Effective Aids Treatment and Support in the Free State (FEATS)

The study has three main objectives, each with a set of more specific aims:

Objective 1: Present a broader view of treatment success

In order to achieve this objective, the study aims to:

• Investigate how access to antiretroviral treatment and to nutritional supplementation impacts labour productivity and time allocation of patients and other household members
• Investigate how access to antiretroviral treatment and to nutritional supplementation impacts household welfare
• Investigate how access to antiretroviral treatment and to nutritional supplementation impacts educational and health outcomes for children in households with patients on antiretroviral treatment

Objective 2: Develop a model of the determinants of treatment success

In order to achieve this objective, the study aims to:

• Determine how various individual, household and facility-level characteristics impact adherence to antiretroviral treatment, including access to a disability grant
• Determine how access to a peer adherence supporter and nutritional supplementation impacts adherence to antiretroviral treatment
• Estimate the cost-effectiveness of these interventions in enhancement in adherence, from both a fiscal and a social perspective

Objective 3: Understand the links between treatment and prevention

In order to achieve this objective, the study aims to:

• Determine how access to antiretroviral treatment impacts on the uptake of voluntary counselling and testing among members of households including patients on ARV treatment and among members of households in the general community
• Determine how access to antiretroviral treatment, to a peer adherence supporter and to nutritional supplementation impacts on the sexual behaviour of patients on antiretroviral treatment, their household members, and members of households in the general community

The study aims to investigate effective AIDS treatment and support in settings where free ARV treatment has been introduced already, with issues of scale-up and sustainability as a result representing an important issue. The study therefore is being conducted in twelve phase I ARV assessment sites in the Free State province, i.e. sites where ARV treatment first became available when the ARV treatment programme was launched in the Free State province in 2004/05.

1. Type of study The study is a prospective, cohort study with a patient, household, facility and provider survey component. The study has a social scientific focus, with a clinical component, and comprises two behavioural interventions.
2. Description of experimental design The study is a combination of a group time-series quasi-experimental design and a a variation of a double randomised consent design. The study comprises four groups of households, three of which include patients on ARV treatment (Figure 1). The provision of ARV treatment represents a quasi- or field experiment, while the peer adherence support and nutritional interventions represent a randomised-control experiment executed in accordance with a Zelen-type double randomised consent design. This study design is deemed appropriate given the fact that: blinding is not practicable or possible; the use of classical randomisation and informed consent procedures significantly threatens internal validity; the interventions are highly attractive; the control group receives standard care; the study focuses on a clinically relevant objective(s) and offers important new insights (Kaptchuk, 2001; MacLehose et al, 2001; Rains \& Penzien, 2005). Figure 1: Experimental study design Group A: Households including ARV patients who receive the ARV treatment and associated support provided as part of government's ARV treatment programme Sample size: n~216] Group B: Households including ARV patients who receive the ARV treatment and associated support provided as part of government's ARV treatment programme PLUS Adherence support provided by a trained peer adherence supporter during twice weekly visits to the patient [Sample size: n~216] Group C: Households including ARV patients who receive the ARV treatment and associated support provided as part of government's ARV treatment programme PLUS Adherence support provided by a trained peer adherence supporter during twice weekly visits to the patient PLUS Nutritional supplementation: weekly delivery of two 400g cans of meatballs and spaghetti in tomato sauce by peer adherence supporter [Sample size: n~216] Group D: Randomly selected households from the general community served by the selected health facility, excluding households where someone is known to receive ARV treatment [Sample size: n~180] Sample size estimation: The sample size calculation for the arms of the study is based on the method proposed by Freedman (1982). The sample size estimation is based on the following assumptions:

• The proportion of people showing an improvement in a specific outcome, are compared across two arms of the study at a time.
• Attrition in the control group of comparison households is assumed to be minimal or none (since there is no ARV patient in the household), while attrition in the intervention arms (where ARV treatment is considered an intervention) is assumed to be much higher, given a reported mortality of 10% among ARV patients in the first months of treatment (Free State Department of Health, 2006).
• The required sample sizes for a 4-equal-arms-at-end-of-study design were calculated taking into account the attrition rate of 17% in the three intervention groups and a ratio of 1 for the treatment: control arms at the end of the study is expected.
• Assuming a two-tailed test with Type I error of 5%, Type II error of 10%, and at least 17% improvement in any outcome. The affordable sample size was 801 households with 177 for the control arm and 208 for each of the intervention arms. This sample is expected to yield statistically significant results for the analysis of all measures contemplated. Although the statistical power of the study design with this sample size can be considered good the sample size was rounded up to ~828 households, with ~180 households in the control arm and ~216 households in each of the intervention arms. This was done to represent multiples of 12 since the study will be conducted in 12 facilities.

1. Method of randomisation From among the ~648 ARV patients recruited into the study [~54/study site], individual patients will be assigned randomly to the control group [Group A; n~216] and to the two experimental arms of the study [Group B \& C; n~216 each], using the relevant random sampling selection commands in version 10 of the Stata software programme.