The POWWeR Health Study included additional 30- and 36-month visits. Below is a summary of all study procedures as per the protocol. All study procedures ended on 30 August 2026.
Enrollment
A total of 473 participants were initially screened (against 503 expected) and 407 were enrolled at baseline (against 402 expected). Enrollment by month during the enrollment period was as follows (Table 1).
| Month, dates | Expected enrollment | Actual enrollment |
|---|---|---|
| Month 1, Feb 9 - 28 | 67 | 41 |
| Month 2, March 1 - 27 | 67 | 51 |
| Month 3, April 3 - 28 | 67 | 18 |
| Month 4, May 10 - 31 | 67 | 51 |
| Month 5, June 1 - 30 | 67 | 50 |
| Month 6, July 1 - 31 | 67 | 52 |
| Month 7, Aug 1 - 30 | - | 84 |
| Month 8, Sept 1 - 30 | - | 60 |
| Total | 402 | 407 |
Follow-up and study forms completed
Completed database forms by time point (Table 2):
| Database form | Baseline | 6 months | 12 months | 18 months | 24 months | 30/36 months |
|---|---|---|---|---|---|---|
| Screening and eligibility | 473 | X | X | X | X | X |
| Locator form1 | 407 | X | X | X | X | X |
| Behavioral survey | 407 | 332 | 332 | 331 | 347 | 315 |
| Medical exam form | 407 | 332 | 332 | 331 | 347 | 315 |
| Clinical diagnosis and treatment form | 407 | 3 | 332 | 6 | 347 | 0 |
| Lab requisition and results | 407 | 326 | 332 | 331 | 347 | 315 |
| HIV C and T form | 407 | 0 | 332 | 0 | 347 | 0 |
| Withdrawal form (prior to the stated visit) | - | 11 | 6 | 6* | 11 | 2 |
| Etiological diagnosis and treatment form | 147 | 95 | 167 | 115 | 179 | 146 |
| WASH assessment | 47 | 0 | 0 | 0 | 49 | 0 |
1 Contact information is confirmed and updated as needed at each study visit.
Retention and withdrawals
Early withdrawals from the trial are summarised below (Table 4).
| Stage and reasons | Withdrawn from study, N |
|---|---|
| Screen failure: after consent, ineligible (gave incorrect birthdate n=1; disclosed amenorrhea n=4; declined pelvic exam n=2; died n=1) | 8 |
| After baseline, before 6 months (no longer interested n=2; relocated n=8; incarcerated n=1) | 11 |
| After 6-month follow-up, before 12 months (no longer interested n=4; relocated n=2) | 6 |
| After 12-month follow-up, before intervention (co-enrolled in another study n=1) | 1 |
| Completed 12-month follow-up and withdrawn by investigators from intervention2 | 22 (withdrawn from intervention only, continuing in study) |
| After completion of intervention and before 18-month follow-up (co-enrolled in another study n=1; inserted an IUD n=2; no longer interested n=1; relocated n=1) | 5 |
| After 18-month follow-up and before 24-month follow-up (no longer interested n=8; relocated n=1; died n=1; hysterectomy n=1) | 11 |
| After 24-month follow-up and before 30/36-month follow-up (death n=2) | 2 |
| After 30/36-month follow-up (death n=1, prior to treatment) | 1 |
2 Reasons withdrawn from intervention: taking part in a PrEP cervical ring study (n=2); currently pregnant (n=14); recently delivered (n=6).
Protocol deviations and safety
Table 5 records a total of 243 protocol deviations, with 150 reported to the IRB. There were no enrollment issues, no serious adverse events that went unreported beyond 48 hours, no duplicated participant IDs that were not corrected, no missing forms, no delayed sample results and no impersonated visits. Eighty-five cases of wrong treatment given for BV were reported in full to the IRBs, NIH and the DSMB. Six instances of intervention delayed by more than four weeks were reported to the IRB, and two instances of appropriate specimens or tests not taken are noted as not reported. Minor deviations are being batched for report to the IRB.
On adverse events (Table 9), no vaginal or cervical lacerations and no menstrual toxic shock syndrome were recorded at any visit. No partner violence attributable to cup use (refused to pay, paid less than agreed, verbal assault, physical assault) was reported at the 13, 14, 15, 18 or 24 month time points.
Baseline profile of participants
Baseline characteristics (Table 6): median age 27.3 years (IQR 23.7-31.1); 365 (89.7%) Luo, 42 (10.3%) other; 241 (59.2%) had less than secondary school education and 166 (40.8%) secondary school or more; 238 (58.5%) reported sex work only and 169 (41.5%) sex work plus other work; 283 (69.5%) had a non-paying male partner.
Baseline laboratory and clinical findings (Table 7): HSV-2 seropositive 253 (62.8%); composite STI 90 (22.1%), comprising Trichomonas vaginalis 30 (7.4%), Neisseria gonorrhoeae 22 (5.4%) and Chlamydia trachomatis 53 (13.0%); bacterial vaginosis 171 (42.0%); HIV 100 (24.6%). Community state type IV (mixed, G. vaginalis dominated) was the most common vaginal microbiome profile at 255 (62.8%).
Outcomes
Primary and secondary outcomes measured every six months (Table 10):
| Measure | Baseline | 6 months | 12 months | 18 months | 24 months |
|---|---|---|---|---|---|
| BV prevalence | 171 (42.0%) | 144 (42.9%) | 145 (43.7%) | 135 (40.8%) | 153 (44.5%) |
| Proportion with prior BV | - | 71 (49.7%) | 108 (74.5%) | 115 (85.2%) | 123 (80.4%) |
| STI composite prevalence | 90 (22.1%) | - | 78 (23.4%) | - | 68 (19.8%) |
| STI composite incidence | - | - | 71 (21.3%) | - | 61 (17.8%) |
| CST-I (L. crispatus dominated) | 32 (7.9%) | 16 (4.8%) | 25 (7.5%) | 13 (3.9%) | 33 (9.5%) |
| Mean relative abundance of L. crispatus (SD) | 7.2 (23.7) | 4.7 (18.5) | 5.6 (20.3) | 3.3 (15.0) | 8.2 (24.9) |
For the individual STI outcomes, prevalence at 24 months was 24 (7.0%) for Trichomonas vaginalis, 17 (5.0%) for Neisseria gonorrhoeae and 37 (10.8%) for Chlamydia trachomatis. Infections following previously treated infections are included; infections for which prior treatment is not documented are not counted as incident infection.
Menstrual health practices
Reported unsafe menstrual management practices declined across the study (Table 11). Insertion of objects into the vagina for sex during menses (for example cotton wool, tissue, sponge) fell from 23 (5.6%) at baseline to 2 (0.6%) at 18 months and 1 (0.3%) at 24 months. Reported use of drying or tightening agents inside the vagina for sex during menses fell from 71 participants answering "always" at baseline to none at 24 months, with 343 (98.9%) answering "never" at 24 months. Wiping inside the vagina with cloth, tissue, paper or cotton "always" fell from 313 (76.9%) at baseline to 111 (32.0%) at 24 months.
Post-trial cup distribution
Of 83 participants who did not receive a cup at the 12-month visit and completed the 24-month visit, all were contacted, none were ineligible, and all 83 were scheduled for post-trial cup intervention; 66 (79.5%) attended. A further 21 participants who had received a cup that was lost or stolen were scheduled for replacement, of whom 17 (81.0%) attended. Seventy-seven post-cup receipt phone surveys were completed and six were not. Among the 60 of 66 attendees without a prior cup, 59 (98.3%) had menstruated since using the cup, 51 (86.4%) had used the cup, and there were no reports of cup retention problems, pain during or after cup use, or partner violence due to cup use.
30- and 36-month follow-up
Outcomes at the additional 30-month (N=70) and 36-month (N=245) visits (Table 13) were broadly consistent with earlier time points:
| Measure | 30 months (N=70) | 36 months (N=245) |
|---|---|---|
| Any physical abuse (composite of 6 questions) | 2 (2.9%) | 11 (4.5%) |
| Any sexual abuse (composite of 4 questions) | 1 (1.4%) | 6 (2.5%) |
| Raped or forced sex | 0 (0.0%) | 0 (0.0%) |
| Any emotional or financial abuse (composite of 2 questions) | 20 (29.0%) | 51 (20.8%) |
| Received cup | 64 (91.4%) | 238 (97.1%) |
| Still have cup | 55 (85.9%) | 207 (87.0%) |
| Used cup | 57 (89.1%) | 203 (85.3%) |
| Cup retention | 0 (0.0%) | 1 (0.5%) |
| Pain during or after cup use | 0 (0.0%) | 0 (0.0%) |
| Partner violence due to cup use | 0 (0.0%) | 0 (0.0%) |
| Inserted objects inside vagina for sex during menses | 1 (1.5%) | 2 (1.0%) |
| Used drying or tightening agents inside vagina - never | 67 (95.7%) | 244 (99.6%) |
| Wipes inside vagina with cloth, tissue, paper or cotton - never | 49 (70.0%) | 200 (81.6%) |
No vaginal or cervical lacerations and no menstrual toxic shock syndrome were recorded at either visit. BV cumulative prevalence at 30/36 months is reported as 146 (46.4%), of whom 128 (40.6%) had prior BV (at baseline, 6, 12, 18 or 24 months).
About the source
Source: POWWeR-Health-Updates.docx, sections 1-4 (reporting tables 1-13). The source document also contains flow diagrams - Screening and Enrollment Flow, Figure 3A Participant Follow-up, Participant Intervention Delivery and Use, Analysis Flow and Extended Observation Flow - which are not reproduced here, as they were not extractable from the Word file.
