We conducted a cross-sectional observational study nested within a large ongoing longitudinal parent study among PWID and their partners in Kenya, the protocol for which was recently published . The parent study, entitled “SHARP,” enrolls HIV-positive people who inject drugs as “index” participants and then employs assisted partner services to identify, test, and link to care the index’s sexual and injecting partners. Inclusion criteria for SHARP differed based on whether individuals were enrolling as index participants or as partners. Index participants in SHARP were HIV-positive PWID 18 years of age or older who had injected drugs at least once in the past year. Inclusion criteria for partners in SHARP were individuals 18 years of age or older who had engaged in sexual intercourse with index cases at least once in the previous 3 years, or who had injected drugs in the presence of index cases at least once in the previous 3 years. All index participants and partners are tested for HCV at baseline enrollment using RDTs for HCV antibodies.
The nested study we describe herein was a cross-sectional sub-study that recruited from any participant newly enrolling as an index or as a partner participant in SHARP.
Setting & population
Participant recruitment and data collection took place at multiple settings located in Mombasa and Kilifi counties. Study procedures took place in partnership with three organizations that operate drop-in centers and provide harm reduction (needle and syringe programmes), social services, and HIV testing and care services to PWID on Kenya’s coast: ReachOut in the city of Mombasa, Muslim Education and Welfare Association in Mtwapa and Kilifi towns, and Omari Project in Malindi town. Outreach services are conducted by community-embedded peer educators who are former drug users.
Participants in this study were individuals aged 18 years or older who were either HIV-positive people who had injected drugs in the previous year, or were individuals aged 18 years or older who had had sexual intercourse or had injected with an HIV-positive person who injects drugs in the previous 3 years... The target sample size for the nested study was 150 participants, calculated to detect at least 50% HCVST acceptance rate with a 10% margin of error.
Sampling & recruitment
Individuals considered for participation this study were identified during screening and enrollment for SHARP, the parent study. Any individual newly enrolling in SHARP was considered for enrollment in the HCVST sub-study on a first-come first-served basis between August and November, 2020. Potential participants were informed of a secondary study to evaluate self-testing for HCV by study staff, and those participants who were interested in enrolling in the sub-study underwent a secondary screening process immediately following screening for the parent study.
Inclusion & exclusion criteria
In addition to meeting the eligibility criteria for the parent study, participants in this study were excluded if they had previously ever tested positive for HCV antibody, or if they had tested negative for HCV antibody within the past year. They were also ineligible if they had ever completed a self-test for HIV or HCV. The purpose of these criteria was to exclude any potential participants who might be biased in their interpretation of the HCVST results or acceptance of HCVST, either due to previous testing or due to previous experiences with self-testing. While participants in the sub-study were required to have also enrolled in the parent study, participation in the sub-study was not a requirement for enrollment in the parent study.
Following screening, eligible participants provided written informed consent to participate in both the parent study and the sub-study. Participants received 400 Kenya Shillings (KSh, about $4) as reimbursement for time and travel for enrollment into the parent study, and an additional KSh 200 (about $2) for time spent participating in the HCVST sub-study. Enrolled participants completed a baseline questionnaire that covered socio-demographic characteristics, sexual and injecting risk behaviors, and previous experiences with HIV and HCV testing and care. Sexual and injecting risk behavior questions included questions about ever engaging in unprotected anal intercourse, any injecting unprescribed drugs, any needle sharing, surgical or dental procedures, tattoos, and sharing shaving equipment.
Following the baseline questionnaire, participants performed the HCV self-test while being observed by study staff in private rooms within the drop-in centers. We used prototype self-testing kits that included a professional use OraQuick® HCV Rapid Antibody Test (OraSure, Inc., US), a plastic stand and instructions for use (IFU) adapted for self-testing by the test manufacturer. Each participant was given the kit and told to follow the procedures written on the IFU. Instructions were mostly pictorial, but included key words or phrases written in the local language Swahili (Additional file 1: Fig. S1). Study staff noted both errors and difficulties with self-testing steps according to a standardized checklist, while observing the participant complete the self-testing procedure. The testing checklist included questions on pre-testing steps (opening the pouch, removing contents, reading instructions, removing the test tube from its package, removing the cap from the test tube, placing the test tube in the stand, and removing the test device), testing steps (correct handling of the test device without touching the flat pad, collection of oral fluid specimen, placing the test device in the test tube and keeping time correctly) and test interpretation steps (interpreting the results correctly).
Participants were encouraged to perform the self-test on their own without assistance. If participants had trouble with steps in the self-testing procedure and requested assistance, study staff provided assistance only after the participant had made multiple attempts at completing the step, usually after 15 min. Staff also noted which steps participants required assistance with. Self-testing results were first read and interpreted by the participant, and then the same results were read and interpreted by the study staff member, and both interpretations were recorded in a test result form.
Participants who self-tested also completed a questionnaire to ascertain opinions and experiences of the self-testing procedure. Questions included the participants’ rating of the test’s ease of use, their willingness to use the test again or to recommend it to family or friends, preferred testing modalities (oral fluid or blood-based), and their understanding of actions they need to take if they tested positive.
Finally, all participants were tested by a trained healthcare worker with the OraQuick® HCV Rapid Antibody Test for professional use. Results from this second test were read and interpreted by study staff, communicated to the participants, and recorded in the test result form. Post-test counselling was provided to all participants, and those found to be positive were linked to confirmatory RNA testing and HCV treatment as part of routine SHARP procedures.
Data management & analysis
Data were all collected electronically and kept in two separate databases. The baseline questionnaire was administered by trained study staff using the Open Data Kit platform on handheld tablets, and these data were stored in the SHARP database on servers within Kenya’s Ministry of Health (MoH). All other data including the self-testing checklist and post-testing questionnaire were collected through OpenClinica, an online data collection system, and stored in the OpenClinica database. Variables of interest from the baseline questionnaire were extracted from the MoH server and merged with the OpenClinica dataset. Data cleaning and analysis were performed using STATA v.14 (Texas, USA).
Usability was assessed through a combination of percentage of observed errors and difficulties and calculation of inter-reader and inter-operator concordance. Baseline characteristics and errors, difficulties, and assistance required in self-testing were compiled and analyzed using simple frequencies and percentages. Errors, difficulties and assistance variables were also compounded into aggregate variables describing whether any errors (problems or omissions noted with any step on the checklist), difficulties (troubles with steps that did not cause errors) or assistance (help provided with any step and noted in the checklist) occurred. Inter-reader concordance was calculated as the percentage agreement between the participant’s and the staff’s interpretation of the participant’s self-test result, including invalid results but excluding participants who received assistance reading the results. Inter-operator concordance was calculated as the percentage agreement between the participant’s interpretation of the self-test result and the staff’s interpretation of the professional use test result, excluding invalid results and excluding participants who received assistance reading the results. Cohen’s Kappa coefficient was calculated for both. Acceptability was assessed through percentages of self-reported opinions and experiences of the self-testing procedures.
About half-way through enrollment in the study, study staff noted that individuals who were experiencing symptoms or showing signs of opioid withdrawal appeared to have greater difficulty with self-testing procedures. After this observation, the baseline questionnaire was modified to include a question on whether each person was experiencing symptoms or showing signs of withdrawal. We then conducted Pearson chi-square and Fisher’s exact tests to assess whether experiencing withdrawal symptoms was associated with making errors, experiencing difficulties, or requiring assistance during the testing process, respectively. Fisher’s exact test was used when the expected frequency was less than 5 in any given category, whereas Pearson chi-square tests were used for the remaining comparisons.