BACKGROUND: Little information exists regarding how health maintenance organizations (HMOs) analyze medical technologies using technology assessment (TA). OBJECTIVE: To describe characteristics, processes, and roles of TA in large HMOs. METHODS: The target survey population was large HMOs (enrollments (100,000). 165 HMOs met this criterion. Corporate consolidation of HMOs reduced the cohort to 96 HMOs, representing 77% of total U.S. covered lives. A questionnaire was developed and field tested. The survey was mailed to pre-identified HMO contacts in early 1998. RESULTS: Survey response rate was 43%. 85% (35/41) of respondents routinely conduct TA. More than 50% of the time, TA in HMOs includes representatives from medical, pharmacy, quality improvement, legal and benefits functions. TA is initiated by HMO medical affairs departments (>70% of the time). HMOs conduct TA on equipment, procedures, devices, and several types of drugs. All respondents conduct TA either at time of product introduction, adoption into practice, or upon FDA approval. HMOs conduct TA again for one of three reasons: competitive alternatives, new indication, or replacement. Legislative, societal, and accreditation demands can influence HMOs to conduct TA. HMOs consider randomized controlled trials, published in peer reviewed journals as the most desirable data source. Other data are used, but were rated less favorably. Lack of timely data can limit TA. Over 90% of respondents seek outside assistance for TA. HMOs typically conduct TA to determine benefits coverage policy. CONCLUSION: Most HMOs surveyed conduct TA. Medical staff influence TA in HMOs. HMOs perceive randomized controlled trials as more desirable than other data. Outcomes data is often not available at time of TA. HMOs use TA consultants as one source of information.