payment or directing increased volume to those providers with recognized and measurable quality improvement initiatives. Both of those options reflect the need for resources to shift if broad-based and lasting quality improvement is to be achieved.

Adverse Risk Selection

Adverse risk selection occurs when an organization that bears financial risk (whether a health plan or provider group) attempts to avoid enrolling or caring for sick patients who have conditions that result in high and/or continuing costs. Both insurers and providers are concerned that undertaking quality improvement and publicizing quality outcomes will attract patients more likely to have conditions that make them high users of care and high-cost (Dudley et al., 1998). For example, if a health plan has a good program for identifying and managing those with diabetes, it is likely to enroll more diabetic patients. Since few purchasers adjust payment for this type of risk, the health plan bears the burden. Such health plans in turn, are unlikely to give their providers incentives to design programs that will attract a disproportionate share of people with diabetes. If financial risk is delegated to the provider group, the providers bear the financial burden of care for this population. The concern is related mainly to chronic conditions rather than acute care needs since the former represent ongoing expenses.

Risk-adjustment methods are an attempt to provide payment to health plans and providers that is commensurate with the health risks of the population served so that the organizations compete on efficiency, service, and quality instead of risk selection (Bowen, 1995). In the context of payment policy, risk is defined as how precisely future medical costs of an enrollee or group can be predicted (Gauthier et al., 1995). Risk adjustment is important because the distribution of medical expenditures is highly skewed, with a small fraction of individuals accounting for a substantial proportion of expenditures in any given year (Luft, 1995; Maguire et al., 1998). It has been estimated that the top 1 percent of spenders account for 30 percent of health care expenditures, whereas the bottom 50 percent account for only 3 percent (Berk and Monheit, 1992). Although some expenditures are unpredictable (such as trauma related to an auto accident), some are predictable (such as people with a chronic illness who are recognized as incurring continuing costs). Since some expenditures are predictable, organizations that assume financial risk for the care of a group can potentially avoid recognized high users of care and their costs.

There are a number of different approaches to adjust for risk, such as use of specific adjustment methods, withholds and risk pools, carve-outs, and reinsurance (Newhouse et al., 1997). Prospective risk-adjustment methods, several of which have been developed in recent years, have gained attention (particularly by the Health Care Financing Administration). Two leading models are adjusted



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement