A savings intervention to reduce men’s engagement in HIV/STI risk behaviors

In Kenyan communities with high prevalence of HIV and other STIs, participants will be randomized to two groups: a control group that receives basic health education or a savings intervention group that receives a mobile banking account with incentives to save. The savings intervention group will receive bank accounts and (a) lottery-based incentives to save in the bank account, (b) opportunities to develop savings goals and strategies, and (c) periodic reminders about the savings incentives and goals.

We will develop a partnership with a local bank with branches in the study communities. Participants will receive assistance from bank staff and research assistants in opening and using a savings account. They will also receive an education session that emphasizes the importance of saving for the future, provides a tutorial on how to earn interest on savings, and addresses common barriers to saving (lack of trust, lack of information, etc.). We will use educational materials developed in our pilot study, with modifications based on participant feedback. In addition, participants will receive 500 Shillings from the study team as an initial deposit into their accounts.

At the time of enrollment, research assistants will discuss with participants their reasons for saving, their savings goals, and the challenges to saving money each month. The research assistants will encourage participants to develop a savings goal amount and to define a purpose(s) for saving (e.g. future investment related to farming or businesses, household expenses on durable goods, children’s education, etc.) and concrete strategies to save, including developing a time frame and amount needed to save to achieve investment goals. This activity is based on prior research demonstrating that setting savings goals and implementation plans can increase the motivation to save.

Participants will be told about lottery-based incentives for saving money in their account. Specifically, they will have a chance of winning monetary prizes each month based on the amount by which their account balance goes up during the month. For example, as in our pilot study, for every 100 Kenyan Shillings (~US$1) of new savings each month, participants will receive 1 ticket for the monthly lottery. Each month’s lottery will have several prizes, including a limited number of larger value prizes and many smaller prizes of 500 Shillings. Prizes will be deposited directly into bank accounts and winning participants will be informed by text message. In addition, participants will receive text messages each month reminding them of savings goals and strategies, as well as the lottery-based incentives for saving. Participants who do not win any prize in a given month will receive text messages encouraging them to save and reminding them of the goal they set.

2022-07-22

2025-04-30

Composite incidence of HIV and other STIs over 24 months

The primary outcome is HIV/STI incidence, defined as the composite incidence of at least one of HIV or other STIs (herpes simplex virus type 2 (HSV-2), Neisseria gonorrhoeae (NG), and Chlamydia trachomatis (CT)) over 24 months. The combination of all STIs (including HIV) into a composite variable is justified as the risk factors for these STI’s are similar. Compared to a strategy of considering the individual infection as separate variables, using the composite variable (any HIV/STI infection) should increase the number of events and hence our ability to detect an effect of the intervention. For participants who are HIV-positive and/or HSV-2-positive at baseline, the outcome will be defined over all STIs other than HIV and/or HSV-2.

• The total amount of savings and investment
• Total formal savings, and total household assets
• Investment behavior such as housing improvements, and schooling expenditures
• Expenditures on food and household needs
• Expenditures on alcohol and transactional sex
• Expenditures on gambling
• Self-reported sexual behavior, including number of transactional sex partners, total number of sexual partners and condom use.
• Alcohol use
• Survey-based measures of future orientation
• Survey-based measures of stress
• Perpetration of intimate partner violence

Alcohol use will be measured using the AUDIT-C scale. Future orientation will be measured using the Cantril scale. Stress will be measured using Cohen's Perceived Stress scale.

The Akiba study is a single-center two-arm parallel group trial designed to test the superiority of a behavioral intervention over control. To be eligible to participate in the study an individual must be a male aged 18-39 with a mobile phone and steady cash income source that results in earnings every week who also resides in study community and plans to remain for the next 2 years. The individual must have used alcohol or other substances in the past month and engaged in transactional sex (defined as payment of money, goods, or services in exchange for sex) in the past 3 months. And finally, they must have or be willing to open a bank account with partner financial institution, have a national identification card, and have or be willing to create a Kenya Revenue Authority PIN (the latter two criteria are required for opening bank account).

Eligible individuals who initially agree to participate will provide written informed consent before enrollment in the trial. This will be followed by a questionnaire and HIV/STI testing. Participants will be stratified on baseline HIV status, and within each stratum they will be randomized (1:1) to an intervention or control group using a computer-generated sequence.

Participants enrolled in the study will be administered a questionnaire at baseline, 6, 12,18, and 24 months. We will evaluate the
presence of HIV, NG, and CT at baseline, 12 months, and 24 months. We will also test for HSV-2 at baseline and 24 months.

We will use a computer-generated sequence with blocks of sizes of 10 that will be generated by the investigators before the start of enrollment. Treatment allocation will be revealed to participants and research assistants after completing the baseline questionnaire and obtaining the results of their HIV test by opening a sealed envelope that corresponds with the results of their HIV status (one set of envelopes for those that test negative and one for those that test positive/indeterminate). Each envelope will contain a sheet of paper that states whether a participant is in the control group or intervention group.

Individual

No

We will not be using clusters

1,500 participants

Roughly 750 in control group and 750 in intervention group

With the expectation that annual HIV/STI incidence in the control group will be 15-20% (based on existing literature), a relative reduction of 21-25% in composite HIV/STI incidence is detectable with 80% power (two-sided, α=0.05) for ~1,500 participants (~750 per group). We included a conservative assumption of 20% attrition over 24 months (expected to be lower given our prior studies). Calculations are based on a two-group Chi-squared test of proportions.

SAP, version 1