Hypothesis 1: The proportion of pre-ART patients whose CD4 cell counts decline to ART-eligibility within 48 weeks will be lower in intervention compared to the non-intervention arm.

Hypothesis 2: PLHIV who receive the PSCB intervention will experience lower rates of morbidity or death over the follow-up period compared to patients not receiving the intervention

Hypothesis 3: PLHIV who receive the PCSB intervention will have better adherence to scheduled clinic appointments compared to those not receiving the intervention

A randomized trial to assess the effect of patient-selected care buddies (PSCB) on adherence to pre-ART clinic appointments, clinical and immunologic outcomes of patients receiving pre-Anti-retroviral therapy (pre-ART) HIV care in a rural community HIV care program

Standard of care: Patients enrolled for pre-ART care (CD4> 250 cells/ul) received general health education, scheduled and unscheduled clinical monitoring, CD4 testing and other clinically indicated investigations, treatment of opportunistic infections, and cotrimoxazole prophylaxis. A client was expected to come to the clinic at least once in three months.

Patient selected Care buddy intervention: In addition to standard of care (above), pre-ART patients randomized to PSCB arm were requested to choose a care buddy who was aware of the patient's HIV infection and resides in the same household or in close proximity. Care buddies attended at least two HIV health education sessions similar to those provided to study participants. Information on HIV, and the importance of adhering to scheduled clinic visits and to prescribed medications were emphasized. Buddies were requested to remind participants to take their prophylactic treatments, and remind them of clinic appointments. No compensation for participation was given to buddies.

Randomization procedure Randomization and concealment This was done to minimize systematic bias in the allocation of patients to the intervention or non-intervention study arms. I used stratified block randomization to restrict chance imbalances so as to ensure that the study arms were as alike as possible for patient enrollment factors, including key factors such as sex and CD4 which can have an effect on the outcomes such as adherence to pre-ART scheduled clinic appointments. A set of permuted blocks were generated for each combination of stratification factors. In this study, these factors included patient sex and CD4 cell count.

Concealment of the patient allocation to study arms was done to avoid both conscious and unconscious selection of patients into the study. The research assistants assessed the eligibility of the patients, sought their consent for study participation, and then the enrollment officer with the sealed trial randomization envelopes let the patient pick the envelope containing the study arm. To further ensure unbiased allocation of patients to study arms, I used varying block size, specifically size 6 so that the recruiting officer does not guess which study arm follows (Concealment through sequence generation). I used a computer-generated allocation sequence with a randomization ratio of 1:1, to PSCB or SOC We collected data using interviewer-administered questionnaires and HIV clinic forms. Patients received a questionnaire at baseline, six months and 12 months follow up visits. All visits were conducted at the HIV clinics. At baseline visit, we collected socio-demographic information including level of education, occupation, marital status, distance to the HIV clinic, willingness to select a buddy as well as readiness to disclose HIV status to a buddy. At both the 6 and 12 month follow up visits, information on continued possession of a buddy, change or loss of a buddy, relationship and perceived helpfulness of their buddy (if in intervention arm), self-reported adherence to clinic appointments and cotrimoxazole, quality of life and sexual behaviors including sexual activity, condom use and number of sexual partners (marital and non-marital) were collected. As per clinic schedules, patients were required to come to the clinic for cotrimoxazole refills at least 3 monthly and to have a blood draw for CD4 testing as follows: at 3 months if previous CD4 count was 251-350 cells/ul or after six months if CD4 was greater than 350 cells/ul. Routine clinic data, collected on clinic forms included patient visit data (visit date, number of cotrimoxazole pills dispensed, blood samples taken for testing), health status (e.g. opportunistic infections diagnosed, Karnofsky score, WHO staging) and laboratory results, including CD4 counts and any other investigations as clinically indicated. Blood for CD4 testing was collected from participants at the various community clinics and transported to the central Rakai Health Sciences Program laboratory in Kalisizo.CD4 counts were assessed by flow cytometry using a FACS Calibur (Becton Dickinson, San Jose, CA, USA).

Inclusion Criteria:

• pre-ART patients aged 15 years or older, with CD4 count of 251+ cells/ul (not eligible to start ART according to the Rakai program criteria of CD4≤250 or WHO stage IV)
• provide informed consent or assent
• able to provide a care buddy

Exclusion Criteria:

• on ART
• cd4 count <250 cells/ul
• unable to provide consent/assent
• unable to provide a buddy