While many of the planned data collection methods were used, some modifications had to be made in response to a broadening focus and the need to adapt to the local culture. However, the main emphasis remained on using a mix of qualitative and quantitative methodologies to collect the information. This strategy was chosen because triangulation of methods would yield different types of information, and such a mixture would not only enable a cross-validation of data, but also minimize bias. Some tools provided both quantitative and qualitative data.

4.7.1 Quantitative data

Quantitative data were collected using a coding manual which had been developed earlier and answers were recorded using the manual. Information on demographics, timing of drug administration, and some default interruptions leading to non-adherence were entered in the manual. Additional quantitative data were collected using SSIs, observation (consultations), exit interviews, and pharmacy records. Information collected on demographics included age, religious background, educational level, knowledge, attitude, perception and experience in the use of ARVs.

4.7.2 Qualitative data

The bulk of the data for the study were based on qualitative methodologies because the key problem studied, sub-optimal adherence to ART, could best be captured in this way. The following qualitative methods were used: FGDs, observations, in-depth interviews, SSIs and exit interviews. FGDs were administered to the following study populations: four groups in the community categorized by sex, five groups of ARV users, and three groups of health workers who were not categorized according to type since it was assumed that all health workers face similar challenges as they interact with PLWHIV and those on ART. Key informant interviews were conducted with local council and opinion leaders, cultural leaders, religious leaders, traditional healers and PLWHIV support groups. Semi-structured interviews were conducted with ARV users, service providers from both facilities, and ethnographic observations were conducted in both facilities with notes compiled on themes related to adherence. Other techniques used included: exit interviews with ARV users, checking pharmacy records with staff, and the use of the adherence tool with ARV users.

Focus group discussions

Focus group discussions, organized by age and sex, were conducted both with community members and with people on ART and enrolled at the selected sites. The aim was to identify difficulties that were being encountered by people on ART. The moderator had an FGD guide, used to keep the research focused on the main themes of the study. Ten FGDs were conducted (four with the community, five with ARV users and one with health workers). The location was considered when selecting participants for the FGDs (i.e. urban, peri-urban and rural setting). The FGDs were used to: determine community knowledge, beliefs, attitudes and behaviour in relation to the use of ARVs; investigate social support given to PLWHIV; and to get suggestions on ways of improving adherence to ARVs. Four of the 10 FGDs focused on getting the views of community leaders and other opinion leaders on the use of ARVs as well as the community's perception of and solutions to the problem. At each facility the counsellor helped select participants for the discussion.

In-depth interviews

These involved the use of semi-structured, open-ended interview guides with flexible probing, ideal for investigating personal experiences of ART from the subjective perspective of each respondent. The exit interviews were helpful in assessing the quality of care. They served as a back-up to the FGD findings. Twenty key informant interviews were conducted, 10 at each facility. The aim was to establish: beliefs about HIV and ARVs; community participation in HIV-related activities; support systems in place for people on ARVs; and the problem of sub-optimal adherence.

Observation (consultation)

Ten observations, five at each facility, were conducted with a doctor, pharmacist, nurse, counsellor/social worker and receptionist. The aim was to explore aspects such as interactions between clients and service providers in health facilities, the availability of ARV stocks, stigma, and the length of time spent at the facility, privacy, and organizational procedures. Observational notes were taken and later used in data analysis. The notes were used to help fill in any gaps in the data obtained during FGDs or in-depth interviews, and to triangulate data.