that require further public agency investigation into actual care delivery or administrative systems. Multiple sources of information on satisfaction, such as routine enrollee surveys, targeted surveys of vulnerable groups, disenrollee "exit interviews," and information from appeals and grievance processes, are more likely than any single source to provide an accurate picture of reasons for satisfaction and dissatisfaction.

National and plan-specific data on the prevalence of complaints and problems could be important to government monitoring agencies as well as plans that want to improve their performance. Valid and reliable measures of the construct of consumer satisfaction are needed. NCQA has begun a study to determine what information Medicare consumers would like in order to choose among health plans. Some plans have developed their own Medicare enrollee assessment tools (Hanchak et al., 1996). Both HCFA and the Prospective Payment Assessment Commission are developing surveys for Medicare managed care enrollees and disenrollees. It would be useful to collect such information routinely over time. Under a grant from the Agency for Health Care Policy and Research, the Research Triangle Institute is developing survey modules on consumer satisfaction that will generate information that consumers can use in choosing among health plans. Although that project does not include Medicare beneficiaries, the instruments could provide models that could be tested on them in order to develop model Medicare consumer satisfaction survey tools. These current research efforts may be able to shed light on an issue not addressed in the literature, the cost to administer consumer satisfaction surveys and to compile and distribute their results.

Perhaps because of the competitive nature of the current market, it was difficult to obtain information from Medicare HMOs regarding enrollee complaints and policies to solicit and address them. (Only three of seven plans contacted for this paper responded to a request for a telephone interview.) A few plan administrators, however, did report that their plans conducted Medicare member surveys (some as often as quarterly) that reveal very high levels of satisfaction. These plans use