which is caused by a The main objective of this study is to fi nd virus). Moreover, Pneumonia can be nonin- out the main determinants of differences in fectious and is normally known as “aspira- pneumonia mortality between the UK and tion pneumonia.” The major cause of this France. The main research question is “what type of pneumonia includes inhaling sub- are the determinants of pneumonia mortal- stances, such as caustic chemicals, food, or ity in the UK and France?” The specifi c vomit, into the lungs. research questions are: The symptoms of pneumonia include fever, dry cough, headache, muscle pain, • What is pneumonia, and what are its weakness, and increasing breathlessness, causes and prevalence? which are similar to infl uenza symptoms. • What are the differences in the socio- The symptoms may be quite acute begin- economic and demographic characteris- ning with a dry cough and eventually pro- tics of populations of the UK and France? gressing to a cough with a green/yellow or • What are the differences in behavioral rust-colored smelly phlegm. and environmental factors in the subject A few immunizations are available to populations and how do these differ- prevent the causes of pneumonia. For exam- ences relate to differences in pneumo- ple, Haemophilus infl uenza type B vaccine nia mortality? (“Hib” vaccine) prevents fl u. Normally a • How do the health care systems of the single dose of the vaccination is given on two countries differ? annual basis to babies, to the elderly, and to The article is organized in to fi ve sections. those with chronic lung, heart, or kidney dis- In section one research questions with the eases, or with a weakened immune system.5 objective of the article are discussed. Sec- 2.2. Theory tion two deals with theory and conceptual framework. Data sources and methods are Pneumonia morbidity and mortal- discussed in section three. Section four con- ity may vary with age, sex, ethnicity, and tains the results followed by the conclusion related demographic characteristics. As in section fi ve. mentioned above, the elderly, children and those with certain health problem, includ- 2. Theory and Conceptual Framework ing chronic obstructive pulmonary disease (COPD), diabetes, diabetes mellitus, 2.1. What Is Pneumonia? congestive heart failure, and sickle cell anemia. High vulnerable groups include Bacteria are the main cause of the infec- people with AIDS, those undertaking can- tion leading to “typical pneumonia,” includ- cer therapy, those who have had an organ ing Streptococcus pneumonia, which causes transplant, and chronically ill patients.6 Determinants of Differentials in Pneumonia Mortality in the UK and France 103

Thus, pneumonia triggers with a viral upper- associations between smoking and res- respiratory-tract infection or, among the piratory and cardiovascular diseases.10 elderly, with fl u, and in most case people However, smoking and pneumonia demon- with some pre-existing conditions, infec- strated a weak but statistically signifi cant tions, or weakened immune systems fall relation for people belonging to various age prey to this disease. groups.11 Studies on the infl uence of gender as Like smoking, consuming excess alco- a risk factor for pneumonia provide con- hol is also a predisposing factor to various tradictory results. According to “An Atlas diseases. on Mortality in the European Union,” The vulnerability of alcohol users to produced by the European communities, various infectious diseases – among those, infl uenza and pneumonia are responsible bacterial pneumonia exerts the strongest for 3 percent of male deaths and 4 percent association with alcohol use – increases due of female deaths in the European Union.7 to its suppressive effects on the immune sys- These infectious diseases have a greater tem; however, the underlying mechanism of impact on female mortality than chronic such a relationship continues to evolve.12 diseases, in contrast to the situation for men. Based on animal experiments, which The large number of older — and hence may not be demonstrated epidemiologi- more susceptible — women explains why cally in humans, Pistelli, et al., attrib- the differences in mortality between the uted the ambient air pollution along with sexes are less marked for these pathologies. smoking as main reason for increas- Others suggest that males as compared with ing incidents of respiratory infections.13 females and Blacks as compared with Cau- Components of air pollution, such as ozone casians are more likely to get pneumonia.8 and nitrous oxide, and exposure to tobacco The latter is mainly attributed to socioeco- smoke damage clearance of bacteria from nomic differences between the two groups. mouse lungs and increase fatality rates fol- There is a common argument and exten- lowing acute infection. sive literature suggesting that behavioral (lifestyle) factors such as smoking, alcohol 2.3. Conceptual Framework consumption, diets, and exercise contribute to health inequalities among populations. Figure 1, below, shows the conceptual Smoking is one of the behaviors that pre- framework of the study. The demographic disposes individuals to various diseases and factors of the two subject populations, the causes premature deaths. Cigarette smok- UK and France, like the age distribution of ing leads to impaired immune defenses the population, have an effect on individual and increases the risk of certain infections.9 behavior. It has been observed that age is Further, smoking alters the responsiveness quite important as children and the elderly of infl ammatory cells and lung function suf- populations are the most vulnerable groups. fers an accelerated rate of decline with age. The elements of age and gender also have In a prospective cohort study using data been included in the model because it is from 40 years of follow-up of smokers, Doll, assumed that both infl uence habits, nutri- et al., observed the strongest cause-specifi c tion, lifestyles (sedentary or active), and 104 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

Figure 1. Conceptual Framework of the Study

Individual Demographic Behaviors Factors (lifestyle)

Vulnerablity to Mortality Socio-Economic Physical Pneumonia from Environment Environment Infection Pneumonia

Health Data Health Care System Policy Quality risky behaviors (smoking, drinking, and so care system is assumed to decrease the vul- on).This behavior affects the vulnerability nerability of the individual, while an inef- of the individual to acquire pneumonia and fi cient and out-of-date health care system is eventually leads to the higher or lower level supposed to increase the vulnerability of the of mortality caused by pneumonia. individual, which has an ultimate effect on It is assumed that the physical environ- mortality by pneumonia. It is also pertinent ment, including climate and topography of to mention that the vulnerability for pneu- the country, will have a direct effect on the monia infection also affects the health care population’s vulnerability to pneumonia. The system. For instance, if there are a large vulnerability increases in a cold and wet cli- number of vulnerable people, the health mate, which ultimately increases mortality. care system may fail to provide better health Other environmental factors such as pollu- care facilities and services to them. tion, density of population, and congested The connection of vulnerability to housing facilities, all are infl uenced by the pneumonia to mortality by pneumonia socio-economic environment. These factors can also pass through the health care sys- determine the vulnerability of the individual to tem. If the vulnerable population (per- acquire pneumonia and eventually will affect sons having a greater risk of acquiring the individual’s mortality by pneumonia. pneumonia) is provided with better health The health policies of the countries under facilities and a better care system, they study has been given importance, and they can be cured, or the chances of mortality are included in the model. The favorable can be reduced. and conducive health policies are supposed to create a better health care system (hos- 3. Data and Methods pitals, doctors, paramedical staff, clinics, immunizations, and sanitation system). An As mentioned earlier, age-standardized effi cient, need-based, and modern health mortality from respiratory diseases was Determinants of Differentials in Pneumonia Mortality in the UK and France 105

about 67 per 100,000 people in the UK and 4. Results 27 per 100,000 people in France in 2010. These statistics are taken from the Euro- 4.1. Demographic Factors stat, and all the diseases in categories J12 Mortality from pneumonia in the UK to J18 are classifi ed as due to pneumonia are strikingly high among the European from the International Classifi cation of Union (25 countries) according to data Death (ICD). The actual data is provided from Eurostat. Death rates caused by res- by the member countries to the Eurostat. piratory disease are signifi cantly lower Eurostat has a quality check on the data, for France as compared to the UK in which includes consistency checks, to 2010: 27 as compared to 67 per 100,000 identify incoherent data. The quality of the inhabitants.14 Eurostat data is subject to classifi cation In addition, it is observed that the age and the coding procedure in each country, composition of the two populations (France especially in certifi cation processes and and the UK), as shown in Table 1, is almost diagnosis. similar. In both populations the proportion To explain the differences in the death of children and the elderly is 18 percent rates due to pneumonia in the UK and and 16 percent respectively in 2012. But, France, we used literature review and data female life expectancy is slightly higher in on health care utilization, behavior and life- France than in the UK. Life expectancy at style, physical environment, socio-economic birth is 78years for males and 85 years for status, and demographic characteristics. females in France, while it is 78 years and Health policies in the two countries were 82 years for males and females respectively also studied. Data quality and comparability in the UK. This means that age does not was also analyzed. help to explain the difference in reported Secondary data was taken mainly from pneumonia mortality between France and the World Health Organization (WHO), the UK. Eurostat databases, the Offi ce for National Though data on age-specifi c incidence Statistics for the UK, and the Institute for of pneumonia is limited for this analy- Alcohol Studies. Since the data are taken sis, it is observed from studies that both from different sources comparability may be the incidence and severity of pneumo- questionable. nia infections are higher among children

Table 1. Population Age Distribution & Life Expectancy at Birth, France & UK, 2004

Deaths Life per 1000 % age less Life Expectancy, Expectancy, Country Total population population than 15 % age 65+ Male Female France 63.6 millions 9 19 17 78 85 UK 63.2 millions 9 18 17 78 82

Source: PRB, 2012, http://www.prb.org/pdf12/2012-population-data-sheet_eng.pdf. 106 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

younger than fi ve years of age and in of heavy episodic drinkers as compared to the elderly (above 60 years old). In chil- France.20 dren, the majority of deaths occur among According to an estimate, about 8,000 to infants. In the UK, the incidence rate of 40,000 people per year die due to alcohol community acquired Pneumonia is esti- use.21 From 1991 to 2005 the death rate for mated at 34 per 1,000 in people aged 75 diseases directly related to alcohol use dou- years and older.15 bled in the UK from 6.9 per 100,000 popula- Differences in pneumonia mortality tions to 12.9, and the majority of the deaths between men and women was observed in occurred among males. both the UK and France. Mortality from pneumonia was higher among men than Physical Exercise women in both the UK and France. Physical activity and physical fi tness are known to be associated with a number of 4.2. Behavioral (Lifestyle) Factors health-promoting behaviors. France and the UK differ in the kinds of sports most com- Smoking and Pneumonia monly practiced by their people. The sports popular in France include basketball, football In the UK half of about 13 million smokers (soccer), handball, and both codes of rugby may die prematurely due to smoking with an football, while in UK the most popular sport average loss of eight years of life.16 Further, is cricket. The sports in France require more about 17 percent of pneumonia deaths were activity and alacrity, which can explain the caused by smoking in 2002.17 In France, it is more healthy population in France. estimated that 66,000 people die as a result of smoking, and 90 percent of lung cancers Education and Employment are caused by smoking.18 With an increase in age Britons’ par- Alcohol and Pneumonia ticipation in full-time education or training declines, and for young Britons the fi gure is In both France and the UK, alcohol con- lowest for the entire European Union.22 sumption is among the highest in world, but The male unemployment rate in the UK with a different trend in the past years. The is close to the European average of about 25 UK has been a relatively moderate consumer percent, and the female rate of unemploy- compared to other Western European coun- ment, at fewer than 10 percent is the best in tries previously. France had one of the high- the EU. Further, young British males work est known levels of alcohol consumption in longer hours as compared to their European the world. But, in the last decades, alcohol counterparts. consumption has fallen in France, while it continued to rise in the UK. Yet alcohol consumption is still higher in France than the 4.3. Physical Environment UK. Recently it is reported that almost 11 percent of the French population is exces- The physical environment of the two coun- sive or dependent drinkers.19 Young popu- tries differs greatly. The UK is a lowland ter- lation in the UK have higher percentages rain with some mountains in the Northwest Determinants of Differentials in Pneumonia Mortality in the UK and France 107

while France is ringed with mountains. The nity. A referral is required from these GPs, UK has a temperate climate and temperature who work as the gatekeepers, to consult with ranges from −4 degrees Celsius to 32 degrees a specialist. In contrast to the UK, in France Celsius. In France, the average tempera- self-employed physicians along with salaried ture lies between 7 and 23 degrees Celsius. staff in hospitals provide primary and sec- Similarly, in the UK the atmosphere is more ondary health care, and patients have a free humid than France. The difference is greater choice of physicians. Despite all efforts, at night and in winter. In the UK rainfall is including monetary incentives, the newly much higher in the winter as compared to introduced gatekeeping system did not show France. Average daily sunlight is also quite any success.26 low in the UK as compared to France.23 So In terms of the number of hospital beds, from the above information it is clear that the France provides better facilities as compared UK has more rainfall and more humidity as to the UK. In 2002, in France there were 4.2 compared to France. And it is suggested that acute hospital beds per 1,000 population as Pneumonia causing microorganisms favor compared to the UK, which had 3.9 beds in this humid conditions. about 240 private acute hospitals, which is There are considerable environmental less than 5 percent of total beds. On the other differences between the UK and France. In hand, France has public (covering two-third December 1991, for example, London expe- of all inpatients beds) and private profi t and rienced a four-day air pollution episode. nonprofi t hospitals. Private for-profi t hospi- During the episode, deaths were 23 percent tals only entertain minor injuries, whereas higher than expected, caused by respiratory public and private nonprofi t hospitals focus infections. Consultations with general prac- on all other treatments, including emergen- titioners rose by 10 percent for upper res- cies, rehabilitation, long-term care, and psy- piratory and 4 percent for lower respiratory chiatric treatment.27 conditions, and hospital admissions of peo- In France there are about 1.6 million health ple older than 65 years rose by 19 percent.24 care professionals who constitute 6.2 percent In the UK national emission of nitrous oxide of the total working population. In 2001, is 130,391 tons as compared to 240,608 tons France had 3.3 physicians and 6.9 nurses per in France in 2003. If we look at the emission 1,000 population as compared to 0.6 GPs per of nitrous oxide from the energy industry, it 1,000 population in the UK, which faces a is 8,937 tons in the UK as compared to 3,290 shortage of skilled staff in the NHS.28 tons in France. Similarly if we look at the Furthermore, both countries have some pollution by ozone, it is much lower in the geographical disparities regarding the organ- UK (983) as compared to France (4,313).25 izational structures of the health service providers. In the UK, for example, public health 4.4. Health Care Facilities staff may serve at central, regional, strategic health authorities, and the primary care trusts. In the UK, general practitioners (GPs) in Similarly, there are geographical disparities groups of an average of three GPs provide in France in the distribution of physicians in primary care. In 2002, a GP had to take care favor of urban as compared to rural areas. of about 1,800 members of the local commu- France’s health system is institutionally 108 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

complex, which causes tensions between pneumonia was mentioned on 86 percent of various components of the system.29 the death certifi cates of patients diagnosed The French health care system was ranked with it during hospitalization; only 38 per- at the top by the World Health Organization cent mentioned it as the underlying cause of in 2000 with the population entirely free death.33 The main reason for this difference from chronic diseases.30 is misdiagnosis of cause of death mainly on Part I of the certifi cate, but it is improved Less Funding for Research in the UK after the subsequent International Classifi - In the UK, despite the obvious severity of cation of Diseases (ICD) coding. Second, lung disease, only 3.8 percent of all money the category of respiratory diseases such as spent by the Medical Research Council on pneumonia is certainly one of the disease medical research is spent on respiratory dis- groups most diffi cult to certify. ease.31 As a result there is a lack of funds to Biases potentially affect all causes of death, support a great deal of excellent research. and this can be particularly serious for diseases It seems that in the UK the current health of the respiratory system, which are subject debate is not about “evidence” but about the to greater variability in certifi cation.34 These political decisions. biases have already been recognized by some studies that outline differences in the UK and 4.5 Data Quality France (in the UK, the rates of chronic lung diseases are particularly high, while in France Data on cause-specifi c mortality rates there is a high frequency of nonspecifi c res- greatly depends on the reliability of the certi- piratory diseases). So it will be diffi cult to fi cation process of deaths. There is evidence compare specifi c pneumonia disease rates in that the differences observed in chronic Europe on the basis of data routines.35 pneumonia between the UK and France depend less on the differences in prevalence between the two countries than on the dif- 5. Conclusion ferences in certifi cation practices. Some of these differences are due to variations in In the foregoing sections, an attempt was the ways doctors certify deaths, others are made to explain the differences in pneumo- due to the way certifi cates are coded in each nia mortality between the UK and France. country.32 Some of the reasons are due to We found that pneumonia, a disease caused co-morbidity. When the cause of death is by an infection of the lung tissue by bacte- mentioned, chronic diseases like chronic ria and virus, is more prevalent in the UK obstructive pulmonary disease (COPD) are than France. Age-adjusted mortality rates mentioned more often than the acute one from pneumonia were estimated at 33 deaths (pneumonia). Similarly, in death certifi cates per 100,000 people in the UK and 9 deaths it is diffi cult to identify all cases of pneumo- per 100,000 people in France in 2004.36 To nia as the sole underlying cause of death. In explain this difference, we used data from a longitudinal study performed one month Eurostat and secondary literature from dif- after hospitalization in six districts in the ferent sources. We reviewed literature which Oxford Regional Health Authority area, explained that some groups of people may Determinants of Differentials in Pneumonia Mortality in the UK and France 109

be more susceptible to pneumonia morbid- differ signifi cantly in their population com- ity and mortality than others, including position. Smoking and alcohol consumption demographic factors, health-related behav- are probably important factors to explain iors, physical and social environments, and differences in pneumonia mortality between access and quality of health care. the two populations. However, it was important to fi nd out Differences in sports behavior, education, whether the differences in pneumonia mor- and employment between the two countries tality between the two countries were real also favor France, although the wet and differences or the result of data quality and humid climate of the UK provides a suitable comparability problems. According to a condition for infectious agents. study by Eurostat, the largest variation for More important are the differences in the pneumonia mortality between France and health care system of the two populations the UK is attributed more to differences in in which France was found to have a better death certifi cation of the disease between the health care system than the UK. In 2002, the two countries than real differences in preva- UK had 3.9 hospital beds per 1,000 popu- lence. The certifi cation process has led to lation while that of France was 8.4 hospital a reduction in the reliability percentages in beds per 1,000 inhabitants. At the same time, France in particular. there were 0.6 GPs per 1,000 population in Apart from the data problem, little differ- the UK, while there were 3.3 physicians per ences were found with regard to the deter- 1,000 population in France in 2002. This minants of pneumonia between the two would mean that there is less access to sec- countries. With regard to demographic fac- ondary and tertiary care in the UK than in tors, for instance, although it was observed France. Similarly, while Pneumonia is more that the very old and the very young popu- prevalent in the UK, it was also observed that lation are at particular risk from pneumo- there was less attention given to research on nia, possibly due to their weakened (low) respiratory diseases and primary care such defense system, the UK and France do not as childhood immunization in the UK.

REFERENCES

1. NHS Direct (2007), pneumonia, http://www. 5. National Library for Medicine, 2013, http:// nhsdirect.nhs.uk/articles/article.aspx, last vis- www.nlm.nih.gov/medlineplus/druginfo/ ited on June 1 2007. meds/a607015.html. 2. http://epp.eurostat.ec.europa.eu/statistics_ 6. Finch, R, “Community acquired Pneumonia; explained/index.php/Causes_of_death_statis- the evolving challenge,” Clinical Microbiol- tics. ogy Infection, 2001(7):30–38 (2001). 3. Mustafa MS, & SV Murthy, “Alcohol- 7. “Atlas on mortality in the European Union,” associated admissions to an adult intensive Health Statistics, Respiratory Diseases care unit: an audit,” European Journal of (2004). Anesthesiology, 2002 (19): 193–196, Cam- 8. Woodhead, M, Mamun M, Rogers T, and bridge University Press (2002). Wort, JS, “Community acquired pneumonia 4. NHS (2013), http://www.nhs.uk/Conditions/ in elderly people,” British Medical Journal, Pneumonia/Pages/Complications.aspx. 1998 (317): 1524 (1998). 110 JOURNAL OF HEALTH CARE FINANCE/Spring 2014

9. Burns DN, Hillman D, Neaton, JD, et al. nhs.uk/systemsandservices/etd,last visited on Cigarette smoking, bacterial pneumonia, and June 2, 2007. other clinical outcomes in HIV-1 infection: 23. http://www.pettortoise.co.uk/tortoise_cli- Terry Beirn Community Programs for Clini- mate.php. cal Research on AIDS. J Acquir Immune Defi c 24. Macfarlane, A, et al., “Air pollution, climate Syndr Hum Retrovirol 1996, 13:374 –383 and health: short-term effects and long-term 10. Doll R, et al., “Mortality in relation to smok- prospects,” Charlton J, Murphy M, eds., The ing: 40 years’ observations on male British Health of Ault Britain, 1841–1994, Vol.1, Lon- doctors,” British Medical Journal, 1994, 309: don, The Stationery Offi ce, 1997; pp. 187– 901–11 (1994). 204 (1997). 11. Doll, R, Peto, R, Boreham, J, & Sutherland, 25. Supra, n.2. I, (2004). Mortality in relation to smoking: 26. Supra, n.25. 50 years’ observations on male British doc- 27. http://www.euro.who.int/document/ tors. Bmj, 328 (7455), 1519.12 Supra, n.3. e88530.pdf?. 13. Pistelli, R, et al., “Determinants of Prognosis of 28. http://www.euro.who.int/document/ COPD in the elderly: Mucus Hypersecretion, e88530.pdf?. Infections, Cardiovascular Co-morbidity,” Euro- 29. Grosse-Tebbe, S & Figuera, J, Snap- pean Respiratory Journal 21(40): 10–14 (2003). shots of Heath Systems (2005), http:// 14. http://epp.eurostat.ec.europa.eu/statis- www.euro.who.int/__data/assets/pdf_ tics_explained/index.php?title=File:Causes_ fi le/0010/110242/E87303.pdf. of_death_-_standardised_death_rate,_2010_ 30. http://www.who.int/whr/2000/media_cen- (1)_(per_100_000_inhabitants).png&fi letime tre/press_release/en/. stamp=20121022145128. 31. British Lung Foundation (2007), http://www. 15. Hoare, Z & Lim, WS, “Pneumonia: Update on britishlungfoundation.org.uk/documents/ Diagnosis and Management,” British Medical BLF_Research.pdf. Journal, 2006 (332): 1077–1079 (2006). 32. Kelson, MC, and Heller, RF, “The Effect of 16. Peto, R, et al., Mortality from smoking in Death Certifi cation and Coding Practices on developed countries 1950–2000 (2nd edi- Observed Differences in Respiratory Disease tion), Oxford University Press, Oxford (2000). Mortality in 8 EEC Countries,” Revue Epide- 17. “Mortality statistics,” Offi ce for National miologie et de Sante Publique, 31(4): 42332 Statistics, 2002; General Register Offi ce for (1983). Scotland, 2002; Registrar General Northern 33. Goldacre, MJ, “Cause-specifi c mortality: Ireland, Annual Report, Statistics & Research understanding uncertain tips of the disease Agency, 2002 (2002). iceberg,” Journal of Epidemiology and Com- 18. Medical News, available at http://www.med- munity Health, 47:491–96 (1993). icalnewstoday.com/medicalnews.php?. 34. Hatton, F, et al. “Bridge coding between 19. http://www.ias.org.uk/What-we-do/Publi- ICD-9 and ICD-10 on 5,179 cases.” Meeting cation-archive/Alcohol-Alert/Issue-2-1999/ of Heads of WHO Collaborating Centres for Health-benefi ts-of-alcohol-debunked.aspx. the Classifi cation of Diseases, Tokyo, Japan 20. http://www.who.int/substance_abuse/publi- (1996). cations/alcohol_gender_drinking_problems. 35. Rican, F, “Respiratory Diseases,” Compa- pdf. rability and Quality Improvement of Euro- 21. Institute of Alcohol Studies, Fact Sheet 2006. pean Cause of Death Statistics, pp. 164–183 22. NHS, Education, Training and Development (2001). (ETD) (2007); http://www.connectingforhealth. 36. Supra, n.2.