Assisted partner notification services to augment HIV testing and linkage to care in Kenya: study protocol for a cluster randomized trial

Abstract

Background: HIV case-finding and linkage to care are critical for control of HIV transmission. In Kenya, >50% of seropositive individuals are unaware of their status. Assisted partner notification is a public health strategy that provides HIV testing to individuals with sexual exposure to HIV and are at risk of infection and disease. This parallel, cluster-randomized controlled trial will evaluate the effectiveness, cost-effectiveness, and feasibility of implementing HIV assisted partner notification services at HIV testing sites (clusters) in Kenya.

Methods/design: Eighteen sites were selected among health facilities in Kenya with well-established, high-volume HIV testing programs, to reflect diverse communities and health-care settings. Restricted randomization was used to balance site characteristics between study arms (n = 9 per arm). Sixty individuals testing HIV positive ('index partners') will be enrolled per site (inclusion criteria: ≥18 years, positive HIV test at a study site, willing to disclose sexual partners, and never enrolled for HIV care; exclusion criteria: pregnancy or high risk of intimate partner violence). Index partners provide names and contact information for all sexual partners in the past 3 years. At intervention sites, study staff immediately contact sexual partners to notify them of exposure, offer HIV testing, and link to care if HIV seropositive. At control sites, passive partner referral is performed according to national guidelines, and assisted partner notification is delayed by 6 weeks. Primary outcomes, assessed 6 weeks after index partner enrollment and analyzed at the cluster level, are the number of partners accepting HIV testing and number of HIV infections diagnosed and linked to care per index partner. Secondary outcomes are the incremental cost-effectiveness of partner notification and the costs of identifying >1 partner per index case. Participants are closely monitored for adverse outcomes, particularly intimate partner violence. The study is unblinded due to practical limitations.

Discussion: This rigorously designed trial will inform policy decisions regarding implementation of HIV partner notification services in Kenya, with possible application to other parts of sub-Saharan Africa. Examination of effectiveness and cost-effectiveness in diverse settings will enable targeted application and define best practices.

Trial registration: ClinicalTrials.gov NCT01616420 .

Figures

Figure 1

Distribution of APS study sites in Kenya. The 18 HIV testing and counselling (HTC) sites participating in the APS study are shown on a map of Kenya, with estimated HIV prevalence among adults for each county [16]. A star (*) indicates that the site was assigned to the intervention (immediate APS) arm. U, urban; P, peri-urban; R, rural.

Figure 2

Restricted cluster randomization of sites in APS study. To ensure balance between study arms for key site characteristics, the 18 sites (clusters) were categorized by county and proximity to an urban area. Randomization was performed for each sub-group, and all possible combinations of these randomizations were generated. One combination was then selected using a random number generator. Note: the star (*) indicates that the three Kisumu peri-urban sites were randomized to two in one arm and one in the other; the Murang’a site then occupied the unfilled spot.

Figure 3

Study flow for intervention versus control sites, starting from when an index participant tests HIV-seropositive. Illustrated here are the study procedures for the intervention arm (immediate APS following index participant enrollment) and the control arm (APS delayed for 6 weeks after index participant enrollment). The star (*) indicates that the partner services involve: 1) partner tracing and enrollment; 2) HIV testing and counselling of partners; and 3) referral of HIV-seropositive partners to HIV care.