摘要:Objectives . We assessed the ability of managed care gatekeeping strategies (i.e., requiring a designated primary care provider to authorize referrals) to control health care costs in the mid-1990s. Methods . We analyzed expenditure data from 8195 privately insured adults sampled in the nationally representative 1996 Medical Expenditure Panel Survey. Managed care gatekeeping plan enrollees included those in health maintenance organizations and other plans requiring a primary care gatekeeper. All others were considered indemnity plan enrollees. Results . In 1996, total per capita annual health expenditures for adult gatekeeping enrollees were about $50 less than those of indemnity enrollees, primarily owing to lower out-of-pocket expenditures. After multivariate adjustment, mean per capita expenditures were approximately 6% lower for gatekeeping enrollees than for indemnity enrollees. Conclusions . In the private sector, gatekeeping strategies resulted in modest cost savings over indemnity plans. Gatekeeping arrangements have been used for decades by some health maintenance organizations (HMOs) to facilitate the provision of integrated health care. 1– 3 This system requires a designated primary care provider to authorize subspecialist referrals. In the 1980s, managed care organizations (MCOs) began instituting gatekeeping in an effort to control costs and reduce inappropriate utilization of subspe-cialists. 1– 4 On the basis of preliminary evidence that gatekeeping was cost-effective, 5– 7 this model became the dominant mode of health care delivery in the United States in the 1990s. At the same time, MCOs greatly expanded market share 8, 9 by offering highly attractive premiums to health insurance purchasers. These lower premium levels led to the widely held belief that gatekeeping was an effective cost-containment strategy. 2– 4, 10– 13 However, recent upward trends in private health insurance premiums have raised doubts about the ability of managed care strategies to control health expenditures. 10, 14, 15 Evidence that managed care gatekeeping arrangements controlled health expenditures in the mid-1990s is scarce. The lull in health care inflation from 1994 to 1997, which coincided with peak enrollment in gatekeeping plans, is often cited as evidence that MCO strategies to control costs were effective. 2– 4, 10– 13 Prior studies that examined this issue are now outdated 5, 6, 16 or have limited generalizability. 17– 19 Difficulties in estimating the cost of services in MCOs that use gatekeeping arrangements, including the proprietary nature of this information, have hampered direct comparisons with expenditures in non-gatekeeping plans. 20 A recent study of privately insured adults from a Midwestern metropolitan area showed no difference in expenditures between gatekeeping and point-of-service (nongatekeeping) arrangements for 1994–1995. 17, 18 In this study, we analyzed data from the 1996 Medical Expenditure Panel Survey 21 (MEPS) of the Agency for Healthcare Research and Quality (AHRQ) to determine whether total health expenditures in the mid-1990s were lower for privately insured adults enrolled in managed care gatekeeping plans than for those in indemnity (non-gatekeeping) plans. In contrast to previous studies, this data source supports a more comprehensive approach by providing nationally representative data on total costs to all purchasers for all types of services. We also examined costs from an insurer’s perspective to determine whether third-party payments for managed care gatekeeping beneficiaries were lower than payments for indemnity beneficiaries.