Sampling and accessing people with AIDS. Implications for program evaluation.

This article describes issues that arose in attempting to conduct a survey of people with acquired immune deficiency syndrome (AIDS) as part of an evaluation of a program to deliver health and social services to this population. Demands to maintain the confidentiality of people with human immunodeficiency virus (HIV) infection posed a large impediment to randomly sampling and accessing program recipients. Efforts to contact people with AIDS through the mediation of health service providers encountered problems of nonimplementation and slow accrual. Comparisons of the obtained sample with a more comprehensive data base of program clients suggest that clients who were more accessible and compliant were overrepresented in the sample. People with AIDS themselves, however, were willing to be interviewed, as demonstrated by refusal rates less than 11%. Future studies of people with AIDS must overcome direct service providers' lack of time to contact and recruit respondents; it may be wise to allocate funds to support recruitment activities conducted by an administrative staff person in the service delivery agency.