kaiserpapers.com/businesspractices

The Rise and Fall of a Kaiser Permanente Expansion Region

DANIEL P. GITTERMAN, BRYAN J. WEINER, MARISA ELENA DOMINO, AARON N. McKETHAN, and ALAIN C. ENTHOVEN 

University of North Carolina, Chapel Hill; Stanford University Prepaid group practices (PGPs) — multispecialty groups that vertically integrate the organization, financing, and delivery of health services to a specific population—were once viewed as the most cost-effective and efficient model for achieving national health care reform (e.g., McNeil and Schlenker 1981; Saward and Greenlick 1981). 

Policy reformers who extolled the benefits of health maintenance organizations (HMOs) in the late 1970s and early 1980s emphasized in particular the cost and quality advantages of PGPs vis- `a-vis solo and single-specialty fee-for-service (FFS) providers. A comprehensive review of comparative empirical studies (HMOs versus FFS) since 1950 concluded that the total costs for HMO enrollees were 10 to 40 percent lower than those for comparable enrollees with conventional indemnity insurance (Luft 1978). Although PGPs did not originate as a competitive response to fee-for-service indemnity healthinsurance, many proponents viewed them as a promising means of helping contain rising medical costs, encouraging a more rational allocation of health care resources, and improving the access to and delivery of quality services (McNeil and Schlenker 1981). Enthoven (1978a and b), for example, argued that PGPs would promote competition within the health care financing and delivery system by stimulating conventional providers to restructure medical practice and insurance benefits, acting as a catalyst for overall cost containment. 

The Milbank Quarterly, Vol. 81, No. 4, 2003 c 2003 Milbank Memorial Fund. Published by Blackwell Publishing, 350 Main Street, Malden, MA 02148, USA, and PO Box 1354, 9600 Garsington Road, Oxford OX4 2DQ, UK. Daniel P. Gitterman et al. 

PGPs were a central pillar in the “managed competition” reform proposal, in which employers would offer employees a choice of plans (perhaps two or three HMOs in the same geographic area, to allow competition) and provide a defined employer contribution, with employees responsible for the premium differential as part of their health benefit packages (Enthoven 1993; Moran 1981). 

Policy experts theorized, for example, that multispecialty groups could achieve greater economies of scale than other provider types could, by buying supplies and equipment in volume, spreading out the risk that accompanies capitated payments, gaining access to financial capital at lower interest rates, achieving a prominent brand name in the community, and attracting experienced physician managers. In their operation, PGPs in particular could offer more efficient clinical care by combining the services of primary care physicians, specialists, and nonphysician providers; by avoiding an undercapacity in primary care practitioners and an overcapacity in specialists; and by retaining clinical responsibility for their patients in both outpatient and inpatient settings (Robinson 2004). 

Despite these supposed advantages, however, the penetration and performance of prepaid group practice have fallen far short of its key proponents’ expectations. Today, with the upward trend in health premiums, growing disil- lusionment with the various forms of highly “managed” care, and the rising number of uninsured people, another round of debate over national health reform seems likely. It is timely, therefore, to reflect on the potential, performance, and prospects of prepaid group practice as a system delivery model for the future.

To do this, we examine the rise and fall of the Kaiser Permanente (KP) expansion effort in North Carolina in order to gain insight into making prepaid group practice work. We trace and analyze the key events in the North Carolina KP’s entry and start-up, its performance and growth over time, and its exit. We use interview data collected from former and current KP national, regional, and local managers; North Carolina public officials; state and national health care experts; and supplemental analyses of enrollment, financial, and other secondary data. We conclude that KP’s failed North Carolina expansion resulted not from an inherent flaw in the PGP model but from a complex interaction of political, economic, and organizational factors. Finally, we offer some policy recommendations for PGPs’ role in future reform efforts. We believe that PGPs should still remain of special interest to those reformers who see the health system changing through marketplace competition. 

The Origins of Kaiser Permanente s Expansion into North Carolina Kaiser Permanente (KP) began on the West Coast as a company-funded and company-managed means of providing medical care services to workers in Henry J. Kaiser s industrial enterprises. Its corporate headquarters began and remain in Oakland, California. 

By 1955, KP had a major presence in three regions (Northern California, Southern California, and Oregon), with a growing network of hospitals and clinics and a combined membership of 500,000. 

KP expanded to Hawaii in 1958 and a decade later established regions in Colorado (1969) and Ohio (1969). After the passage of the Health Maintenance Organization Act in 1973, the KP plans in all six regions became federally qualified HMOs. 

By the early 1980s, KP had additional regions in Texas (1979), the greater Washington, D.C., etropolitan area(1980), and greater Hartford, Connecticut (1982). 

The plan in Texas began as a 50–50 joint venture of KP and Prudential, and in 1983 KP took over Prudential’s interest. 

KP tried (and failed) to acquire a plan in Chicago but eventually bought an existing plan in Kansas City. 

Whereas our interview participants described KP’s earlier expansions as “opportunistic,” KP’s expansion into the Southeast appeared to be more purposeful. KP wanted a national presence in order to compete more effectively for national corporate accounts and to position itself more favorably in the event of national health care reform. 

Accordingly, it chose to expand in regions where it did not have a presence, namely, the South. KP conducted comprehensive analyses of several markets in the South and Midwest and narrowed its final expansion choices to Raleigh-Durham, NorthCarolina, and Atlanta, Georgia.  Because half the organization’s joint national health plan/physician leadership committee preferred one market, and the other half preferred the other market, KP decided to enter both markets (KP Institute 2002). 

At the same time that KP was trying to expand regionally, elected officials and policy reformers in North Carolina were trying to make the state an attractive operating environment for healthcare delivery systems that were not fee-for-service plans. Impressed by the cost-containment achievements of PGPs such as Kaiser Permanente, the Group Health Cooperative of Puget Sound, and the Health Insurance Plan of NewYork, Congress had already passed the Health Maintenance Organization Act in 1973. 

The act defined HMOs, provided grants and loans for the start-ups of nonprofit HMOs, and required all employers of 25 or more employees to offer at least one PGP and one Independent Practice Association – based HMO as health insurance options wherever they were available and desired. This requirement was a big boost to the development of HMOs, including PGPs, and a spark for state-level reforms in the late 1970s and early 1980s. 

Faced with rapidly escalating health care costs, the North Carolina General Assembly, with the support of Governor James Hunt, created the Commission on Prepaid Health Plans to study alternatives for organizing and financing health care (North Carolina Commission 1979a and b). The commission examined several models of prepaid health care delivery; took testimony from experts such as Paul Ellwood, who was credited with inventing the concept of HMOs; and concluded that delivery system reforms could introduce effective competition and control costs within the traditional health care system. 

The General Assembly then enacted several legislative reforms to create a more supportive regulatory environment for alternative delivery systems. These reforms enabled prepaid health plans to acquire facilities; make loans to contracting medical groups; contract to provide health care services; contract to provide marketing, enrollment, and administrative services; contract with traditional insurers for insurance, indemnity, or reimbursement for costs of health services; and offer and contract for additional health services (North Carolina Commission 1979 a and b). 

In 1982, Governor Hunt established the Foundation for Prepaid Health Plans to attract new HMOs to North Carolina. “Kaiser Permanente was clearly the dominant HMO in the country and had a good reputation,” the foundation’s senior vice president noted in an interview, “so it became the initial target of our [recruitment] efforts” (personal communication, May 2002). 

The foundation recruited KP in much the same way as the state recruited large employers, offering KP information about demographics, employers, and health care utilization in the state’s major metropolitan areas; and facilitating contacts for KP in the employer and provider communities. This aggressive recruitment strategy coincided nicely with KP’s interest in expanding into the Southeast. Once KP entered the state as KP–Carolina, it sought state employees as its new enrollees. The existing State Health Plan (SHP) offered a conventional self-insured indemnity health plan—the North Carolina Teachers’ and State Employees’ Comprehensive Major Medical Plan (CMMP)—with an administrative services–only arrangement with Blue Cross and Blue Shield of North Carolina. 

With an effort by the General Assembly in 1985 to encourage more “choices and competition” among the plans, for the first time the SHP’s annual enrollment included HMOs in the 1986 contract year. 

KP–Carolina was one of the first alternative plan choices offered, along with Blue Cross Blue Shield (Personal Care Plan) and Prudential (PruCare). By the mid-1990s, seven options were offered in addition to the CMMP. 

A Brief History of KP–Carolina’s Operations and Performance The Kaiser Foundation Health Plan of North Carolina (KP–Carolina) was incorporated on May 1, 1984. Licensed as a not-for-profit 502(c)(3) organization and run by a board of directors in North Carolina as a subsidiary of the national Kaiser Foundation Health Plan, it began operating in January 1985 in Raleigh, Durham, and Chapel Hill, a region of the state known as the Research Triangle. 

At its inception, KP–Carolina operated as a group-model HMO, and it remained in that form for most of its tenure. In the standard group-model HMO, a health plan contracts with physicians who are organized as a partnership, professional corporation, or other association. 

Unlike the Independent Practice Association (IPA) model, in which physicians may participate in numerous competing networks and still bill on a fee-for-service (FFS) basis, the physicians in the new Carolina Permanente Medical Group (CPMG) served exclusively KP patients. 

The health plan compensates this medical group for contracted services at a negotiated rate, and the group is responsible for compensating its physicians and contracting with hospitals to care for its patients (UnitedHealthcare 1994). 

KP–Carolina was the marketing entity that attracted and encouraged employer groups to offer the choice of enrollment to their employees and contracted with physicians and health facilities. It contracted with the CPMG, a for-profit professional corporation of physicians who staffed KP outpatient clinics and exercised full responsibility for medical services provided to enrollees. The CPMG hired new physicians to provide primary (and some specialty) care as salaried employees for an initial three-year contract (after which time they were eligible to become shareholders), and KP–Carolina reimbursed the CPMG a fixed per capita rate for all physicians’ services (primary and specialty care) based on the number of members enrolled at the beginning of each  month (fixed per member, per month). 

KP transferred physicians from other regions to serve as middle and upper managers. All primary care physicians were part of the CPMG. Some outside specialists with whom the CPMG contracted were also paid on a capitated basis, but many received fee-for-service or discounted FFS payments. 

As the enrollment base grew, the CPMG was able to directly offer more specialty care. The CPMG was financially liable for emergency department visits, physicians’ services during hospital admissions, inpatient evaluations, and the management of hospital care. When KP first entered the market, it intended to target only the Research Triangle area. Over time, however, KP–Carolina’s service area grew to encompass 22 of the state’s 100 counties, including the Charlotte metropolitan area. 

Its provider network included more than 200 physicians, nurses, and physician extenders in nine medical offices, plus more than 400 contracting community providers (Silberman 1996). In contrast to its California regions, KP never owned any hospitals in North Carolina. Instead, it contracted with hospitals on a per diem or case rate basis for regular inpatient hospitalizations and on a discounted FFS basis for outpatient hospital services (Silberman 1996). 

KP–Carolina and the CPMG had a mutual risk-sharing arrangement that allowed both parties to share in both the positive and negative financial results. In general, however, KP’s health plan absorbed any medical group losses that would have resulted in the medical groups’ insolvency. 

Individual physician shareholders were eligible for bonus payments when both the CPMG and the health plan had positive financial div results. KP–Carolina, one of the first HMOs to enter the state, introduced the first and only true prepaid group practice in North Carolina. Although other medical groups and IPAs relied on capitation revenue from HMOs, they did not exhibit the mutual exclusivity of the group or staff model. 

In the late 1980s and early 1990s, only one company applied to the North Carolina Department of Insurance to set up HMOs in the state. Then a flurry of HMO activity took place, dramatically altering North Carolina’s competitive health care landscape in the mid-to-late 1990s. Numerous network-model HMOs, with primary care contracts with individual physicians in the local communities, began to compete against KP–Carolina. For example, the Prudential Health Care System moved into the Research Triangle in 1993 with a network-model HMO product and attracted several large employers, including the SHP. 

An important factor in KP’s successful entry into other regions was the strong backing of influential local organizations. On the West Coast and in Colorado, Kaiser was strongly supported by the AFL-CIO, which liked its emphasis on comprehensive benefits and preventive medicine and demanded that employers offer Kaiser as an alternative to the traditional insurance.

Given this history, KP–Carolina targeted North Carolina public employees as the initial foundation of its enrollee base. KP–Carolina was one of the first alternative plans that the SHP offered, and it proved to be a popular choice. In fact, in 1993, KP–Carolina had the highest enrollment of any alternative plan to the SHP’s traditional indemnity option, with enrollment peaking at 26,928 SHP members. 

This constituted 22.9 percent of KP–Carolina’s total enrollment in 1993 (NCSHP 1993–1998). Although KP–Carolina depended heavily on the SHP for enrollment, it also was initially successful among private-sector employers within the Research Triangle market. Enrollment figures show that KP–Carolina steadily grew during the first years of its operation, attracting 28,328 members by the end of 1986 and reaching 56,140 enrollees by 1987 (NCDOI 2000). 

By the end of 1989, KP–Carolina had 97,464 enrollees, a number that rose to 117,811 by the end of 1993 and reaching a peak of 134,081 by the end of 1997. Then in the mid -1990s the rate of growth slowed substantially as competition from other health plans heated up in an increasingly crowded market. Even within the SHP, KP–Carolina lost members because of the intensifying competition. 

By 1996, KP–Carolina served slightly more than 20,000 SHP members, who constituted 15.8 percent of KP–Carolina’s total enrollment (NC- SHP 2000). By 1998, KP–Carolina served only 18,157 SHP members, or 15.3 percent of KP–Carolina’s total enrollment. 

Finally, in December 1999, KP–Carolina closed down and sold its North Carolina operations, leaving only 4,399 members who had opted to remain in the plan until their contracts expired (NCDOI 2000). 

KP–Carolina’s financial performance had three distinct phases: (1) the entry phase, including the accumulation of “start-up” losses; (2) the steady growth phase, including the achievement of profitability; and (3) the final phase of declining profits and eventual exit from the marketplace. KP–Carolina reported a profit in just four of its 14 years of operation in North Carolina (NCDOI 2000). In the start-up phase (1984 to 1991), KP–Carolina enrolled 115,196 members and incurred nearly $113 million in total debt, including the income (or loss) from administering the plan and the associated fixed costs. 

From 1992 until 1995, KP–Carolina was profitable. In 1992, KP–Carolina’s net income totaled approximately $6.4 million, with 118,790 mem- bers enrolled (NCDOI 2000). KP–Carolina’s annual net income peaked in 1993 at approximately $6.5 million before returning to net annual losses in 1996 totaling $15.9 million. In its profitable phase (1992 to 1995), KP–Carolina’s total net income was approximately $17 million. 

After this brief period of profitability, KP–Carolina’s financial perfor- mance worsened considerably. In the decline phase (1996 to 1999), its net losses totaled $125.8 million, and its enrollment peaked at slightly more than 134,000 members. In all, KP–Carolina incurred approximately $280 million in aggregate losses (NCDOI 2000). In 1999, KP–Carolina sold its employer-sponsored and Medicare membership in the Research Triangle market to Partners National Health Plans of North Carolina, and it sold its Charlotte membership to Principal Health Care of the Carolinas (now Coventry). 

After KP–Carolina’s Triangle operations were sold to Partners, the Triangle-area physicians affiliated with the Carolina Permanente Medical Group, which had an exclusive contract with KP–Carolina before its sale, continued as an independent physicians’ practice under a new name (Carolina Premier Medical Group). But in 2000, the Carolina Premier Medical Group filed for bankruptcy. 

Regulators, Purchasers, and Providers 

What were the barriers to building a viable prepaid group practice? In this section, we highlight the major political factors (as identified by the participants) that impeded KP’s effort to build a viable prepaid group practice in North Carolina: regulatory uncertainty, the politics of the SHP’s structure and operations, and resistance from provider groups to a prepaid group practice. 

Although some factors might seem unique to North Carolina, reports and evidence from the field suggest that similar barriers exist in other regions of the country. Regulatory Uncertainty after Market Entry In the mid-1980s, KP–Carolina, as one of the first HMOs and the first prepaid group practice in the market, faced a fluid and uncertain regulatory environment. In the mid-1980s, the North Carolina Depart ment of Insurance’s regulation of HMOs was limited to financial and solvency issues.

That changed, however, when employer purchasers, seeking assurance that the new, undercapitalized HMOs would be similar to indemnity insurance in stability and predictability, asked the department to set new requirements because there had been three HMO insolvencies from 1985 to 1987 in the market areas where KP had begun operating. 

At first, the department persuaded all the HMOs to open their enrollment to clients who had been left without health coverage by a bankruptcy. In 1987, the General Assembly defined prepaid health plans as risk-bearing entities, subjecting them to market conduct requirements and new minimum capital requirements. The 1987 Health Maintenance Organization Solvency Act created new financial reserves rules, insolvency protections, and cost and utilization review requirements. To protect enrollees, the department further increased solvency standards in 1989 and set more stringent standards for the operation of HMOs. 

One critical challenge (specific to prepaid group practice) was whether KP–Carolina would be regulated as an insurance entity or as a health care plan or provider. KP always had seen itself as basically a health care delivery system rather than a health insurance plan, and KP–Carolina argued that certain regulatory requirements did not “fit” the PGP model.

AccordingtoAllen Feezor, a former North Carolina deputy insurance commissioner, “The insolvencies of other HMOs clearly subjected KP–Carolina, without any doubt, to new jurisdiction under [the] DOI [Department of Insurance]. And [the] DOI, which knew how to measure financing in the classic indemnity kind, was not very adept about understanding laying-off risk and downstream risk to physician groups, which is the classic KP model” (personal communication, May 2002). 

Proponents of community-sponsored, prepaid direct-service plans contended that examining the stability of the medical practice would have been more important than imposing solvency requirements on the health plan. A key difference between KP and other HMOs in North Carolina was that KP was a true group practice and none of its competitors were. 

That fact presented challenges in regard to network adequacy and construction and the definition of service areas and ratings, with which the regulators were unfamiliar. KP’s specific challenges were administrative and compliance issues. Rules on how service areas were defined and how networks had to be built - such as the requirement that KP had to have certain types of providers within certain geographic areas—meant that KP could not effectively negotiate pricing with particular specialties and subspecialties. 

For example, KP had to contract with the specialists in Raleigh and could not send its patients outside that area, even though the specialists there demanded higher prices. Nor could they send Charlotte patients to the Research Triangle for transplants. Thus, despite the best efforts of the local sponsors and proponents of the PGP model to create a favorable regulatory environment before entering the market, KP–Carolina faced some unusual regulatory challenges during its start-up phase. The Politics of a Purchaser: The North Carolina State Health Plan KP achieved some of its greatest successes in other regional expansions by enrolling large numbers of members from either private-sector unions or public employee systems. 

Although North Carolina had few unions to spur enrollment, it did have a large public employee system. KP– Carolina believed that if given a choice of plans and an opportunity to pay lower premiums, state employees enrolled in the traditional indemnity plan within the State Health Plan would transfer to KP’s low-cost, comprehensive prepaid group-practice plan. 

In other public employee systems in which KP prospered, all competitors were required to offer similar, if not the same, benefits. The SHP’s underlying politics and structure, however, posed significant challenges to KP–Carolina’s ability to capitalize on a large and growing base of potential state employee enrollees. 

Under its enacting statute,all SHP benefits were enumerated inlaw, and any benefits offered under all prepaid hospital and benefit plans had to be at least comparable to those offered under the SHP’s indemnity plan. This provision limited KP’s ability to tailor its benefits design. Moreover, an implicit agreement existed between the General Assembly and state employees that employees would always have health coverage through the North Carolina Teachers’ and State Employees’ Comprehensive Major Medical Plan (CMMP) free of a premium contribution, although they would pay the full average cost for covering their spouse and dependents in the plan. 

In the 1990s, employees had to pay the price difference to enroll in health plans with higher total premiums than the CMMP, but they were not rewarded with savings by selecting health plans with lower total  premiums. 

In 1987, for example, KP had offered an employee-only policy for 9 percent less than the CMMP, but employees were not given any incentive to enroll in KP, despite the cost savings to the state. TheCMMP monthly premium that year was $93.82, and the KP–Carolina premium was $85.74. By 1992, the CMMP monthly premium was $144.60 and the KP–Carolina premium was $174.06, with the employee responsible for paying the difference of almost $30 per month. 

This worked against KP–Carolina and its traditionally lower-cost business model. Thus over time, KP’s early and overall penetration into the SHP was significantly lower than its penetration into other federal and state employee health programs (see figure 1). The SHP required all new health plan entrants to offer low co payments for office visits and drugs (at first free and then $5) even as the CMMP’s copayments, coinsurance, and deductibles increased over time. It took a couple of years of intense negotiating with the SHP before KP and other health plans could raise their office copayment from nothing to $5 per visit. The SHP also refused to allow KP and the other health plans to place any limits on mental health or substance abuse treatment, although theirs was a more generous benefit than that of the self-insured indemnity plan, despite clear evidence of adverse selection in similar circumstances (Frank et al. 1996). 

Consequently, employees who expected to use the health care services often were more likely to select KP’s more comprehensive services and lower cost-sharing (copayments, etc.) over those of the CMMP. By the late 1990s, the Fiscal Research Office of the North Carolina General Assembly, which maintained strict oversight (and, in some years, de facto control) of the SHP, was determined not to let HMOs enjoy what they perceived as favorable (rather than negative) risk selection. 

The SHP concluded that the people exiting the indemnity plan were younger, and thus presumably healthier, than those remaining enrolled. Accordingly, as a way of adjusting the risk, it levied a monthly surcharge of $10 per member on all HMO enrollees aged one to 40. The state’s risk adjustment was based on age, ensuring that the younger employees choosing HMOs would end up subsidizing older workers who stayed with the Comprehensive Major Medical Plan. According to Dr. Bill Gillespie, who served as the CPMG’s medical director in North Carolina, the legislature’s risk adjustment created a“perverse incentive”that significantly undercut KP’s strategy to be the low-cost option in the state health plan (personal communication, June 2002). 

As another KP regional manager concluded, “It was a political thing more than a medical thing” (UNC 2002). 1 

The KP national office presented the SHP with a five-factor method of risk adjustment, but the state officials preferred their own methodology. 

KP–Carolina opposed the surcharge, claiming that risk should not be computed solely as a function of age and that the SHP did not compensate KP for the older members that it attracted. 

KP–Carolina’s national perspective—that there should be no preexisting exclusion period of any sort and no lifetime cap on benefits—ended up hurting its bottom line. Most of the other HMO competitors had both an exclusion period and a lifetime cap on benefits, which meant that those people who had reached their lifetime maximum in another plan or who had serious medical conditions that could not wait out the usual six-month preex- isting condition period enrolled in KP. The age-only adjustment, on top of zero copayments and benefit mandates, reduced KP’s and other plans’ ability to offer competitive bids to SHP participants. A risk adjustment that also considered the patient’s sex and severity of illness would have enabled KP to receive more than the standard monthly payment when the employee’s health status was less favorable than average. 

The SHP used the risk adjustment funds, however, to create a surplus pool for the CMMP, but the HMOs did not have that luxury and eventually raised their prices to cover costs. 

According to Paul Sebo, the health plan program manager of the North Carolina SHP, the more mature HMOs were discriminated against because of strategies adopted by new market entrants. Newer companies such as Wellpath gained market share in the state health plan membership by initially offering their plans at prices significantly lower than KP–Carolina’s. 

Because younger and healthier employees were motivated primarily by price, these new options were able to attract better risks, leaving both KP and the older HMOs with disproportionately more bad risks (UNC 2002). 

Other SHP policy choices also exacerbated the adverse selection problem over time. 

Unlike some other public employee health programs, the State Health Plan pools retirees and active employees. The state (as the employer) pays the full premiums of active employees (who chose the CMMP) but does not contribute to their dependents’ cov- erage. These policies had at least two negative effects. First, the premiums for dependents in the SHP were very expensive compared with similar coverage outside the SHP because they were subsidizing the retirees’ higher costs. 

Second, this situation probably created an adverse  $0 $50 $100 $150 $200 $250 1987 1989 1991 1993 1995 1997 1999 KP-Carolina Average across all other health plans Comprehensive Major Medical Plan Premium Per Member Per Month Note: All premiums and rates are for non–Medicare employees only. “Average” represents the average of all non-Medicare employee-only premiums for all the North Carolina State Health Plan options available to members each year. fig. 1. Comparison of monthly health plan premium per member in North Carolina State Health Plan, 1987–1999. selection, since dependents would have a strong incentive to obtain coverage elsewhere and those who could not would likely include some members who were unable to do so because of their health status. 

As a result of these problems, by 1988 the premium for KP and other health plans was actually greater than that of the fully subsidized (100 percent employer paid) CMMP option in the SHP (see figure 1). Although KP initially entered the SHP as a plan option with a premium that undercut even the CMMP’s premium, the combination of low enrollment incentives to state employees and the enforced subsidization of the CMMP translated into higher premiums after the first year, further reducing enrollment incentives for state employees. Over time, as KP’s premiums rose, most of its members who left went back to the CMMP, which did not raise its base premiums at all from 1991 to1998.

A smaller number of members transferred to less costly competitor plans. During this period, KP’s premium, paid by the employees, was between 20 and 46 percent greater than the CMMP’s employer-paid premium. In 1999, the employer-paid CMMP premium (employee-only) was $187.98 per month, whereas KP’s was $220.46; the CMMP was the least expensive of the seven other health plan options offered that year.Thus, at some point, the cost differential became too high, first for the healthier members  and then even for the sicker members. 

All the “competing” health plans eventually withdrew from the SHP, but KP–Carolina was hurt the most, based on its historical reliance on public employees for its enrollments. Resistance from the Medical Community Cohesive physician organizations can find ways to inhibit the entry and growth of prepaid group practices. This can include informal means, such as the public characterization of PGPs as dispensing low-quality medicine. Physicians and hospitals in the traditional medical community in North Carolina were quite hostile to prepaid group practice. 

KP–Carolina did not adequately anticipate the magnitude of this resistance and so was slow to counter it. The North Carolina Medical Society also was generally opposed to KP–Carolina, and in fact, some specialist medical societies organized against prepaid health plans. The North Carolina Hospital Association and some of the larger hospital systems opposed “managedcare”as well. Several other professional organizations, such as the North Carolina Psychological Association, which objected to health plans’ utilization review of providers’ practices or that wanted to require employers to cover their services, were opposed to prepaid group practice. 

Many local medical providers perceived KP not as high-quality medicine but as a “doc-in-the-box” operation. Other comments by participants that were especially damaging to KP’s efforts to attract new enrollees were “Kaiser attracts only inept doctors,” “Kaiser is ‘cut-rate’ medicine,” and “Kaiser is a scheme to withhold medicine.” 

KP–Carolina may have been a more visible target than the other HMOs because it entered the market early and its form of managed care offered the most restricted choice of providers. As one former KP manager observed, “There was a lot of antipathy for anyone limiting doctors in terms of the numbers of prescriptions they write or procedures they ask to do, and the incentives that KP might have given to their physicians were probably overcome by the North Carolina environment and the antipathy” (UNC 2002). 

Challenges for the KP Model in the Market Marketing KP-Carolina 

The nature of the market itself may have contributed to KP’s failure to expand in this new region. Looking back, many interview participants  questioned whether the “Kaiser model” ever had a realistic chance of succeeding in North Carolina, given the economic and market challenges it faced in terms of its volume and cost. PGPs require sufficient population density so that they can enroll a critical mass of members within a referral area sufficient to support a multispecialty group practice that represents most of the secondary care specialists. 

The Research Triangle had the smallest and least dense population base of any market in which KP operated. The market analysis on which KP based its decision to enter the state suggested that KP–Carolina would have had to enroll 40,000 members in order to achieve financial viability, a figure that, in retrospect, looks astonishingly low. As it turned out, KP–Carolina needed a much higher enrollment, perhaps as many as 100,000 members in the Triangle alone, to reap the economies of scale on which its business model depended. Although it exceeded that number in the state, it never gained that market share in the Triangle. 

Two important marketing and financial assumptions proved difficult to meet: (1) people would join KP because it was KP (even though its provider network was tiny and a closed-group practice was an unfamiliar arrangement), and (2) membership would grow because their premiums would be 20 to 25 percent below those of their competitors. For KP to market its comparative advantage, employers would need to be willing to offer choices among plans and also be willing to structure the choice so that employees could share the savings if they selected a less expensive plan. 

As is the case with any new health plan, KP–Carolina faced a two-tiered marketing challenge. First, it had to convince employers to offer its product. Then, if employers offered several different options to their employees, KP had to convince the employees to choose its product over other health plans. When KP entered the market in the mid-1980s,there was no organized employer/purchasing-group presence in North Carolina—that is, no “exchange” or entity to link employers, employees, and health plans and to arrange a greater choice of plans for beneficiaries. Most employees worked in firms in which multiple choices would be very costly to administer, such as national firms with a small concentration of employees in many regional markets. 

KP entered the North Carolina market well before employers were able to act collectively to make KP’s PGP attractive to their employees. Not until 1995 did employers in Burlington and Greensboro, North Carolina, including the lab services giant LabCorp, form the Piedmont Health Coalition. And not until 2002, facing the rising costs of employee health benefits, did the Research Triangle’s largest employers, like IBM and Cisco Systems, become members of the newly formed Triangle (now North Carolina) Business Group on Health (personal communication from Jack Rodman, May 20, 2002). 

Unfortunately for KP–Carolina, many employers (especially small firms) preferred contracting with a single carrier or insurer that could offer a menu of health plan options to contracting with multiple carriers. This model of health plan/employer-exclusive contracting is known as “single-plan replacement” because the employer contracts with a single carrier at a time, replacing it when necessary with another single carrier. 

Major commercial carriers and insurers have capitalized on the administrative convenience of single-plan replacement and have developed employee-choice products for small employers. Carriers can use these products to satisfy employers’ need for a range of coverage options and price points, avoid adverse selection, and maintain an exclusive contract with each employer. Because of adverse selection and other concerns, carriers often refuse to offer a comprehensive plan along-side a different carrier’s plan that has a markedly different level of coverage. 

Therefore, because it had only one product to offer, KP– Carolina could not compete effectively in the “single-plan replacement” market. Very late in the game, KP–Carolina developed and marketed a point-of-service and IPA product, but interview participants characterized the efforts as “too little, too late.” They also noted that KP–Carolina was having trouble competing for some larger employers’ accounts. For example, a KP manager noted that many of the large banks in Charlotte claimed that commercial insurance carriers offered better deals, broader geographic coverage, and greater physician choice than KP–Carolina could. 

In addition, some large employers were national firms with centralized buying offices that negotiated only with those health plans that had a presence in several states, including those where KP did not have a presence. KP–Carolina faced two other challenges in marketing its product to private-sector enrollees. First, employers that contracted with KP–Carolina often continued a historical practice of paying all or a high fixed percentage of the monthly premiums. KP’s initial market analysis signaled the pervasiveness of this practice when it noted that only 35 percent of the 134 employers it surveyed required employees to pay all or part of the monthly premium (KPAS 1980). 

That is, two-thirds of  the surveyed firms dampened one of KP–Carolina’s key selling points to prospective enrollees: comprehensive coverage at a low cost. Although we could not obtain historical data on employee health insurance premium contributions from employer/purchasers sources other than the SHP, former KP–Carolina executives told us that many large employers effectively subsidized more costly plans by failing to allow employees to keep all their savings if they chose lower-cost alternatives, such as KP. In addition, KP’s premiums were higher than those of some of its competitors. Second, KP’s own market analysis also warned that most of the people surveyed by the Commission on Prepaid Health Plans reported having a “regular doctor.” 

Given KP–Carolina’s closed-panel medical group–practice structure, for many prospective enrollees, choosing KP meant switching primary care providers. Consequently, KP–Carolina attempted to enroll recent arrivals to the state, the majority of whom relocatedtotheCharlotteandTrianglemarkets.Eveninthesemore“concentrated”employer markets, the geographic dispersion of the towns and cities meant that many prospective enrollees would have had to drive longer distances to visit a KP–Carolina physician or facility than they were used to doing or would have had to do with other health plan options. 

Carolina Permanente Medical Group and Organizing the Provision of Medical Care 

In addition to meeting the two-tiered marketing challenge, KP’s success depended heavily on efficiently providing medical care through the group-practice model. Here, too, the Carolina Permanente Medical Group encountered difficulty. As one interview participant noted: In those markets where [KP] had created a favorable cost structure, the local medical group has essentially taken and executed the responsibility for making that happen. They have found ways to perform more effectively than community physicians, often requiring some sacrifices on the part of the physicians to achieve that level of performance either in how hard they work or in how much they get paid or in how they utilize resources. 

The medical groups have themselves created those efficiencies. Whether it’s will or skill or circumstances, the medical group in Raleigh was never able to achieve the level of efficiency that would afford that kind of cost structure. (KP Institute 2002) 2 The CPMG also faced the dilemma of having to operate many small clinics in order to maintain convenient access for its geographically dispersed enrollees, even though the high fixed costs of doing so under- cut its ability to achieve and sustain the scale efficiencies necessary to support its cost-leadership strategy. 

One former KP–Carolina manager pointed out that “in Raleigh, there was the veritable one-stop shop, and it was [similar] in Charlotte. Those clinics were filled. So there were places in North Carolina where they were able to put the prod- uct almost together” (UNC 2002). However, not all of KP–Carolina’s nine medical offices had such a high volume of patients. KP–Carolina and the CPMG could not achieve the economies of scale necessary to generate and maintain market share by offering lower premiums, lower out-of-pocket expenses, and more comprehensive benefits to prospective enrollees. PGPs require a supply of high-quality providers (hospitals and spe- cialists) willing to contract on terms similar to those granted to other carriers. The CPMG did not employ its own specialists or own its own hospitals, as KP and its medical groups did in other parts of the coun- try. Even obstetric and gynecological services, often considered primary care specialties elsewhere, were not provided by the CPMG until the early 1990s, perhaps because KP–Carolina lacked the volume necessary to support the fixed costs of internalizing these services. This meant that the CPMG exerted less control over utilization and costs than it might have. And many interview participants suggested that the CPMG com- pounded the problem by contracting with specialists and hospitals at unfavorable fee-for-service rates. The high fixed costs of maintaining a PGP infrastructure became an increasing liability as new market entrants increased the level of price competition. Prospective enrollees did not see a significant enough price difference to offset the restricted choice and geographic inconvenience of KP–Carolina physicians and facilities. In addition, as one former CPMG employee observed, 
 

“The middle ’90s is when everybody was starting to be pushed into HMOs, and because it was a forced choice, it wasn’t something that people were voluntarily choosing. People were looking to the health plan where they could maintain their provider relations. And that wasn’t [KP]” (UNC 2002). 

Organizational Constraints and the Failure to Adapt Beyond these “external” political and market obstacles, KP–Carolina faced additional difficulties resulting from “internal” organizational factors. 

National Corporate Constraints Like other organizations with multidivisional structures, KP’s corporate headquarters struggled to find the right balance between giving new regions the flexibility and autonomy they needed to respond to local market conditions and advancing KP’s corporate goals and maintaining consistent policies. 

Some of KP’s corporate decisions benefited the organization as a whole but constrained the region’s ability to respond to start-up demands in a challenging market. For example, the company required the region to repay its start-up debt with interest, at rates that some former KP–Carolina managers described as above market. From the KP corporate point of view, the decision made sense given its experience in Texas, where the West Coast medical groups had been incensed about having to subsidize an operation that lost money for 15 years. 

Moreover, it needed its investment repaid quickly, as its West Coast regions were clamoring for funds for information technology and facility upgrades. 

In our interviews, current and former KP corporate executives downplayed the significance of the debt service requirement in KP–Carolina’s performance problems, but former KP–Carolina regional managers expressed a different point of view. As one stated, “Our price points were not consistent with what a peer marketplace would have us handle because we were managing a very substantial interest load on the debt that we had accumulated over those additional years” (UNC 2002). 

As the operation struggled to win market share and generate economies of scale in the midst of intense price competition, local managers viewed internally imposed debt service, which some estimated to be as much as $10 per member per month, as making a tough situation worse. Similarly, in the early 1990s, KP’s corporate headquarters urged KP– Carolina and other regions to develop business plans for achieving and sustaining a 15 to 20 percent price differential from competitors. Faced with heated competition from network-model managed care plans, KP corporate headquarters jointly conducted a study with a major consulting firm and affirmed its commitment to the cost-leadership strategy. 

The consulting firm encouraged KP to think about market share as an important success factor. The conventional wisdom was that with a good economic base, marketplace presence was a key factor. According to one KP national executive, 
 

“If a given plan was not one of the several large players in any local market, it was unlikely to secure the discounts and other important contractual provisions from providers that would make it successful” (KP Institute 2002). 

At the time, KP–Carolina had just reported its first profitable year and had priced its product above the average for the market in order to reflect its true costs of operation (including service on the start-up debt). Again, the opinions on the wisdom of this corporate decision differed. As one CPMG leader stated: We were encouraged, to use a mild word, to reduce our operating infrastructures enough to support a 15 to 20 percent lower price point. But we were unable to cover the cost of what we were doing. The solution was applied uniformly across the organization, which I think in retrospect everybody believes was a problem. (UNC 2002) 

Current and former KP national executives disagreed, arguing that the problem was not the wisdom of the strategy but, rather, its poor execution by the medical group. Corporate executives cited the inability or unwillingness of regional leaders—especially in the CPMG—to build a sustainable business model based on tight cost control. 

Regional managers questioned whether the classic “Kaiser model” could work in North Carolina, claimed they had done their best, and railed against company-imposed constraints that, they argued, limited their flexibility to respond adaptively to difficult local political and market conditions (which differed from KP’s previous experience). These constraints were said to include limits on plan and benefit design as well as on advertising, sales, and marketing. Interview participants disagreed about whether the failure to achieve the requisite operating efficiencies to sustain this corporate objective resulted from problems of ability, willingness, or circumstances. The general business literature indicates, however, that simultaneously achieving revenue growth and cost control in a start-up is daunting even under ideal market circumstances (e.g., Kaplan and Norton 1996). 

Inexperienced Regional Management Like many organizations, KP routinely transferred personnel laterally in order to shift needed expertise from one region to another and to enhance career development. Several interview participants reported, however, that this did not work well in the case of KP–Carolina. In particular, some suggested that KP staffed the North Carolina region with senior managers who had little start-up experience or entrepreneurial spirit:
 

“Just because someone was successful managing a medical center in Northern California did not make him the right person to come in and build the practice in North Carolina” (KP Institute 2002). 

Not surprisingly, senior KP–Carolina managers disagreed, contending that the lack of entrepreneurship stemmed not from inexperienced management but from national corporate constraints on innovation. As one said: The original leadership brought in to start this region was supposed to replicate, not innovate. The job was to replicate the KP model in North Carolina. There was no room for entrepreneurship, nor would it have been welcomed. We were given binders of material that we were supposed to use—staffing ratios, financial reporting formats, marketing materials, benefit plans, graphic standards, staff training tools, even floor plans. They were very interested in taking our start- up, using the materials they gave us, and coming out of this experience with a “cookbook” (their words) for additional expansions around the country. They wanted us to stick to the recipe. (UNC 2002) 

Most former KP local managers agreed, however, that some of the mid-level managers and medical directors transferred from other KP regions “[were] really great about the Kaiser model, but didn’t have the slightest idea how to adapt that locally”(UNC2002).

It was the Oakland way—or no way. Divergence from Core Competence As the competition from less restrictive managed care products intensified, KP–Carolina tried to complement its group-model HMO with a point-of-service product and an IPA-model product. Two problems immediately arose. First, KP–Carolina did not have the organizational capability to effectively manage an extensive network of contracted providers. Its business systems could not track members as they moved through a more open, networked delivery system, pay claims in a timely and accurate manner, or monitor utilization across loosely affiliated physicians and hospitals. 

As one former CPMG employee noted: We didn’t know how to pay a claim. We hadn’t had to pay very many claims. We had a lot of capitated and prepaid specialty arrangements and all of a sudden we started getting claims in, and I remember when 50,000, you know, six months of claims went unpaid. It was outrageous. (UNC 2002) 

Other KP–Carolina managers concurred: “Kaiser's claims system was terrible, and KP–Carolina just could not keep up with what was happening in the marketplace” (UNC 2002). 

KP–Carolina managers and medical directors found it hard enough to build the familiar group- model delivery system from scratch under less than hospitable market conditions. Simultaneously creating and managing a network model so far removed from KP’s core competence proved impossible. As one KP– Carolina manager said:
 

“The problem was that a different business model requires different skills and competencies, different systems, different data, and so forth. Without the skills and competencies and data and systems to manage these changes, the outcome was going to be unfavorable from a financial point of view” (UNC 2002). 

Moreover, by trying to straddle the gap between staff-model and network-model product markets, KP–Carolina diverted precious resources from its core product and its core constituencies. As one former KP–Carolina manager noted:

 
“The resources were not available to do both models well. The original group-health model concept did not get the attention that it needed and started to deteriorate. The baby was sacrificed with the bath water” (UNC 2002). 

In the Research Triangle, for example, KP–Carolina attempted to implement a “concentric circle” strategy in which the inner core—the staff model—was the protected, preferred delivery system around which it hoped to establish a number of rings consisting of physicians contracted to various degrees. This seemed a good way to build enrollment, increase volume, and counter the growth of competing products. Yet as one former KP–Carolina manager put it, 
 

“We could never really execute on that strategy of how to...keep the inner core happy, with the existence of the outer-core physicians out there....I felt there was a real internal struggle” (UNC 2002).

Reflecting on KP’s flirtation with network models in North Carolina and other regions, several interview participants commented that the flawed strategy nearly cost the company its soul. One former KP–Carolina  manager, for example, described it as “a near-death experience” that prompted many in Kaiser to go “back to fundamentals [and ask], ‘Is that what we are really good at?’” (UNC 2002). 

Perhaps this is the most important internal lesson that KP as an organization could learn from its North Carolina experience. It is easy for outside observers to be critical of a business strategy and outcome after the fact. However, reflecting on our analysis of the North Carolina experience, David Lawrence, KP’s CEO at the time, concluded that the regional failure was due to an internal corporate failure to understand how to expand: KP expanded with a missionary zeal that substituted for careful, thoughtful planning and development of the core modules required to incrementally build a viable business. 

We did not learn from other industries, follow established pathways for successful expansion that have occurred in other industries, etc. It is an important lesson for us....The failure was not of prepaid group practice per se but of planning and execution of a prepaid model in a way that would increase the likelihood of a viable outcome. We operated with a California bias and had no real understanding of what was required to accomplish, execute a start-up, and to build a successful business. I do not think the model was wrong; rather, it was in the execution. 

Stated differently, I do not believe we have tested whether or not the model can be successful yet. (e-mail communication, KP Institute on Health Policy, October 23, 2002) We thus conclude that the KP experience in North Carolina illustrates in microcosm the complex interdependencies that determined the fate of a KP expansion effort, not to mention similar efforts by other PGPs around the country. A Mismatch of Model and Market? The demise of several KP regional expansions reinforces the importance of the numerous interlocking pieces that are necessary to foster a market in which a prepaid group practice can exercise its competitive advantage. It is clear that in North Carolina none of the important factors was pointing in the right direction. There was no “smoking gun” behind KP-Carolina’s demise, just as no single factor accounted for KP’s more successful expansion efforts elsewhere. However,while KP used to be able to charge less for more comprehensive benefits than others were charging for less comprehensive benefits, KP now has a smaller price advantage. 
 

KP’s historical business model attempts to build efficient-scale operations and vigorously pursue cost control while maintaining acceptable levels of quality and service. Achieving low overall costs requires enough enrollees to support the internalization of most specialties into the medical group as well as access to production inputs (e.g., hospital services) on terms as favorable as those of one’s competitors. By achieving low overall costs through efficient, high-volume operations, KP can offer low premiums and low out-of-pocket expenses as well as more comprehensive benefits. Furthermore, in its fully perfected version, the Kaiser model would offer more coordinated care delivery through a group-practice culture and internal coordination among medical and hospital service providers in exchange for a more restricted choice and, sometimes, less convenient access. 

The historical business model clearly did not succeed in North Carolina. Kaiser Permanente's expansion model also failed in Texas, Kansas City, New York, and New England. 

Where KP has been successful in entering markets, an important factor has been strong backing from influential local organizations (e.g., unions) and institutions (medical center affiliation). Local sponsors such as unions have provided an enrollee base and lent important political support. On the West Coast and in Colorado, Kaiser had strong backing from the AFL-CIO, which liked its emphasis on comprehensive benefits and preventive medicine and demanded that employers offer Kaiser as an alternative to traditional insurance. Regional expansions have also been successful in Georgia (with more than 270,000 members) and in the Washington, D.C., area (with more than 500,000 members), both of which Kaiser entered via acquisitions. In Atlanta, the Kaiser Foundation Health Plan of Georgia bought the assets of Maxicare Georgia in 1988, making Kaiser the second largest HMO in the city. Although this acquisition was accompanied by initial losses—$27.5 million in 1990, and $17.7 million in 1991—the plan had a net income of $6 million by 1992 and, with 155,000 members, had become Atlanta’s largest HMO. While there were widespread losses in the competitive Atlanta market, KP’s ability to consolidate allowed it to reach a critical mass of enrollees. Kaiser entered Washington by acquiring the Georgetown University Community Health Plan (GUCHP) in 1980 and gained important institutional support and affiliation from the Georgetown University Medical Center. Kaiser has also continued to grow steadily in Colorado, Hawaii, and on the West Coast. 

Kaiser Permanente s successes in California and Oregon and its failures in Dallas, Raleigh-Durham, and Kansas City were due in part to the large scale it was able to achieve in the former before the arrival of the large “carrier HMO” companies, but not in the latter. 

It was able to build on the West Coast from the 1950s through the 1970s, when the managed care industry was young, and independent competing medical groups were scarce. It achieved a network scale and scope that would be difficult to replicate today, when the industry is mature and competitors abound. When Kaiser expanded outside its core markets in the 1980s, as was the case in North Carolina, the industry was maturing, and sophisticated competitors were plentiful. The North Carolina case illustrates the difficulties of replicating the vertically integrated model in new geographic markets under these circumstances. 

Kaiser Permanente maintains a dominant position on the West Coast, and hybrid entities that embody some but not all the elements of prepaid group practice can be found in many metropolitan areas. In fact, the national Kaiser Foundation Health Plan reported a positive financial div performance in 2003, its net income growing from $161 million in the second quarter of 2002 to $306 million in the second quarter of 2003. Its operating income rose to nearly $300 million from $168 million a year earlier (Kaiser Permanente 2002). In its most recent report, Kaiser div Permanente s national membership (in nine states and the District of Columbia) remained flat at 8.3 million. But the trend in the health care marketplace generally is toward broad-network insurance products divorced from provider systems, retrospective rather than prospective payment, a purchasing framework that emphasizes copayments at the time of service rather than a cost-conscious choice at the time of insurance enrollment, and an institutional framework hostile to the principles and practices of managed competition. 

In six of the eight geographic regions that Kaiser Permanentediv serves, the two largest customer groups are state and federal employees. In the other two, the largest enrollee groups are federal employees and a public school system. Why has Kaiser done so well in the California Public Employees’ Retirement System (CalPERS) (more than 410,000 members) and Federal Employees Health Benefits Program (FEHBP) but fared so poorly in the North Carolina SHP.  First, a state employer purchaser’s policy of offering several choices of carriers and a fixed dollar contribution to public employees’ insurance programs has almost always been important in the success of Kaiser Permanente.

The dominant presence of the FEHBP and its enrollees was doubtless a major contributor to the success of Kaiser in the Washington, D.C., area. Second, within both the CalPERS and FEHBP, the playing field is reasonably level. CalPERS has standardized the benefit packages offered by all HMOs, and the statewide preferred provider organizations (PPOs) have coverage sufficiently comprehensive that their premiums are higher than those of the HMOs, reflecting PPOs’ inherently greater cost for the same covered benefits. 

In the FEHBP, the benefits are not as standardized, but the Office of Personnel Management has required that the benefit packages offered be fairly comprehensive. In addition, in both the CalPERS and FEHBP, retirees are treated as a separate risk pool, in recognition of their higher costs, and the employer contributes to dependents’ coverage quite generously. However, not all is perfect even in the ideal purchasing environment. In the CalPERS program, Kaiser is concentrated in urban areas in California, where health plan competition still works, which means that with a statewide premium, the PGPs enjoy the luxury of not having to operate in rural areas that are costly because of local provider monopolies but where the self-funded PPOs do operate. It is likely that certain markets, such as most rural or commuter areas, may not have the geographic conditions to sustain a profitable private-sector PGP, even when these conditions are met.

What can private (employer) and public policymakers do to make market environments more hospitable to the PGP model of care delivery? Is managed competition vital to the PGPs’ future success? The essential insight of managed competition, as a reform, is to divide the provider community into competing economic units and then to offer employees a responsible choice with premiums that reflect the differences in per capita cost, in order to give them an incentive to choose the efficient providers (Enthoven 1993). 

If a critical mass of employers were able to do this in any market area, managed competition advocates still claim, they would create the environmental conditions in which efficient de- livery systems could enter, market their superior value for the money, and achieve economies of scale. By “critical mass,” we mean enough that each of several competitors could grow to a point at which they achieved economies of scale (such as 500,000 enrollees in a metropolitan area). The following elements must be present if a PGP is to have access to the employee (not just employer) market: a broad choice of health plans; risk adjustment to mitigate adverse selection; an employer contribution that allows employees to retain any savings resulting from an economical choice; a level regulatory playing field among HMOs, insurers, and self-insured plans; and reliable, comparable information about the quality of care and consumer satisfaction. Recent research has shown wide variations among providers in the resources used to treat the same conditions and produce the same outcomes (Fisher et al. 2003). 

Employers might offer employees a price-sensitive choice among existing HMOs and encourage HMOs to develop selective networks to improve their performance. (The all-inclusive network favored by employers in the single-source model is sure to be ineffective.) Alternatively, under the protection of the Employee Retirement Income Security Act, employers might develop several selective PPOs, each of which would offer the preferred services of a different network of providers. 

If there are effective IPAs in a market, employers might build their choices on them or even share the gains created by the most efficient providers by offering them bonuses for achieving performance goals. The focal points for these networks would likely be different hos- pitals and their staffs. The idea is to be sure that those employees who choose efficient providers realize the savings generated by their choices. Prepaid Group Practice, Whither the Future? Despite the high expectations, PGPs have fared poorly in the market in recent decades. Group and staff-model HMOs have survived only where they constitute a large portion of the local market, offer an adequate choice of physicians, and gain from economies of scale similar to those of nonintegrated competitors (Hurley et al. 2002; Robinson 2004). 

PGPs have had a great impact on the American health care system and continue to offer high-quality, cost-effective care to millions of patients in particular regions and communities. Yet their future remains unclear. KP was not the only HMO to have faced competitive challenges. Nationally, no group and staff-model HMOs have performed well over the past two decades. In June 1980, the group and staff models had 7.4 million members, or 81 percent of the total HMO membership. By July 1990, with 13.1 million members, they accounted for 39 percent of the total HMO membership. In July 2002, group and staff models served 7.5 million members, 10.1 percent of the HMO total (InterStudy 2003). Membership in staff models actually declined, while membership in group models (mainly Kaiser Permanente) grew slowly and lost market share to faster-growing models, such as IPAs and mixed models. 

The models that rely on established providers and facilities can grow much faster than group and staff models that must recruit and develop their own doctors and facilities. With only limited success in particular regions, this strongly suggests that the vertically integrated HMOs have structural features—such as a narrow network and a market appeal based in part on lower costs—that put them at variance with common employer policies of purchasing from a single source and paying all or most of their employees’ premiums. By the early 1990s, some group and staff-model HMOs, seeking to attract employers that wanted a single source of health insurance, sought innovations and merged with or acquired wide networks of traditional providers to offer a long side their groups.

For example, the Harvard Community Health Plan, a “flagship” staff model, grew rapidly in the 1980s, peaked in 1992, and then began to decline. In 1993, it acquired a group- network model and entered the “mixed-model” category. In 1996, it merged with the Pilgrim IPA to become Harvard Pilgrim Health Care, now a “mixed-staff, group, IPA” model. Similarly, the Group Health Co-operative of Puget Sound, another “flagship” HMO, was a staff model as of January 1993. After experiencing three years of no growth, it brought another group into its network and created a wide network of fee-for-service, solo-practice doctors. By July 1994, it too was a “mixed-staff, group, network, IPA” model. 

In these cases, market conditions created by the employer single-source and employee contribution policies that did not highlight cost differences forced health plans to abandon the pure staff model (InterStudy 1998). Some health plans have tried to combine the virtues of organizational integration with the attractions of contractual promiscuity by wrapping a network of independent physicians around a core of an integrated group practice. 

In some cases, “mixed-model” hybrids, like Harvard Pilgrim Health Care, proved to be resting points on the road to the vertical disintegration of the insurance and delivery components. In Washington and Idaho, however, the Group Health Cooperative has combined a core prepaid group practice with a contracted network of solo and small-group practices, thereby preserving its market share. But it has not been able to leverage the distinct virtues of integrated efficiency and broad choice into a comparative advantage and so remains a niche player in a market increasingly dominated by broad network-insurance products and fee-for-service payment (Robinson 2004). As a vertically integrated organization that combines an insurance en- tity with multispecialty group practices and, in some regions, hospitals, KP possesses particular strengths and weaknesses.  

A PGP cannot be built overnight. For decades, KP and the Group Health Cooperative have been working out the kinks and links in the financing and delivery systems. The KP model is not always a winner, as evidenced by the fail- ures of KP–Carolina and of similarly structured group/staff HMOs in some other markets, nor a clear loser, as evidenced by the continuing success of KP and similarly structured entities in California and elsewhere (Robinson 2004). Hence, the most general lesson of our analysis is that the successful introduction and proliferation of prepaid group practice into markets with little or no experience with the model depend on the conjuncture of several supportive conditions. These include employers willing to offer a choice of carriers (because PGPs cannot succeed as a single source); employers willing to structure the offering to employees so that the employee making the choice gets most, if not all, of the savings; a framework that mitigates adverse selection; a regulatory framework that imposes equal burdens on PGPs and their competitors; a supply of high-quality providers (hospitals and specialists) willing to contract with PGPs on terms similar to those granted to others; and a high enough population density to permit the enrollment of a critical mass of members within a referral area sufficient to support the multispecialty group practice with most of the secondary care specialists represented. 

All this suggests that while PGPs can play an important role in market-driven reform, without the right mix of supporting factors, a variety of other, more flexible and robust models will also be needed. endnotes 1. Former KP–Carolina and CPMG managers, physicians, and employees; North Carolina Department of Insurance officials; and Duke University and University of North Carolina health policy experts participated in a discussion on the “rise and fall of KP” at the University of North Carolina. The KP–Carolina and CPMG participants’ comments are cited as UNC 2002 (March 19, 2002). 2. Former and current KP national officials in Oakland participated in a conference call organized by the Kaiser Institute of Health Policy on “lessons learned” from previous regional expansion efforts. These comments are cited as KP Institute 2002 (May 6, 2002). 

References Enthoven, A. 1978a. Consumer-Choice Health Plan (first of two parts): Inflation and Inequity in Health Care Today: Alternatives for Cost Control and Analysis of Proposals for National Health Insurance. New England Journal of Medicine 298(12):650–8. Enthoven, A. 1978b. Consumer-Choice Health Plan (second of two parts): A National Health Insurance Proposal Based on Regulated Competition in the Private Sector. New England Journal of Medicine 298(13):709–20. Enthoven, A. 1993. The History and Principles of Managed Competition. Health Affairs 12(suppl.):24–48. Fisher, E.S., D.E. Wennberg, T. Stukel, D.J. Gottlieb, F.L. Lucas, and E.L. Pinder. 2003. The Implications of Regional Variations in Medi- care Spending. Part 1: The Content, Quality, and Accessibility of Care; Part 2: Health Outcomes and Satisfaction with Care. Annals of Internal Medicine 138(4):273–98. Frank,R.,H.Huskamp,T.McGuire,andJ.Newhouse.1996.SomeEco- nomics of Mental Health “Carve-Outs.” Archives of General Psychiatry 53(October): 933–7. Hurley, R., J. Grossman, T. Lake, and L. Casalino. 2002. A Longitudi- nal Perspective on Health Plan–Provider Risk Contracting. Health Affairs 21(4):144–53. InterStudy. 1990. The InterStudy Edge, Managed Care: A Decade in Review, 1980–1990. Excelsior, Minn. InterStudy. 1993. The InterStudy Competitive Edge, Part I. HMO Directory, 5.1, 1993. Excelsior, Minn. InterStudy. 1994. The InterStudy Competitive Edge, Part I. HMO Directory, 6.1, 1994. Excelsior, Minn. InterStudy. 1995. The InterStudy Competitive Edge, Part I. HMO Directory, 7.1, 1995. Excelsior, Minn. InterStudy. 1996. The InterStudy Competitive Edge, Part I. HMO Directory, 8.1, 1996. Excelsior, Minn. InterStudy. 1998. The InterStudy HMO Trend Report, 1987–97, August 1998. Excelsior, Minn. InterStudy. 2003. The InterStudy Competitive Edge, Part II: HMO Industry Report, 13.1. Excelsior, Minn. Kaiser Permanente Advisory Services (KPAS). 1980. A Feasibility Study of a Prepaid Group Practice Health Plan in the Research Triangle Area. Oakland, Cal. Kaiser Permanente Institute for Health Policy (KP Institute). 2002. Lessons Learned from KP’s Expansions and Divestitures. Meeting notes, May 6. Kaplan, R.S., and D.P. Norton. 1996. The Balanced Scorecard. Boston: Harvard Business School Press. Luft, H. 1978. How Do HMOs Achieve Their Savings? Rhetoric and Evidence. New England Journal of Medicine 98:1336–43. McNeil, R., Jr., and R.E. Schlenker. 1981. HMOs, Competition and Government. In Health Maintenance Organizations: A Milbank Reader, edited by John B. McKinlay, 31–60. Cambridge, Mass.: MIT Press. Moran, D.W. 1981. HMOs, Competition, and the Politics of Minimum Benefits. In Health Maintenance Organizations: A Milbank Reader, edited by John B. McKinlay, 72-92. Cambridge, Mass.: MIT Press. North Carolina Commission on Prepaid Health Plans. 1979a. Interim Report: Findings, Recommendations and Executive Summary. Vol- ume 1. Raleigh, N.C., April 1. North Carolina Commission on Prepaid Health Plans. 1979b. Interim 

Report:TechnicalReportsandSupportingData.Volume2.Raleigh, N.C., April 1. North Carolina Department of Insurance (NCDOI) Financial Evalua- tion Division. 2000. Kaiser Foundation Health Plan of NC Annual Statements (1985–2000). Raleigh, N.C. NorthCarolinaDepartmentofInsurance(NCDOI).2002a.KaiserFoun- dation Health Plan of NC Annual Statements (1985–2000) submit- ted to NCDOI Financial Evaluation Division. Raleigh, N.C. NorthCarolinaDepartmentofInsurance(NCDOI).2002b.KaiserFoun- dation Health Plan of NC Annual Statements (1993–1998) submit- ted to NCDOI Financial Evaluation Division. Raleigh, N.C. North Carolina State Health Plan (NCSHP). 1993–1998. Enrollment by Health Plan Option. Raleigh, N.C. Robinson, J. 2004 (forthcoming). The Limits of Prepaid Group Practice. In Reforming Health Care: The Contributions and Promise of Prepaid Group Practice, edited by A. Enthoven and L. Tollen. New York: Jossey-Bass. Saward, E.W., and M.W. Greenlick. 1981. Health Policy and the HMO. In Health Maintenance Organizations: A Milbank Reader, edited by John B. McKinlay, 1-30. Cambridge, Mass.: MIT Press. Silberman, P. 1996. Kaiser Foundation Health Plan of North Carolina. Ph.D. diss., University of North Carolina, Chapel Hill. UnitedHealthcare. 1994. A Glossary of Terms: The Language of Managed Care. Minneapolis, Minn. University of North Carolina (UNC) (Chapel Hill). 2002. The Entry and Exit of KP. Meeting notes, March 19. Acknowledgments: We gratefully acknowledge the numerous interview participants who gave generously of their time and insights, especially Dr. Eugenie Komives (CPMG), Ann Lore, and Dan Soper (KP–Carolina) for their analysis. The Kaiser Permanente Institute for Health Policy provided grant support and information, but the analysis and conclusions are solely our own. Address correspondence to: Daniel Gitterman, Department of Public Policy Abernethy Hall, CB#3435, University of North Carolina, Chapel Hill, NC 27599– 3435 (e-mail: danielg@email.unc.edu).  

Appendix List of Interview Participants Bradley Adcock, Vice President, Government Affairs, Blue Cross Blue Shield of North Carolina (March 19, 2002). James Bernstein, M.H.A., Director, North Carolina Office of Research, Demonstrations and Rural Health Development; Assistant Secretary for Health, North Carolina Department of Health and Human Ser- vices; former President, Foundation for Advanced Health Programs, Inc. (formerly North Carolina Foundation for Prepaid Health Plans) (March 18, 2002). William Brandon, Ph.D., Metrolina Medical Foundation Distinguished Professor of Health Policy, University of North Carolina, Charlotte (March 19, 2002). Barbara Burke, Senior Deputy Commissioner, Technical Service Group, and former Deputy Commissioner, Managed Care and Health Benefits Division, North Carolina Department of Insurance (June 21, 2002). Christopher Conover, Ph.D., Assistant Research Professor, Duke University; former consultant, North Carolina Health Reform Commission (March 19, 2002). Ray Coppedge, M.D., Executive Director, Key Physician IPA, Wake County,N.C.;ExecutiveDirector,Patient’sChoice,ResearchTriangle, N.C. (June 20, 2002). David Coulter, M.H.A., Vice President, Rex Hospital, Raleigh, N.C. (June 19, 2002). Wally Dawson, Vice President, Aon Consulting, Raleigh, N.C.; former Director of Sales and Marketing (Triangle region), KP–Carolina (March 19, 2002). Allen Feezor, M.A., Health Benefits Administrator, California Public Employees’ Retirement System (CalPERS); former North Carolina Deputy Insurance Commissioner and Executive Administrator, North Carolina State Health Plan (June 11, 2002). Claudia Ghianni, M.P.H., Director, Health and Wellness Center, Belmont Abbey College; formerly with Health Appraisal Unit, KP– Carolina (March 19, 2002). Bill Gillespie, M.D., former President, Regions Outside of California, Kaiser Foundation Health Plan, Inc.; former Senior Vice President for Quality, Kaiser Foundation Health Plan, Inc. Kaiser Permanente;  former Medical Director, CPMG and Texas Permanente Medical Group (May 6; June 20, 2002). Karol Gioannini, Director of Center Operations for Wake Health Services, Inc.; former Medical Center Administrator, KP–Carolina (March 19, 2002). Russell Guerin, President, Managed Health Resources and Senior Vice President, Carolinas HealthCare System. Sam Havens, former President and Chief Operating Officer, Prudential Health System. Nancy Henley, M.D., Clinical Assistant Professor of Medicine, University of North Carolina, Chapel Hill; former Medical Director, KP– Carolina; former Medical Center Administrator, KP (March19,2002).  James Hunt, Governor of North Carolina (1977–85, 1993–2001) (June 19, 2002). Harrison Kaplan, Attorney and Lobbyist, Raleigh, N.C.; former Director of Government Relations and Counsel, KP–Carolina. Eugenie Komives, M.D., Medical Director, North Carolina State Health Plan; former Associate Medical Director for Health Policy and Utilization Management, CPMG (March 19, 2002; September 4, 2003). Jim Lane, Senior Vice President for Policy and Planning, Kaiser Foun- dation Health Plan, Inc. (May 6, 2002). Anna Lore, B.S.N., Government Affairs Representative, Duke Univer- sity and Duke University Health System; former Executive Director (Triangle Operations), KP–Carolina; former Health Plan Manager, Wellpath (March 19, 2002; August 22, 2003). Don Madison, M.D., Professor of Social Medicine, University of North Carolina, Chapel Hill (March 19, 2002). Paul Mahoney, Executive Director, North Carolina Association of Health Plans (May 2002). Ray Martinez, Deputy Commissioner, Financial Evaluation Division, North Carolina Department of Insurance (March 19, 2002). Dave Morgan, National Practice Leader, Permanente Company, LLC (May 6, 2002). Joe Morrissey, Ph.D., Deputy Director, Cecil G. Sheps Center for Health Services Research; Professor of Health Policy and Administration, University of North Carolina, Chapel Hill (March 19, 2002). Sandra Newton, M.D., Regional Medical Director, Blue Cross Blue Shield of North Carolina; former physician, KP–Carolina (March 19, 2002). 

Nancy O’Dowd, Deputy Commissioner, Managed Care and Health Ben- efits Division, North Carolina Department of Insurance (March 19, 2002). Lynette Omar, M.H.A., Assistant Director for Administration, Injury Prevention Research Center, University of North Carolina, Chapel Hill; former Medical Office Manager, KP–Carolina (March19,2002). David Pockell, Chief Program Officer, California Healthcare Foundation; former Executive Vice President and Northern California Regional Manager, Kaiser Foundation Health Plan, Inc. (May 6, 2002). Derek Prentice, M.D., Senior Medical Director, Medical Resource Management, Blue Cross Blue Shield of North Carolina; former Associate Medical Director, Health Care Policy, KP–Carolina (June 14, 2002). Tony Rand, North Carolina State Senate Majority Leader; Chairman, North Carolina State Health Plan (Legislative) Oversight Committee (June 20, 2002). Jack Rodman, Executive Director, Triangle (North Carolina) Business Group on Health (May 20, 2002). Carol Scheele, Associate General Counsel, North Carolina Medical Soci- ety; formerly with EQUICOR/CIGNA HealthCare Corporation (May 20, 2002). Paul Sebo, Health Plan Program Manager, North Carolina State Health Plan, Raleigh, N.C. (March 19, 2002). Pam Silberman, Ph.D., Associate Director, Policy Analysis, Cecil G. Sheps Center for Health Services Research; Vice President, North Carolina Institute of Medicine; Clinical Associate Professor, Depart- ment of Health Policy and Administration, School of Public Health, University of North Carolina, Chapel Hill (March 19, 2002). Dan Soper, President and Chief Operating Officer, Elsinore Technolo- gies, Inc., Raleigh, N.C.; former Executive Director of Health Plan Operations, KP–Carolina (March 19; June 14, 2002). Lynn Spragens, M.B.A., Senior Consultant, The Bard Group; former Chief Financial Officer, Carolina Permanente Medical Group (June 14, 2002). Jill Steinbruegge, M.D., Associate Executive Director, Physician De- velopment, Permanente Federation (Cal.); former Associate Medical Director, Physician Development and Information Technology, KP (June 12, 2002). Stephen Stemkowski, Manager, Premier, Inc., Charlotte, N.C.; for- mer Director of Major Account Services, KP–Carolina; former 

Senior Information Analyst/Product Manager, KP–Carolina (March 19, 2002). Diane Stimson, Senior Vice President, Managed Care, University of North Carolina, Chapel Hill (March 19, 2002). Torlen Wade, North Carolina Office of Research, Demonstrations and Rural Health Development; Senior Vice President, North Carolina Foundation for Advanced Health Programs, Inc. (formerly North Car- olina Foundation for Prepaid Health Plans) (June 26, 2002). AlvinWashington,formerRegionalManager,KaiserPermanente(N.C.) (June 10, 2002). Herman Weil, Ph.D., Senior Vice President, Planning and Analysis (regions outside California), Kaiser Permanente; former Vice President for Marketing, Southeast Division, Kaiser Permanente (June 17, 2002). Glenn Wilson, M.A., Founding Chairman, Department of Social Medicine, University of North Carolina, Chapel Hill; former Chairman, North Carolina Commission on Prepaid Health Plans. Wilson was a leader in the organization and financing of eight community- sponsored prepaid direct-service group medical practices, including the Community Health Program, Cleveland, Ohio, which eventually became Kaiser Permanente of Ohio (KP-OH) (March 19, 2002).