Community-based distribution of oral HIV self-testing kits: experimental evidence from Zambia

Publication Details

Hensen, B, Ayles, H, Mulubwa, C, Floyd, S, Schaap, A, Chiti, B, Phiri, M, Mwenge, L, Simwinga, M, Fidler S, Hayes, R, Bond, V and Mwinga, A, 2018. Community-based distribution of oral HIV self-testing kits: experimental evidence from Zambia, 3ie Impact Evaluation Report 86. New Delhi: International Initiative for Impact Evaluation (3ie)

Link to Source
Bernadette Hensen, Helen Ayles, Chama Mulubwa, Sian Floyd, Ab Schaap, Bwalya Chiti, Mwelwa Phiri, Lawrence Mwenge, Kwame Shanaube, Musonda Simwinga, Sarah Fidler, Richard Hayes, Virginia Bond, and Alwyn Mwinga
Institutional affiliations
London school of hygiene and tropical medicine, The Zambia AIDS Related Tuberculosis (ZAMBART)
Grant-holding institution
None specified
Sub-Saharan Africa (includes East and West Africa)
Health Nutrition and Population
Gender analysis
Gender analysis
Equity Focus
None specified
Evaluation design
Randomised Control Trials (RCT), Mixed Methods
3ie Series Report
3ie Funding Window
HIV self-testing thematic window


Community health workers providing door-to-door testing services through a universal test and treat intervention additionally offered HIV self-tests to assess whether this added choice increases knowledge of HIV status.


HIV testing coverage gaps remain, even in the context of universal test and treat, and door-to-door offers of testing. Some residents do not want a health care provider testing them, and others are never found at home. Men are particularly difficult to reach.

Research questions

  • Does the choice of HIV self-testing increase knowledge of HIV status among residents of communities participating in a universal test and treat intervention?
  • Does the choice of HIV self-testing increase the probability that residents will accept participation in the PopART (universal test and treat) intervention?
  • Does the choice of HIV self-testing increase accepting the offer of HIV testing services among those who do not currently know their HIV status?

Qualitative questions:

  • What factors influence individuals or groups to choose HIV self-testing over the standard provider finger-prick test?
  • Are any social harms associated with choosing the HIV self-test kit?
  • How are HIV self-test kits managed (handled, stored, interpreted, disposed, transported)?
  • What is the impact of HIV self-testing on the role of the community health workers?
  • What is the incremental cost-effectiveness ratio of adding HIV self-testing to the PopART universal test and treat intervention?


Intervention design

The intervention consisted of adding the choice of oral HIV self-testing to the standard health worker-provided HIV finger-stick test. In the PopART universal test and treat intervention, community health workers (called CHiPs) go door-to-door annually to offer HIV testing services. In this round, HIV self-tests were also offered as an option by randomly selected CHiPs. Residents could choose to get tested for HIV as well as decide which method of testing if they did choose to test. For residents who were not present, their partner or family member could agree to receive a self-test and give it to them later. All self-tests that were left behind or performed unsupervised were followed up by the CHiP to see if they were used, if the user was willing to share the result for counselling, and to assist in linking to care as needed. Linkage assistance and counselling were also provided for standard testing services.

Theory of change

The authors theorised that the choice of testing methods, and the option to receive a kit or use a kit outside the presence of a health worker, would increase testing rates overall. The theory is based on the idea that often people choose not to test because they have confidentiality concerns, do not like needles, do not have time, find it inconvenient, or are afraid of finding out the result (among other reasons). HIV self-tests offered at one’s home can overcome these barriers by allowing users to test in private, at a time of their choosing, offering an oral method without the need for a finger stick, and can increase self-efficacy by allowing the user to perform and interpret the test oneself. In addition, secondary distribution, by leaving a kit for household members who are not present, may be more convenient with the possibility of reaching individuals, especially men, who are more frequently absent. Increased testing would therefore result in increased knowledge of HIV status.

Evaluation design

The study was implemented in four communities in the PopART universal testing and access to treatment arm. The four communities were divided into 33 operational zones, with CHiP pairs assigned to each zone. Using a cluster-randomised trial design, half of the 33 zones were randomly assigned to treatment—the CHiPs in those zones would offer HIV self-tests as an option. The proportion of individuals who either tested or self-reported HIV+ was then compared between the two arms, along with secondary outcomes. Effects were adjusted for clustering at the zone level. Researchers used a logistic regression model, adjusting for age, sex and community, and whether the individual was previously resident in the community. They additionally assessed whether there was effect modification by these groups using an interaction term, and obtained p-values using a Wald test.

Main findings


  • The choice of HIV self-testing slightly increased knowledge of HIV status compared to those only offered standard testing (adjust OR = 1.3, p=0.03).
  • The effect differed by sex (p=0.001). The effect was almost entirely driven by men (aOR=1.31, p=0.009). No effect was seen for women (aOR 1.05, p=0.62).
  • Younger people were also more influenced (aOR=1.31 p=0.02), with a weaker effect among older people (p=0.07), although the interaction effect was not significant (p=0.44).
  • The HIV self-test intervention was also most effective in community 2, which had a higher concentration of middle class residents.
  • While linkage to care among self-testers was generally lowest for those who were not seen by a CHiP, it could not be verified whether they received a confirmatory test elsewhere. However, overall linkage to care rates were comparable between self-test and non-self-test zones (41.0% versus 41.2%) at three months.

Cost efficiency analysis

The cost per person tested was $22.06 in standard testing zones and $30.17 in self-test zones. The cost per new tester was $96.89 in standard zones and $102.72 in self-test zones. The incremental cost per additional person tested in HIV self-testing zones was $255.98.

Implications for policy and practice

The door-to-door offer of a choice of whether and how to test for HIV, which included HIV self-testing, increased knowledge of current HIV status among the general population, but largely among men. Secondary distribution reaches men not easily contactable by community health workers.

With time, as self-testing becomes less novel, more people who opted for supervised self-testing may opt for finger-stick tests, which may be more cost-effective by improving targeting for HIV self-testing. HIV self-testing should ideally be reserved for those who are otherwise unwilling to access or accept finger-prick tests.

Careful and detailed communication and information should accompany distribution of self-test kits to ensure proper use and promote linkage to care.

Additional publications

Media coverage

Scroll to Top