Evaluating Demand Generation (Stylish Man , Stylish Living) for HIV Prevention/Family Planning Services, Rakai, Uganda

Assessment strategy:

We propose a pragmatic cluster-randomized trial of the effects of MM/mobile service camps+SMLEvent (intervention) versus MM/mobile service camps without SMLEvents (control) on CHP coverage and on usage of mobile camp services by arm. In addition, RCCS data from before and after trial initiation will provide data on trends in CHP coverage, HIV incidence and key behaviors in both arms before and during the trial. The "pragmatic" approach allows modification of the SMLP components based on feedback and iterative assessments, as would be the case in normal program practice.

Cluster randomized trial component:

• 50 Rakai Community Cohort Study (RCCS) communities will be divided into ~20 clusters of 2-3 villages in close geographic proximity per arm. The total population of all 50 communities is ~18,000, of whom we estimate approximately 15,000 will be interviewed via the RCCS as a component of intervention evaluation. The clusters will be stratified by community characteristics (e.g., fishing villages, main road hubs, agrarian villages), and randomized within strata.

Intervention arm: ~ ten clusters will be randomly assigned to receive the SMLEvent activities once ~ annually for 3 cycles in addition to the ongoing district-wide MM. The SMLEvents will include the community mobilization as described above, carried out in conjunction with mobile service camps conducted at the time of the SMLEvent.

The other ~10 clusters (control arm) will be exposed to MM for 3 cycles but will not receive SMLEvents. All control communities will be offered all CHP services via RHSP/Ministry of Health (MOH) mobile clinics conducted for the same 3 cycles in conjunction with RCCS survey rounds.

Both intervention and control arm clusters will continue to have year-round access to existing MOH/RHSP clinic CHP services. Services in both arms will be offered per Uganda Centers for Disease Control and Prevention (CDC) PEPFAR guidelines. Both arms will thus have equivalent access to CHP.

Pre-post component:

The MM campaign on radio which covers a large proportion of the population of Rakai cannot be randomized. However, the RCCS provides longitudinal information on population-level rates of HIV care/ART for the general population and for pregnant women (p-MTCT), VMMC and family planning, as well as data on behaviors (condom use, numbers of partners, etc…) and HIV incidence, since 1994. Thus, within control arm communities, we have data required to determine what proportion of the population are exposed to the mass media, and whether the MM is associated with increased demand and among which population subgroups. The rate of service uptake in the RCCS has essentially plateaued for over the past 3 RCCS rounds, so an upward inflection following MM startup will suggest program effects. We will also conduct a pre-post comparison in the SMLEvent+MM arm.

As described below, we will add questions to the RCCS to enhance the SMP evaluation. CHP uptake will also be assessed with mobile camp service statistics in both study arms.

Methods of data collection:

1. Annual RCCS community surveillance surveys, among all consenting individuals aged 15-49 in the 50 RCCS communities (aggregated into ~10 clusters per trial arm).

   The RCCS will continue to collect detailed data on sociodemographics, behaviors (including numbers and types of partners, condom use with each, whether the partner is from within the community or another geographic setting), egocentric sexual network, pregnancy since the last RCCS survey round, health status and use of HCT, condoms, ART, P-MTCT, and VMC. We will add a module in the RCCS to evaluate exposure to the SMLP (mass media, SMLEvent), perceptions of components of the SMLP (both positive and negative), perceptions of different CHP services and the degree to which they are acceptable within the respondent's family, peer group and community; whether the respondent has had conversations with partners, family and/or peers about any CHP services; whether an individual accepted any service or adopted any self-reported behavioral change as a result of the campaign, and if so, what was (were) the main influence(s) (i.e., information, found out peers were using services, friends or spouses encouraged uptake as a result of the campaign, easier access to the service, etc….. ) HIV testing will be conducted at each RCCS survey round, using a three rapid test algorithm, with EIA and/or PCR confirmation of all newly identified HIV+ in-migrants and all new seroconverters. Serum samples from the RCCS will be archived at -800C for future studies such as community viral load, under future grants. Women will be asked about pregnancy status; those who are uncertain will be offered a urinary hCG pregnancy test (this is offered routinely in the RCCS). Pregnant HIV+ women will be referred for p-MTCT.

   Since the RCCS routinely collects serum for HIV testing and CD4, we will track HIV incidence, and immunologic status through the RCCS, complementing the clinic-based evaluation described below.

2. Process data:

   Records will be kept of each SMLP activity: radio spots and interviews, scheduling of SMLEvent activities, etc. SMLEvent process data will include estimates of the number of participants each day, with photos taken to assist with the count. Approximately 12 short anonymous "spot interviews" will be conducted daily at the SMLEvent site daily with younger and older men and women, to gauge interest in the activities, positive and negative perceptions, and suggestions for improvement.

3. Clinic-based evaluation:

In both study arms, we will collect service statistics from the mobile CHP camps (mobile camps plus SMLEvents in the intervention arm, and stand-alone service camps in the control arm)by services and referrals provided, client's age, gender, venue and date. Please note: the sand-alone mobile camps int he control arm will be announced to the community approximately two weeks prior to their occurrance, in order to ensure potential clients are aware of them; however, this notification will not include the multimedia and community mobilization associated with the SMLEvents. (Client medical records will include the individual's name, but will be retained separately from research records.) Clients will be asked what motivated them to accept a service (some SMLP component, other people were getting services, encouragement from someone influenced by the SMLP, factors unrelated to the SMLP, etc…) We will track whether persons referred for HIV care/ART present to our HIV clinics (and conversely, we will ask about clinic attendance during each RCCS survey rounds). Routine HIV clinic data collection includes tracking of visits, questions on adherence, and CD4 cell count every 6 months. PEPFAR does not currently fund HIV viral load, but samples from the RHSP/MOH clinics are tested for viral load annually with International Center for Excellence in Research (ICER) funds. These data will complement the RCCS data, providing a comprehensive indication of ART coverage and long term treatment outcomes.

Cost data:

We do not propose a formal cost benefit analysis. However, the cost of the campaign will be closely tracked (cost of each radio spot, personnel/materials for the SMLEvent, fuel, administrative and support costs, additional cost to PEPFAR of extended evening service hours, etc…). The focus will be on service-related costs, with research outlays tracked separately. We will compare overall costs per unit ART, VMC and p-MTCT coverage, in the intervention versus the control arm and in relation to PEPFAR costs per client to date.