Evaluating Demand Creation Strategies for Voluntary Medical Male Circumcision (VMMC) in Kenya

STUDY AIMS Aim 1: Assess the rate of uptake of VMMC services by men aged 25-39 years exposed to Inter-Personal Communication (IPC) vs. Designated (older men only) Service Outlets (DSO) vs. IPC and DSO interventions combined vs. no intervention.

Aim 2: Determine through testing of unlinked de-identified blood from bleeding vessels from the surgical wound, the proportion of men whose HIV status is unknown who are HIV infected (Sub-study 1).

Aim 3: Determine the cost of providing VMMC among the three interventions relative to each other and to no intervention.

MAIN STUDY DESIGN: The study proposes to conduct a cluster randomized-controlled trial, with Locations forming the unit of randomization. Within the selected Locations, the study shall randomly select approximately 16% (209) of the villages which will be sufficient to generate the required sample size.

Sample size Determination:

Sample size for detecting the impact of the interventions on VMMC uptake - in order to be able demonstrate at least a 33.3 % increase in the rate of VMMC among young adult men in intervention arms when compared to those in control arm with 80% statistical power, and Bonferroni adjusted two-sided α=(0.05/3)=0.017 following the implementation of study interventions, the study will need a sample size of 4,932 men reached with the intervention.

Potential participants to be reached - An effect size equal to 33.3% increase in MC from 30% to 40% will require about 1233 participants per study arm (total 4932), while an effect size equal to 50% increase in VMMC from 30% to 45% will require 573 participants per study arm (total 2292). Adjusting for a refusal rate of 15% and a loss to follow up of 15%, suggests contacting 1603 per study arm resulting in a total of 6412 uncircumcised men aged 25-39 years.

PHASE 1: HOUSEHOLD ENUMERATION To ascertain the size of eligible population (uncircumcised men aged 25-39 years), the study will conduct complete household listing in all selected villages and enumerate male members in each household aged 25-39 years. During this exercise, study staff shall physically visit all homes in the study villages and assign a unique identification number to every household (HIN). Enumeration shall be through home visits ONLY and not by any other means. The unique HIN assigned during the enumeration will be entered on Excel database and households with men aged 25-39 years will be sorted out and put on a separate list. The list will be used to create a database of eligible men in each village. Determining the number of households with eligible men will help in knowing the denominator and in planning for recruitment.

PHASE 2: CONDUCTING RAPID FORMATIVE STUDY (Sub-study 1) Justification: This phase is intended to provide information for use in delivering the interventions. The study have already drafted messages based on existing literature for use in IPC and IPC+DSO intervention Locations/villages. However, since almost all the available information on barriers to and facilitators of VMMC in literature were collected among men of all ages and were not disaggregated by age group, the investigators cannot assume they are automatically relevant to older men. As such, it will still be necessary to conduct a rapid formative study to explore reasons why older men aged 25-39 years go or do not go for VMMC services. Information obtained from this phase will be used to modify if indicated, the messages developed from existing literature.

SUB-STUDY DESIGN: This phase shall comprise of Focus Group Discussions (FGDs) and In-Depth Interviews. Twelve FGDs, with 6-10 participants, will be held with six groups of circumcised men and six groups of uncircumcised men in each of the following age categories: 25-29 years, 30-34 years and 35-39 years. A similar number of FGDs will be held with female partners of circumcised and female partners of uncircumcised men in same age categories.

Besides FGDs, the investigators will conduct 48 IDIs to capture confidential information that some men and women would otherwise not divulge in an FGD setting.

PHASE 3: ESTABLISHING BASELINE VMMC PREVALENCE AND PROPORTION OF ELIGIBLE MEN Based on the eligible participants' list, research assistants will only go back to households with men aged 25 to 39 years. Before administering enrollment consent, RAs will ask participants to identify a private location within or outside the house where they can do consenting and interviewing. The RA will then administer consent for participating in the study, enrol consenting participants and assign Participant Unique ID number.

After assigning the participant unique ID number, the RA will conduct the baseline interview, verify the MC status and inform participants of their intervention arm and deliver the intervention if the participant is available (if not, another visit will be made at a later date to deliver the intervention). During the interview, study staff will document participants' circumcision status from the verbal report. After the interview, participants who consented to physical verification will be asked to choose where this can be performed.

At enrollment, participants will be given 3 months within which to go for circumcision, Participants can go whenever they want for VMMC, but the study will only capture their data if they go within 3 months. They will also be informed that a reminder of the expiry date for circumcision will be sent through text message to all participants one month before the expiry of the circumcision window period.

Details of participants enrolled in the study will be entered into an excel Master Log which will contain all participant details; 3 names, ID number, age, location, village, the telephone number of the participants or of close friend and date of enrollment.

PHASE 4: IMPLEMENTING THE INTERVENTIONS The study has four intervention arms: Inter-Personal Communications (IPC), Dedicated Service Outlets (DSOs), a combination of IPC and DSO (IPC+DSO), and Control. The interventions and how each of them will be implemented are described under 'Protocol arms and interventions' section below.

HIV TESTING OF PARTICIPANTS WITH UNKNOWN HIV STATUS (Sub-study 2) In VMMC program undertaken by Impact Research and Development Organization (IRDO) and in other Kenyan programs as well, between 10% and 30% of men decline HIV testing. Given that VMMC is offered principally as an HIV prevention strategy, it is important to determine the HIV status of those who decline testing, and if the prevalence is significantly higher than those who accept testing, there will be need to design approaches to improve uptake of testing. This is particularly important following results from a recently concluded study that indicated that 44% of HIV-infected men resumed sex before the recommended 6 weeks of sexual abstinence.

Sample size for HIV prevalence among uncircumcised men whose HIV status is unknown (Sub-study 2):

Investigators estimated that they will need to test randomly selected unlinked blood samples from 163 men per arm whose HIV status is unknown in order to estimate a HIV prevalence that falls within 10% points of the true population prevalence with 95% confidence. The study staff will collect samples from all of men who refuse testing but consent to Unlinked HIV testing.

With a refusal rate of 23.6% voluntary HIV testing, and a consenting rate of 60%, then the rate of getting a man who refuses testing but consents to the testing of unlinked blood sample from his bleeding vessels is 14.2%. In other words, of the 4932 participants in the main study, investigators expect 1164 (4932*0.236) to refuse HIV testing. With assumed unlinked HIV testing acceptance rate of 60%, the study can obtain 700 such men from a sample size of 4932 in the main study. For a design effect of 2.5, using a sample size of 700 for the sub-study will enable us to estimate the HIV prevalence of 21.6% with a precision of 4.8%=sqrt[(2.5*1.96^2*.216*.784)/700].

Methods/Procedure:

RAs will perform unlinked testing of blood from bleeding vessels or finger prick of those whose HIV status is unknown. While the call for proposals recommended testing blood from the excised foreskin, investigators are concerned that they may not be able to harvest sufficient blood from most foreskins to perform the test. This is because the forceps-guided method being used in Kenya crushes the foreskin and compresses the blood vessels that supply the distal section. This results in little or no bleeding on the excised foreskin. The study, therefore, propose to only collect blood samples from the bleeding vessels or finger prick from participants (n=700).

COSTING VMMC IN THE CONTEXT OF THE INTERVENTIONS (SUB-STUDY 3) Study Design: Data on costing will be collected across all the study arms. Whereas economic evaluation of service delivery has been conducted in several contexts there is limited information on the cost-effectiveness of various approaches used for recruitment to increase service uptake. The study plan to assess the cost-effectiveness of demand creation for VMMC services based on the three interventions. The results will demonstrate the accrued costs and net savings associated with each intervention strategy.

Methods: The primary outcome of the main study is the incremental change in circumcisions performed attributable to the respective demand creation interventions. Costs will be estimated for primary outcome with reference to the control group, in which the study will continue to provide routine recruitment approaches. This will demonstrate how the costs vary depending on the mode of demand creation.

The study shall adopt a societal perspective because of the ability to integrate diverse factors into the model and the role the public sector plays in HIV prevention and management. This approach incorporates all costs borne by the implementers in recruiting participants for services.

Sources of costing data:

The study will collect data from all study sites using a tool adapted from the Male Circumcision Decision-Makers' Program Planning Tool (DMPPT) for each of the demand creation strategies. The DMPPT was developed by the USAID/Health Policy Initiative in collaboration with UNAIDS to enable decision-makers to understand the potential cost and impact of various options for scaling up male circumcision services. The DMPPT has previously been used to undertake costing of male circumcision services in East and Central Africa. The study shall use this instrument as a data collection tool to estimate the cost of demand creation and the unit cost of providing VMMC (i.e. cost per adult circumcised) in each of the four study arms.

PHASE 5: ESTABLISHING ENDLINE MC PREVALENCE/ESTIMATING IMPACT OF INTERVENTIONS At Endline, the study will review Participant Master Log to identify those who did not go for VMMC at the 45 VMMC clinics participating in the study. The study will use the telephone numbers obtained during enrollment to contact the participants. Those who say that they got circumcised but are not in the Participant Master Log, study staff will make an appointment and visit them at home, or another convenient location, to administer consent, conduct Endline interview and verify their circumcision status. Those who report not having gone for VMMC will be asked why they did not go, and their response recorded.