Read the 2002 article Integrated Delivery Networks: A Detour on the Road to Integrated Care by Lawton Bruns and Mark Pauly in Health Affairs, volume 21, issue 4. Compare and contrast the differences between vertical and horizontal integration. Explain the models of integrated healthcare. Indicate the authors findings regarding vertical integration in healthcare.

During the 1990s many hospitals pursued twin strategies of vertical and horizontal integration. Each type of integration assumed multiple forms. Vertical combinations included acquisition of primary care physicians (PCPs), strategic alliances with physicians in physician-hospital organizations (PHOs)and management services organizations (MSOs), and the development of health maintenance organizations (HMOs). Horizontal combinations included the formation of multihospital systems, mergers, and strategic alliances with neighboring hospitals to form local networks. All of these combinations are collectively referred to as integrated delivery networks, or IDNs. While the forms of integration varied across hospitals and markets, their economic performance, after a decade of experience, was generally uniform: Nothing worked.1 This paper briefly summarizes the evidence for that conclusion from the late 1980s through the 1990s and then asks, What went wrong, and what prospects are there for integrated delivery that we can learn from these lessons? Provider Versus Theoretical Rationales For Integration Vertical integration. Proponents of vertical integration between hospitals and other health care providers and payers mentioned several goals underlying these efforts. These objectives reflected a range of efficiency goals (manage global capita128 July/Aug ust 2002 Business Of Health 2002 Project HOPEThe People-to-People Health Foundation, Inc. Rob Burns is the James Joo-Jin Kim Professor and professor of health care systems and management at the Wharton School, University of Pennsylvania, in Philadelphia. Mark Pauly is the Bendheim Professor and professor of health care systems, insurance and risk management, and public policy and management at Wharton. Downloaded from HealthAffairs.org on July 29, 2020. Copyright Project HOPEThe People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org. tion, form large patient and provider pools to diversify risk, reduce cost of payer contracting), access goals (offer a seamless continuum of care, respond to state legislation), and quality goals (assume responsibility for health status of local population). Of course, hospitals pursued private agendas in their vertical combinations, such as purchasing PCP practices out of fear of losing their referral bases or the mistaken belief that they could control the referrals. Some large physician groups engineered mergers with closely affiliated hospitals to tap the latters cash reserves. Finally, some hospitals held the mistaken view that economies of scale and scope resulted from operating a common infrastructure over their diverse provider sites. These hospital rationales only partially align with the academic view of vertical integration.2 The major overlap concerns protecting access to needed inputs (such as referrals). However, the major theoretical rationale for vertical integration rests on gains from improved coordination among trading partners (such as lower costs of contracting and monitoring)that exceed the forgone gains from economies of specialized production. Providers mentioned such considerations rarely, if ever. Moreover, contrary to many providers belief, there are no general results from economic theory that vertical integration leads to greater efficiency or market power on the part of the firm.3 Horizontal integration. Hospitals rationales for horizontal combinations likewise reflected a mixture of efficiency goals (prepare for capitation, reduce excess capacity, strengthen financial position)and access goals (expand the delivery network). In fact, their objective functions for the two broad strategies often overlapped, which suggests that hospitals did not really understand the difference between the two. A major stated objective in horizontal integration was achieving economies of scale. Such economies were believed to flow from several sources, including large patient volumes, sharing of equipment and services, and group purchasing. There were also a host of private motives behind these combinations. These included the need to prepare for the Clinton health plan and the presumed need to be one of the surviving accountable health plans in the local market; respond to stories of investor-owned chains that buy up local hospitals, rationalize their capacity, and out-compete freestanding community hospitals for managed care contracts based on lower costs; and exert greater bargaining power over managed care payers and resist their price-discounting pressures. With regard to this last motive, some local hospital systems may have been pursuing anticompetitive strategies under the banner of improved efficiency and access.4 Reports issued by hospital associations that endorsed IDNs provided legitimacy and political cover for what these systems were doing.5 As with vertical combinations, the rationales for horizontal integration espoused by providers only loosely conformed to the benefits of merger identified in the economics and strategy literatures.6 Many of the finance-related benefits of merger accrue to for-profit firms because they obtain capital in the equity markets, so the benefits do not apply to nonprofit hospitals. Other benefits of horiIntegrated Networks HEALTH AFFAIRS ~ Vo l u m e 2 1 , N u m b e r 4 129 Downloaded from HealthAffairs.org on July 29, 2020. Copyright Project HOPEThe People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org. zontal integration include the acquiring firms belief that it can better use the assets of the target firm, the desire to enter new markets without expanding existing capacity or to speed up entry, to achieve synergies through enhanced performance of the combined firm, to glean economies of scale, and to increase market power by eliminating competitors. As with vertical integration, only the latter two motives parallel providers stated rationales. Economists note that mergers are motivated by both efficiency and bargaining (market power)considerations.7 Performance Of IDNs Vertical integration. Physician practice acquisitions. Hospitals acquired PCP practices by purchasing the practice assets and then placing the physicians on salary. In so doing, hospitals suffered heavy financial losses (because of high acquisition prices, adverse PCP selection, insufficient practice cash flows, lack of productivity and at-risk compensation incentives, and other factors), failed to garner more managed care contracts and covered lives, and failed to greatly increase physicians alignment with their organizations.8 The accumulated losses and the failure of full-risk capitated contracting to develop in local markets led many hospitals to downsize their PCP networks before such benefits could be demonstrated and achieved.9 Physician-hospital strategic alliances. Hospital alliances with physicians took a number of forms, such as PHOs and MSOs, in the 1990s. The proportion of hospitals with these types of alliances peaked in 1996 and has declined ever since. By 2000 only 26 percent of hospitals had a PHO (down from 33 percent in 1996), and less than 13 percent had an MSO (down from 22 percent). The alliances decline was fueled by their lack of managed care infrastructure and failure to attract covered lives in risk contracts and by their failure to improve physician-hospital collaboration and hospital financial performance (costs per day or discharge).10More generally, evidence from the 1980s suggests that affiliating or linking outpatient care with a large and complex inpatient institution tends to raise the marginal and average cost of both outpatient and inpatient care.11 While there surely are economies of scope for some medical services (it would not be efficient to have hospitals each specializing in one diagnosis-related group, or DRG), it is clear that there are limitations on the production side to efficient combinations. Hospital-sponsored HMOs. Hospitals also integrated vertically into the insurance market by establishing their own HMOs. Hospitals experienced widespread failures with insurance products as a result of low capitalization, huge medical loss ratios, conflicting capital needs between hospitals and health plans, lack of expertise in actuarial science (poor pricing of risk)and marketing, and entry into competitive health insurance markets.12 Recently, hospital systems and networks have begun to abandon their PPO and HMO products.13 The handful of successful provider launches of HMOs (Carle Clinic, Marshfield Clinic, Geisinger, Scott and White)may be attributed to some unique advantages 130 July/Aug ust 2002 Business Of Health Downloaded from HealthAffairs.org on July 29, 2020. Copyright Project HOPEThe People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org. that are not easily imitated by others. These include the anchoring of the IDN around a core multispecialty group established seventy or eighty years ago, giving ample time to develop a coherent medical culture; the IDNs location in a rural area that buffers it against competition and market entry by commercial plans; and early development of the health plan during the 1970s that provided an accumulation of managed care experience.14 A related set of advantages account for the success of Kaiser and the Mayo Clinic, which other IDNs have failed to imitate.15 Horizontal integration. The three types of horizontal integration can be distinguished as follows. Multihospital systems feature common asset ownership but separate system versus hospital boards and executives. Mergers feature common asset ownership and consolidated governance and executive functions. Networks maintain the hospitals separate asset ownership and governance structure. Multihospital systems. The literature on multihospital system performance during the 1980s reported little evidence of economies of scale and none for improvements in cost per admission, profitability, service provision to the community, charity care, or patient outcomes.16 Subsequent analyses have generally found that hospital systems do not improve production efficiency (hospital expenses or cost per admission)or market share (inpatient days), may actually incur higher administrative costs, but may achieve some marketing benefits and increased revenues.17 A recent study of organized delivery systems that reported a positive association between integration and financial performance was based on only eight systems arrayed in a scatter plot (with two important outliers).18 Despite including some of the vanguard nonprofit hospital systems undertaking integrated delivery, the investigators concluded that they were still loosely coupled systems in a nascent stage of development. Hospital systems are trending from tightly to loosely coupled structures. Between 1994 and 1998 there has been a decrease in systems organized centrally (from 16 percent to 8 percent), an increase in the proportion that are moderately centralized (from 25 percent to 42 percent), and a decrease in the degree of centralization of five of six products and services studied. Researchers suggest that there are diseconomies from overcentralization of hospital systems and a need to blend both centralized and decentralized forms of management in such systems.19 It is instructive to consider some of the spectacular horizontal successes and failures. Systems such as Partners Health Care, University of Pittsburgh Medical Center (UPMC), and Sutter Health have been successful in extracting greater levels of reimbursement from local payers by combining prestigious flagship and large numbers of high-quality community hospitals.20 On the other hand, several systems have filed for bankruptcy (such as Allegheny Health Education and Research Foundation)or disbanded, while others have experienced double-notch bond downgrades and declines in utilization.21 These results may be due as much to poor management and market instability as to system membership, however.22 Hospital mergers. As was the case with vertical integration, hospital merger activIntegrated Networks HEALTH AFFAIRS ~ Vo l u m e 2 1 , N u m b e r 4 131 Downloaded from HealthAffairs.org on July 29, 2020. Copyright Project HOPEThe People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org. ity peaked in 1996 and has steadily declined ever since.23This decline probably reflects the dwindling number of acquisition targets in local markets and the mounting evidence regarding lackluster postmerger performance in both health care and other industries.24 A major finding from this literature is that economies of scale do not flow automatically from hospital size and merger. Indeed, research on the history of large corporations suggests that economies are achieved by driving higher volumes of output at a faster speed through physically consolidated capacity.25 Such conditions are rarely met in hospital mergers. Early research on the economies associated with larger hospital size found inconsistent evidence. Recent studies have found that economies are achieved at low rather than high levels of hospital scale, with more modest cost savings.26 A recent analysis of academic medical center (AMC)mergers reported savings of only 12 percent.27 There is evidence suggesting that efficiencies flow from clinical consolidation in merging hospitals.28 However, recent market studies of hospital mergers have noted the extreme geographic and political hurdles to getting clinical departments to consolidate.29 Several studies also have found anticompetitive effects resulting from hospital mergers seeking to eliminate acute care competitors and exercise market power over payers.30 In a set of longitudinal studies and simulations, researchers found that hospitals raise prices in response to mergers, price increases grow over time, and prices rise among both merging hospitals and nonmerging hospitals in the market.31 Even studies that find beneficial effects of mergers on price competition report higher prices following mergers in more concentrated hospital markets.32 Hospital networks and alliances. Recent studies question the efficiencies of local hospital networks. Network membership may fail to improve cash flow per bed and operating expense per adjusted discharge and may fail to compete with system membership in terms of hospital costs and margins.33 As with hospital systems, the proportion of centralized networks plummeted between 1994 and 1998 (from 18 percent to 1 percent), while moderately centralized and independent networks spread (from 52 percent to 61 percent, and 27 percent to 34 percent, respectively). The degree of centralization of six products and services also dropped. These figures are important because less centralized networks have higher costs.3