Enhanced Peer-Group strategies to support prevention of Mother-to-Child HIV transmission leads to increased retention in care in Uganda: A Randomized controlled trial

Introduction: Despite scale up of Option B+, long-term retention of women in HIV care during pregnancy and the postpartum period remains an important challenge. We compared adherence to clinic appointments and antiretroviral therapy (ART) at different follow-up time points between enrolment and 24 months postpartum among pregnant women living with HIV and initiating Option B+ randomized to a peer group support, community-based drug distribution and income-generating intervention called “Friends for Life Circles” (FLCs) versus the standard of care (SOC). Methods: Between 16 May 2016 and 12 September 2017, 540 ART-naïve pregnant women living with HIV at urban and rural health facilities in Uganda were enrolled in the study. Participants were randomized 1:1 to the FLC intervention or SOC and assessed for adherence to prevention of mother to child HIV transmission (PMTCT) clinic appointments at 6 weeks, 12 and 24 months postpartum, self-reported adherence to ART at 6 weeks, 6 and 24 months postpartum validated by plasma HIV-1 RNA viral load (VL) measured at the same time points, and HIV status and HIV-free survival of infants at 18 months postpartum. We used Log-rank and Chi-Square p-values to test the equality of Kaplan-Meier survival probabilities and hazard rates (HR) for failure to retain in care for any reason by study arm. Results: There was no significant difference in adherence to PMTCT clinic visits or to ART or in median viral loads between FLC and SOC arms at any follow-up time points. Retention in care through the end of study was high in both arms but significantly higher among participants randomized to FLC (86.7%) compared to SOC (79.3%), p=0.022. The adjusted HR of visit dropout was 2.5 times greater among participants randomized to SOC compared to FLC (aHR=2.498, 95% CI: 1.417 – 4.406, p=0.002). Median VL remained < 400 copies/ml in both arms at 6 weeks, 6 and 24 months postpartum. Conclusions: Our findings suggest that programmatic interventions that provide group support, community based ART distribution and income-generation activities may contribute to retention in PMTCT care, HIV-free survival of children born to women living with HIV, and to the elimination of mother to child HIV transmission (MTCT).

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 4 51 Conclusions: Our findings suggest that programmatic interventions that provide group 52 support, community based ART distribution and income-generation activities may contribute 53 to retention in PMTCT care, HIV-free survival of children born to women living with HIV, 54 and to the elimination of mother to child HIV transmission (MTCT).
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023  taking ART among pregnant women that included financial constraints limiting access to food 73 and transport, and side effects of the therapy (12). Additionally, high dropout rates from 74 PMTCT care have been documented after delivery and at two years of follow-up among 75 mothers living with HIV (13-15), suggesting that service integration and linking mothers to 76 routine ART services were important determinants of retention in care. These findings 77 highlighted the critical need for innovative interventions to promote retention in HIV care and 78 adherence to ART for the successful implementation of option B + .
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Eligible and consenting women were then randomized 1:1 irrespective of the study sites to 118 SOC control arm or FLC intervention arm. The randomization list was computer-generated by 119 the data manager based at MUJHU using random-sized block groups that included consecutive 120 intervention numbers with corresponding random intervention assignments. Each pregnant 121 woman was assigned unique participant identification number and authors had no access to 122 information that could identify individual participants during or after data collection 123 After randomization, participants were assessed for socioeconomic status, ARV drug 124 adherence and stigma. A questionnaire addressing participants' experiences with 125 stigmatization and discrimination at family and community levels was administered at baseline, 126 12 months post-enrolment and at end of study. An additional needs assessment questionnaire . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. ; https://doi.org/10.1101/2023.04.15.23288495 doi: medRxiv preprint 8 127 was administered to participants enrolled in the FLC arm at enrolment, 1-year post enrolment 128 and every 6 months thereafter until the end of study. This was to assess individuals' 129 achievements and to document challenges related to group activities and participation in 130 income-generating activities (IGAs) and its benefits, skills acquired and training needs. Government to support the implementation of livelihood projects for community-based groups.

148
SOC arm participants received counseling from clinic PMTCT counselors and collected their 149 drug refills on an individual basis from the PMTCT clinics where they were enrolled as per . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023.  153 At each study site, all study participants were followed at a dedicated clinic for all scheduled 154 and unscheduled visits and were given transport reimbursement for scheduled study visits.  Participants were terminated from the study if they missed four or more consecutive scheduled 163 study visits, withdrew consent, relocated outside study area, or died.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  As a second primary outcome, we compared maternal ART adherence using self-report at 6 186 weeks and 6 and 24 months postpartum among all study participants. We validated self-187 reports by comparing VL measurements for all participants at 6 weeks, 6 and 24 months 188 postpartum.

Follow-up procedures
189 All participants were asked to self-assess the number of complete ART doses they missed in 190 the past three days. Self-reported ART adherence was then calculated as 1 minus the 191 proportion of self-reported doses missed in the past 3 days.

192
Laboratory measures of ARV adherence were used to validate self-reports using viral 193 suppression below 400 copies/mL. At the time of writing the study protocol, this limit was  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023.    Table 1 provides the number of participants, clinic visits made, and person-years of follow-up 243 by baseline socio-demographic characteristics of the study population stratified by study arm.

244
These results show that despite randomization, there were significant differences between the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. There was no statistically significant difference between the intervention and control arms in  Overall, retention in care was high across both study arms with 83.0% of all participants 257 remaining in care at the end of follow-up (Table 3). Significantly, more women were retained 258 in care at the end of follow-up in the FLC arm (86.7%) compared to the SOC arm (79.3%, p= 259 0.0221). Also, more women were terminated before the end of follow-up due to relocation in 260 the SOC arm (n=27, 10.0%) compared to the FLC arm (n=12, 4.4%). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. ; https://doi.org/10.1101/2023.04.15.23288495 doi: medRxiv preprint  Table 4 shows that participants self-reported optimal adherence to taking >95% of their ART 273 medication in the last 3 days at their 6 weeks, 6 and 24 months postpartum visits. There was 274 no statistically significant difference in optimal adherence to ART between the arms. The 275 median viral loads of participants at each of these time points were <100 copies/ml with no 276 statistically significant differences between the two arms.    p=0.002). Participants enrolled from rural health facilities and those aged 15-24 years were 291 also significantly less likely to be retained in care compared to their urban (p=0.021) or their . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023.

313
This is comparable to the study by Masereka et al, which found that 87.9% of women were 314 retained in care after being initiated on Option B+ in Uganda (26). Our data is also consistent . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 17, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 16 315 with a study from Malawi which found that at 24 months after initiation of Option B+ ART, 316 retention in the peer group models was 80% and 83% in the facility-based and community-317 based arms respectively, compared with 60% in the standard of care arm (17). This may explain 318 the absence of a statistically significant difference between the intervention and control arm 319 when it came to adherence to PMTCT clinic appointments.

320
In both arms of this RCT, 87% or more participants self-reported optimal adherence to taking 321 their ART at 6 weeks, 6-and 24-months postpartum and the median viral load levels of 322 participants were under 100 copies/ml at each of these time points with no significant difference 323 between the FLC and SOC arms. We did not find fluctuations in longitudinal adherence during 324 the postpartum period. is comparable to our finding. Our finding is also consistent with a cohort study done in Malawi 328 that found 90%-100% adherence to ART among women on Option B+ observed from 4 to 21 329 months postpartum (28). That study used pharmacy records to measure adherence as opposed 330 to self-reports and VL validation which may limit comparisons with our study.

331
Although we found high self-reported adherence to ART that corroborated with low median 332 viral loads, we did not find significant differences in these outcomes between the FLC and 333 SOC arms at any follow-up time points. The close correlation between these outcomes echo 334 the findings of a study among 452 women on PMTCT care and treatment in South Africa where 335 it was found that a raised viral load was consistently associated with lower median adherence 336 scores (29). However, the lack of a more distinct outcome from the FLC intervention may result 337 from a study effect or still a 'contamination' between the FLC and SOC arms among both study 338 staff and participants in the study facilities (30).
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Our study's main limitation is that our RCT was individually and not cluster randomized, which 349 may have resulted in cross-contamination between the FLC and the SOC arms in the same 350 clinics over the long-term follow-up period. Contamination could have also occurred when 351 study staff attended to both FLC and SOC participants at the same clinic. Additionally, all 352 study participants were given transport reimbursement for their study visits which may have 353 positively influenced adherence to study visits. Lastly, the feasibility of our intervention would 354 depend on its cost and cost-effectiveness, which we did not assess.

356
Our study found that the FLC intervention significantly increased retention in care and HIV-

362
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 17, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 18 363 364

365
The authors declare that they have no conflict interests. We would like to thank the FLC for Option B+ study team who were vital in the collection of 373 this data and the study participants without whom this study would not have been possible.

378
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 17, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023