SEARCH Randomized Trial
Principal Investigators:
Diane Havlir, MD, UCSF; Moses Kamya, MBChB, PhD, Makerere University

Sponsors and Collaborators:
University of California, San Francisco (UCSF)
Makerere University (MU)
Infectious Diseases Research Collaboration (IDRC)
Kenya Medical Research Institute (KEMRI)
National Institute of Allergy and Infectious Diseases (NIAID)
President's Emergency Plan for AIDS Relief (PEPFAR)
Gilead Sciences
World Bank Identifier: NCT01864603

The first proposed initiative of the SEARCH collaboration is a community cluster randomized trial in Uganda and Kenya of widespread early community wide antiretroviral therapy (ART), where primary endpoints will include both community health and community economic status. The study is supported by the SEARCH consortium (The World Bank, NIAID, PEPFAR, WHO, MOH of Kenya, Uganda; Gilead).

The SEARCH study is designed to inform the current debates on global health investments precipitated by a) mathematical models predicting the HIV epidemic can be halted with widespread ART; and b) the reality of diminishing resources and growing costs of existing programs. The study is designed to leverage investments in HIV to develop sustainable systems for other infectious and NCDs. By examining individual, household, and community level socio-economic outcomes, the SEARCH study will assess whether an integrated approach to addressing health problems serve as an economic development intervention as well. Unique elements of the study a) inclusion of diagnosis and linkage to care of multiple communicable (TB, malaria) and non- communicable disease (hypertension, diabetes) in study design; b) improving and building community health delivery for ART and other diseases using new and efficient care models; c) evaluations that measure the immediate, cumulative and downstream effects of the intervention in the health, education and economic sectors; and d) a focus on sustainability.

Study Design:

1) The SEARCH study will involve the randomization of 32 rural communities in Kenya and Uganda of approximately 10,000 persons each.

2) Selected communities will include those with high HIV rates, representative of rural east Africa.

3) After a baseline community census that includes geospatial mapping and digital imaging, there will be annual community health campaigns that include HIV and multi-disease testing - an efficiency community cohort --for endpoint measurement.

4) The annual community health campaigns will be tailored and evolve to community health needs. For example vaccine access, family planning services, AMC may be added as the campaigns evolve.

5) Districts in the intervention arm will provide antiretroviral therapy to all newly identified cases using streamlined care for those with high CD4+cell count.

6) Districts in the standard arm will offer routine HIV testing, antiretroviral therapy as per country guidelines for adults, pregnant women and children.

7) The primary study endpoint is the cumulative 5 year HIV incidence in men and women over 15 years of age. Incidence will be directly measured via the cohort design.

8) Secondary endpoints for health include mortality (overall, maternal and infant), mother to child HIV transmission, AIDS, TB, and HIV drug resistance. Economic/education secondary endpoints include adult and child employment levels, asset holding, school attendance, programmatic costs, health gains in DALY and cost effectiveness. Assessments of hypertension and diabetes linkage and retention to care, and implementation science agenda focused on the HIV care cascade are incorporated.

9) The study is powered to detect a 40% reduction in 5 year cumulative incidence in treatment versus control communities under conservative assumptions regarding plausible values for cumulative incidence in control communities, evolving prevention and treatment programs, incomplete followup, and between-community variation. Sample-based weights will be used to combine testing data from tracked individuals with testing data from the community health campaigns in order to obtain consistent estimates of community-specific five year cumulative incidence.