Share this article on:

STI Patients Are Effective Recruiters of Undiagnosed Cases of HIV: Results of a Social Contact Recruitment Study in Malawi

Rosenberg, Nora E. PhD, MSPH*,†; Kamanga, Gift MSc, DLSHTM†,‡; Pettifor, Audrey E. PhD, MPH*; Bonongwe, Naomi NMT; Mapanje, Clement DCM; Rutstein, Sarah E. BA; Ward, Michelle MPH§; Hoffman, Irving F. PA, MPH†,§; Martinson, Francis MD, PhD†,§; Miller, William C. MD, PhD*,§

JAIDS Journal of Acquired Immune Deficiency Syndromes: April 15th, 2014 - Volume 65 - Issue 5 - p e162–e169
doi: 10.1097/QAI.0000000000000066
Implementation and Operational Research: Epidemiology and Prevention
IOR articles

Background: Patients with newly diagnosed HIV may be part of social networks with elevated prevalence of undiagnosed HIV infection. Social network recruitment by persons with newly diagnosed HIV may efficiently identify undiagnosed cases of HIV infection. We assessed social network recruitment as a strategy for identifying undiagnosed cases of HIV infection.

Methods: In a sexually transmitted infection (STI) clinic in Lilongwe, Malawi, 3 groups of 45 “seeds” were enrolled: STI patients with newly diagnosed HIV, STI patients who were HIV-uninfected, and community controls. Seeds were asked to recruit up to 5 social “contacts” (sexual or nonsexual). Mean number of contacts recruited per group was calculated. HIV prevalence ratios (PRs) and number of contacts needed to test to identify 1 new case of HIV were compared between groups using generalized estimating equations with exchangeable correlation matrices.

Results: Mean number of contacts recruited was 1.3 for HIV-infected clinic seeds, 1.8 for HIV-uninfected clinic seeds, and 2.3 for community seeds. Contacts of HIV-infected clinic seeds had a higher HIV prevalence (PR: 3.2, 95% confidence interval: 1.3 to 7.8) than contacts of community seeds, but contacts of HIV-uninfected clinic seeds did not (PR: 1.1, 95% confidence interval: 0.4 to 3.3). Results were similar when restricted to nonsexual contacts. To identify 1 new case of HIV, it was necessary to test 8 contacts of HIV-infected clinic seeds, 10 contacts of HIV-uninfected clinic seeds, and 18 contacts of community seeds.

Conclusions: Social contact recruitment by newly diagnosed STI patients efficiently led to new HIV diagnoses. Research to replicate findings and guide implementation is needed.

*Department of Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, NC;

UNC Project, University of North Carolina at Chapel Hill, Lilongwe, Malawi;

Department of Health Policy and Management, University of North Carolina at Chapel Hill, Chapel Hill, NC; and

§Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Correspondence to: Nora E. Rosenberg, PhD, MSPH, University of North Carolina Project, Tidziwe Centre, Private Bag A-104, Lilongwe, Malawi (e-mail:

Previously presented at (1) Rosenberg NE, Pettifor AE, Kamanga G, Bonongwe N, Mapanje C, Hoffman I, Martinson F, Miller WC. Social Networks of STI Patients have Higher STI Prevalence than Social Networks of Community Controls. STI & AIDS World Congress, 2013; (2) Rosenberg NE, Bonongwe N, Pettifor A, Kamanga G, Mapanje C, Hoffman I, Martinson F, Miller WC. Acceptability of a Social Contact Recruitment Program for Identifying HIV and STIs in Lilongwe, Malawi Consortium of Universities for Global Health Annual Meeting, March 2013; (3) Rosenberg NE, Kamanga G, Bonongwe N, Pettifor A, Mapanje C, Nkhata L, Rutstein SE, Hoffman I, Martinson F, Miller WC. Social Network Recruitment by STI Patients is a Promising Strategy for Identifying STIs and HIV in Resource-Limited Settings XIX International AIDS Conference, July 2012.

Supported by a 2010 developmental grant from the University of North Carolina at Chapel Hill Center for AIDS Research (CFAR), an NIH-funded program (P30 AI50410), and by an MP3 study (5-R01 AI083059-03).

N.E.R. was supported by NIH HIV/STD training Grant (T32 AI007001-34) and UNC Hopkins Morehouse Tulane Fogarty Global Health Fellows Program (R25 TW009340). S.E.R. was funded by F30 MH098731–01 and T32 GM008719. The remaining authors have no conflicts of interest to disclose.

Author contributions: G.K., W.C.M., N.E.R., and A.P. conceptualized the study under the guidance of I.F., and F.M., N.E.R., S.R., A.P., and G.K. developed data collection tools. M.W. developed the study database. G.K., N.B., and C.M. oversaw study implementation. N.E.R. conducted all analyses under the guidance of W.C.M. N.E.R. drafted the initial manuscript. All authors provided substantive edits to the manuscript and approved the final draft.

Received November 05, 2013

Accepted November 05, 2013

© 2014 by Lippincott Williams & Wilkins