The sample size calculation for ARV users using the adherence tool was based on the results of the pilot study (mean overall adherence rate = 98%, p*(l-p) = 0.015605, and á at 95% confidence interval) which gave 24 per each health facility (estimated total of 168 ARV users for seven facilities). The adherence tool was used for 107 ARV users, while 100 ARV users were studied through exit interviews and semi-structured interviews (see Table 1). In some health facilities, the sample size was smaller than planned for the adherence tool because some identified users did not meet the criteria of being on treatment for at least three months. ARV users were randomly chosen using the outpatient attendance register. Each ARV user chosen was administered a single tool.

Twenty-six observations were made to see how the patient was received and dealt with in the ARV clinics and 28 health workers were also interviewed. The health workers interviewed were those who were working in the ART clinics. However, the number of staff varied between the different health facilities and therefore the expected sample size was not consistent in all facilities.

Eight FGDs were conducted with ARV users and 10 key informant interviews were carried out. In six of the health care facilities, stocks of ARV medicines were checked to assess availability.

Table 1 shows the total study population in the seven health facilities in both Arusha and Dar es Salaam according to data collection tools.

Table 1: Study population of selected health facilities in Arusha and Dar es Salaam

SSI: Semi-structured interview. FGD: Focus group discussion