As PMTCT and HIV treatment programmes have rolled out, more women are being kept alive and fewer children are being infected with HIV. Most mothers requiring HIV treatment in this community have an opportunity to access it and as such children of HIV-positive mothers are less likely to lose their mothers in early childhood to an untimely death. However they enter middle childhood, and the journey into adulthood, facing a myriad of other associated risk factors, not least of which are parental ill health, hospitalisation, and ultimately possible death and loss, often compounded by stigma and discrimination and a lack of adequate familial and social support network. Work under this heading includes research around depression, adherence, disclosure and resilience; research on depression and disclosure is being taken forward in PhD research by Tamsen Rochat and Dumo Mkwazani.
Antiretroviral adherence (Till Bärnighausen)
Research questions
We conducted a prospective cohort study of antiretroviral adherence at KwaMsane clinic. The study goals were:
· To identify valid and feasible instruments to measure adherence in rural South Africa, and in particular to compare estimates of adherence measured with three subjective instruments (self-report of the proportion of pills missed in the past seven days, self-report of the day of the last pill missed and visual analogue scale) and validate the three subjective instruments to measure adherence against objective criteria (viral load and CD4-count)
· To measure antiretroviral adherence in a government clinic in rural South Africa
· To identify important determinants of adherence in rural South Africa, especially the relationship between adherence and age and whether individual characteristics that could be influenced by interventions affect adherence (disclosure, coping, life style choices, social support, general mental health, concurrent treatment, access to care)
In addition, we have conducted a systematic review of the evidence on interventions to improve antiretroviral adherence in sub-Saharan Africa.
Fit
Antiretroviral adherence is one of the most important determinants of treatment success. The study thus fits in well with the Africa Centre research agenda on determinants treatment success and treatment outcomes.
Data sources and methods
We interviewed 272 patients (mean age 38, age range 20-67) attending the antiretroviral treatment services in KwaMsane clinic over a time period of nine months. In addition, we conducted clinical chart reviews of patient records. Analytical methods include ordinary least squares regressions, panel regressions, and measures of validity and measurement agreement.
Findings
In initial analyses, we have investigated the agreement of different subjective measures of antiretroviral treatment and the prevalence and determinants of depression in the adherence study cohort.
In pooled analyses (jointly with two other sites in South Africa), we found that
• 13% of patients gave discordant responses to two adherence self-report questions administered during a single interview by trained study staff (categorical assessment of adherence, visual analogue scale). Among these patients, we observed similar numbers of ‘yes-no’ and ‘no-yes’ discrepancies.
• A number of factors were associated with discrepancies in self-reported adherence based on the two questions, including younger age, duration of ARV use (<1 year or >3 years), and undisclosed HIV status. Study site was also associated with increased rates of discordance, with subjects from the KwaMsane site having the lowest rates of discordant response.
We further investigated whether the mental health of patients receiving antiretroviral treatment is influenced by social support and strategies to cope with HIV infection. Mental health was assessed in a cross-section of 272 patients with the General Health Questionnaire 12 (GHQ12). A GHQ12 score of 4 or higher indicated mental health pathology (depression), while lower scores indicated normal mental health. We regressed depression on sex, age, marital status, education, household wealth (measured with a principal component (pc) score summarizing information on water source, energy sources, electricity, home ownership, and 27 household assets), social support (measured with two orthogonal pc scores, the first capturing largely questions related to “instrumental social support” and the second capturing largely questions related to “emotional social support), and 6 strategies to cope with HIV infection. Holding the other variables constant, “instrumental social support” was a significant predictor of mental health pathology (P=0.001). Using avoidance of people as a strategy to cope with HIV increased the odds of depression almost threefold (p=0.006), “trying to keep it from bothering” one reduced it by a factor two (p=0.068). 33% of patients were depressed indicating that depression is very common in patients on ART in rural South Africa. In addition to drug treatment, interventions improving instrumental social support and changes in the strategies to cope with HIV infection may be effective in reducing this disease burden among ART patients.
Policy implications
The current and future study results inform the design of optimal instruments to measure antiretroviral adherence in rural KwaZulu-Natal. In addition, the study results will be useful in designing interventions to improve antiretroviral adherence and the general health of patients in this setting.