Soliciting defined populations to recruit samples of high-risk older adults.

Abstract

Generalizable research on high-risk older persons requires samples that are both large enough for adequate statistical power and similar enough to community populations that its results can be generalized to them. We tested the effectiveness and efficiency of mixed-mode (mail-telephone) solicitation of a defined population as a method for recruiting a large, representative sample for a randomized trial of outpatient geriatric evaluation and management (GEM). Fee-for-service, community-dwelling older Medicare beneficiaries were mailed a short self-administered screening questionnaire. Eligible respondents were called to assess eligibility and willingness to give consent; consenters were called again for baseline data. Information about nonrespondents, ineligibles, and refusers was obtained from the Health Care Financing Administration. The response rate to the screening questionnaire was 61.1%. Of the respondents, 13.2% were eligible for the study and, of those, 34.4% agreed to participate. Response rates appeared to be influenced by small financial incentives and by subjects' age, race, sex, location of residence, and use of hospitals in the previous year. Consent rates were influenced by age and sex. The final sample (N = 522) was representative of community high-risk respondents in racial composition, previous use of hospitals, and probability of repeated admission (Pre) in the future, but it was slightly younger and contained a higher percentage of men. Recruitment costs averaged $286.92 per consenting person. Mixed-mode solicitation of defined populations can produce, at reasonable cost, large samples whose representativeness of community high-risk populations can be determined. Procedures that may enhance the success of this approach include: advance communication with members of the target population and their families and physicians; provision of medical and small financial incentives; continuous monitoring of recruitment results; and attention to subjects' needs for convenience, time, transportation, and reassurance.