Trial recruitment
The TOPSY trial recruited 340 participants (10 more than the 330 target) between 16 May 2018 and 7 February 2020, thus completing recruitment before the first COVID-19 lockdown (). The 10 additional women were already on various stages of the screening pathway at the time the initial target of 330 was reached, so all 10 were recruited. The original target was to complete recruitment by July 2019 (see ); however, recruitment was extended by 6 months so that the required number of participants could be randomised.
Projected and actual recruitment.
During the initial 6-month internal pilot, 72 participants were recruited (see the Project Documentation on the project website for a full report of the pilot study). Thereafter, between 12 and 27 participants joined the study during each full month of recruitment (). Participants were recruited at 21 centres across the UK (). Although centres were given targets for recruitment (based on an initial feasibility assessment at each of the centres’ study set-up), the centres that were recruiting well were permitted to recruit over the agreed target. This meant that the cohort of women could be recruited to target despite some centres struggling to recruit their expected number of women. The greater number of women recruited in two of the centres was discussed at the oversight committees (TSC and DMEC), and both committees agreed that this was acceptable.
Treatment of prolapse with self-care pessary recruitment per month.
TOPSY recruitment by centre.
Participant flow
A total of 2174 women were screened for eligibility, with 770 (35.4%) ineligible and 1404 (64.6%) declining to participate. A total of 340 women were randomised (169 to self-management and 171 to clinic-based care).
Questionnaires were completed by 334 participants (98.2%) at baseline (167/169 in the self-management group, 167/171 in the clinic-based care group). At 6 months, 306 participants (90.0%; 149/169 in the self-management group and 157/171 in the clinic-based care group) provided primary outcome (PFIQ-7) data, dropping slightly to 292 participants (85.9%; 144/169 in the self-management group and 148/171 in the clinic-based care group) at 12 months. At 18 months, the target of 264 questionnaires (80%) with valid primary outcome data set out in the sample size calculation was exceeded (291 returned; 85.6%), with 139 out of 169 (82.2%) in the self-management group and 152 out of 171 (88.9%) in the clinic-based care group. Two participants died during the follow-up period (both deaths were unrelated to the trial). The number of participants at each stage of the trial is summarised in the CONSORT diagram (). Data were collected until the trial database was locked on 17 September 2021.
CONSORT flow diagram. Some participants in the self-management group who discontinued the intervention are counted as both discontinuing pessary use and reverting to clinic-based care.
There were 23 withdrawals from the study: 5 because they discontinued pessary use, 6 because they reverted to clinic-based care, 1 because they had difficulties with the questionnaires, 5 because they opted for surgery, and 6 because they had moved/did not feel they could self-manage/were no longer interested in taking part. In addition to the withdrawals, two participants died.
Primary outcome measure
The PFIQ-7 score at 18 months was the primary outcome measure in the trial (lower scores indicate poorer quality of life related to pelvic floor dysfunction). In the self-management group, the unadjusted mean (SD) score was 32.3 (50.9) compared with 32.5 (47.8) in the clinic-based care group, with a mean difference (adjusted for baseline score and minimisation covariates) of –0.03 (95% CI −9.32 to 9.25). There was, therefore, no evidence of a difference between the groups in terms of prolapse-related quality of life. Comparable results were found at the 6- and 12-month time points (Table 7).
Summary of PFIQ-7 responses and analysis of differences between groups
Sensitivity analyses
Analyses of the primary outcome to examine data under differing assumptions relating to non-compliance and missing data (Table 8) all showed very similar results to the primary ITT analysis (see Table 7).
Summary of sensitivity analyses of the PFIQ-7 at 18 months
Subgroup analysis
The prespecified subgroup analysis was conducted by age group, pessary user type (new/existing) and history of hysterectomy at baseline and showed no significant treatment effect by subgroup interactions (), that is there was no evidence that any intervention effects are modified by subgroups. However, it is important to note that the study was not powered to investigate subgroups and effects were considered on the primary outcome only. These analyses were therefore exploratory and potentially for hypothesis generation.
Forest plot of interaction effects.
Secondary outcome measures
Pessary complications
Table 9 shows pessary complications by trial group as reported on the Pessary Complications Questionnaire at 18 months. As specified in the SAP, the proportion of complications reported by each participant, of those complications that were applicable to the participant, was calculated (Table 10).
Pessary complications reported at 18 months
Mean proportion of complications reported
In the mixed-effects linear regression on proportion of complications at 18 months adjusted for baseline, there was evidence that participants in the clinic-based care group had a statistically significant higher proportion of pessary-related complications than those in the self-management group (mean difference in proportion 3.83, 95% CI 0.81 to 6.86).
Pelvic Floor Distress Inventory-20 and subscales
Table 11 shows the descriptive statistics for the PFDI-20 and its subscales at 18 months. There are three subscales (UDI‐6, CRADI‐8, POPDI‐6), with each subscore ranging from 0 to 100 and a total score of 0–300. Higher scores indicate greater symptom severity.
Summary statistics for PFDI-20 total and subscores at 18 months by group
A mixed-effects linear regression showed no significant difference between the groups in the severity of prolapse-related symptoms measured by PFDI-20 total score at 18 months adjusted for covariates (adjusted mean difference –0.55, 95% CI –9.17 to 8.08).
Pessary care
Participants were asked whether they had removed their pessary themselves in the last 6 months. At baseline, 42 (25.15%) of participants in the self-management group and 28 (16.77%) in the clinic-based care group had done so. At the 6-, 12- and 18-month follow-up points, the proportions reporting removing their own pessary were 84.87%, 84.03% and 79.58%, respectively, in the self-management group and 19.75%, 23.03% and 25.00%, respectively, in the clinic-based care group.
Participants who reported removing their pessary were asked how often they did so. In the self-management group, the numbers of those who reported removing their pessary at least once a month in the last 6 months were two at baseline, 20 at 6 months, 20 at 12 months and 22 at 18 months. In the clinic-based care group, those who reported removing their pessary at least monthly numbered seven at baseline, nine at 6 months, five at 12 months and 13 at 18 months.
Participants were asked at each time point, ‘Are you planning to continue using a pessary to manage your prolapse symptoms?’ (Table 12). Most participants in both groups intended to continue using a pessary.
Intentions for continued use of pessary
Participants were asked at each time point to agree or disagree with three statements regarding pessaries. The responses to each statement are presented by group (Tables 13–15). The proportions finding pessary changes comfortable or convenient or pessary care acceptable are similar in the two groups.
‘I find pessary changes comfortable’
‘I find my pessary care acceptable’
‘I find pessary changes convenient’
Participants completed the Patient Global Impression of Improvement at each time point to indicate any change in their perception of their pessary care. Responses to the question ‘Compared to before I took part in this study, my pessary care now is …’ are summarised in Table 16.
Summary of Patient Global Impression of Improvement responses
A mixed-effects ordinal regression model indicated that participants in the clinic-based care group were significantly more likely to respond in lower categories. The clinic-based care group had increased odds of being in a more dissatisfied category (adjusted odds ratio 3.23, 95% CI 1.47 to 7.13), conditional on all other variables in the model and the random effects.
Pessary confidence
Table 17 shows the responses to ‘How confident are you that you can manage problems related to using a pessary?’, ‘How confident are you that you can (or could if asked) remove your pessary on your own?’ and ‘How confident are you that you can (or could if asked) insert your pessary on your own?’. Responses are on a 0–100 scale, where 0 is no confidence and 100 is highly confident.
Summary of responses to pessary confidence questions by group
The adjusted difference between groups at 18 months is statistically significant for all three outcomes. From a mixed-effects analysis there was evidence that those in the selfmanagement group were more confident that they could manage pessary problems (−7.99, 95% CI −14.15 to −1.82), that they could remove their pessary (−32.78, 95% CI –40.45 to –25.10) and that they could insert their pessary (−32.92, 95% CI –40.64 to –25.19).
General self-efficacy
The mean general self-efficacy score (measured on a scale from 10 to 40, where higher scores indicate greater self-efficacy) was similar in the two the groups (Table 18).
Summary of general self-efficiency scale responses by group and time point
In a mixed-effects linear regression, there was no evidence of a difference between the groups in general self-efficacy at 18 months (mean difference 0.77, 95% CI –0.14 to 1.69).
Sexual activity
Participants were asked at baseline and 18 months whether they were sexually active or not (Table 19).
Sexual activity at baseline and 18 months
There was very little difference between the groups in the proportion of participants who were sexually active at each time point: just over half said at both time points that they were not sexually active. The PISQ-IR50 was used to assess participants’ sexual symptoms. There are five subscales for women who are sexually active with or without a partner. Higher scores indicate better sexual function. Responses were very similar in the two groups at both time points (Table 20).
Summary of PISQ-IR responses for sexually active participants at baseline and 18 months
A linear mixed-effects model of the total PISQ-IR score for participants who were sexually active showed no difference between the groups at 18 months (mean difference –0.34, 95% CI –9.87 to 7.19).
Adverse events
No suspected unexpected serious adverse reactions were reported by participants in the trial. There were 32 SAEs in total (17 reported in the self-management group, 14 reported in the clinic-based care group). There was also one SAE reported by a woman in the non-randomised interview group relating to her pessary and her prolapse symptoms. This was reported and included despite her not being in the randomised cohort.
Fourteen of the reported SAEs were for surgery: seven for prolapse surgery and seven for non-urogynaecology surgery. Two deaths were reported: one due to COVID-19 and one due to ischaemic small-bowel infarction. Two women were hospitalised due to COVID-19. The other 14 reported SAEs, including the SAE for the non-randomised woman, were all hospitalisations for a variety of other reasons such as cancer, ectopic pregnancy and chronic obstructive pulmonary disease.
Health of vaginal tissues
In the self-management group 136 women attended their final 18-month clinic visit, and in the clinic-based care group 154 women attended. We have data on health of vaginal tissues for 135 and 153 participants from each group, respectively (Table 21).
Health of vaginal tissues at 18-month appointment, by group
Additional telephone support
In the self-management group 28 participants received at least one additional telephone support call. In the clinic-based care group 26 received at least one additional call. The difference in the number of additional telephone calls between the groups was not statistically significant (effect size 0.01, 95% CI −0.02 to 0.03). The mean number of additional clinic visits per participant was 0.25 in the self-management group and 0.19 in the clinic-based care group (effect size −0.07, 95% CI −0.20 to 0.07).
Adherence to intervention
Crossover
Twenty-eight participants reverted from self-management to clinic-based care before the 6-month follow-up, and six reverted from self-management to clinic-based care between the 6- and 12- month follow-ups. An analysis was conducted in which these participants were recoded as clinic-based care and the groups were compared by treatment received, employing an identical linear mixed model to that used in the main analysis of the primary outcome, and this is shown in row 2 of Table 5. This showed an interaction effect between the groups and the 18-month time point of –7.84 (95% CI –17.51 to 1.83), which was not statistically significant.
A complier-average causal effect estimate was obtained using instrumental variable methods, where randomisation was the exogenous instrument and treatment received under the crossover definition was the endogenous variable. The estimate of the complier-average causal effect was −0.66 (95% CI –12.59 to 11.28), which was not statistically significant.
On treatment
Very few clinic-based care participants met the original 'on treatment' definition (n = 2) because there were large numbers of missing data for the question of whether they had inserted their pessary themselves in the last 6 months. Using a modified definition of ‘on treatment’, which was specified in the SAP, an additional analysis was carried out whereby those in the clinic-based care group who had missing responses to the question about pessary insertion were treated as if they had responded ‘no’. Using this definition, there were 141 (83.4%) self-management participants on treatment and 103 (60.2%) clinic-based care participants on treatment for the 18 months. The low proportion of clinic-based care participants on treatment is primarily due to a sizeable number who had responded ‘yes’ to inserting their pessary for at least one time point. A total of 44 participants in the clinic-based care group (26.0%) reported inserting their pessary themselves at the 6-, 12- or 18-month time point, and a further 16 had discontinued pessary use.
Post hoc analysis: disruption to treatment due to COVID-19 pandemic
Primary outcome
Some participants in the clinic-based care group did not receive their clinic appointments while clinics were cancelled during the COVID-19 pandemic. The primary analysis was re-run excluding participants in the clinic-based care group who missed appointments for this reason. There were 11 participants who missed 6-month appointments and 15 who missed 12-month appointments. All data for these participants were removed from an analysis of the primary outcome at each time point, the results of which are summarised in Table 22.
Summary of PFIQ-7 responses excluding clinic-based care participants with appointments cancelled due to COVID-19 pandemic
The primary analysis by mixed-effects regression showed no difference between the groups at 18 months after these data were removed (interaction effect between trial group and 18-month time point −4.38, 95% CI −14.22 to 5.47).
Pessary complications
The analysis of pessary complications was repeated with participants in clinic-based care who had missed clinic appointments due to COVID-19 removed from the analysis. The proportion of complications experienced by participants in each group at each time point is summarised in Table 23.
Mean proportion of complications reported excluding clinic-based care participants with appointments cancelled due to COVID-19
The analysis of the proportion of complications experienced at 18 months by group showed that the participants in the clinic-based care group reported a significantly higher proportion of complications than the self-management group (4.76, 95% CI 1.56 to 7.96).
COVID-19 questionnaire
Participants in either group who had a clinic appointment postponed or cancelled due to COVID-19 restrictions were sent a survey about their experiences. There were 80 responses to the survey: 26 from the self-management group and 54 from the clinic-based care group.
Thirty-seven participants gave the dates of their cancelled appointments, and these were from 1 March 2020 to 15 December 2020 inclusive.
Participants were asked ‘When your pessary appointment was cancelled were you given clear instructions on what to do if there was a problem with your pessary?’, and 65.4% (17/26) of the self-management participants and 55.6% (30/54) of the clinic-based care participants responded ‘yes’ to this question.
Responses to the question ‘How worried were you about NOT having your pessary changed due to your face-to-face clinic appointment being delayed?’ are shown in Table 24.
Summary of level of worry reported related to missed pessary change appointment
A higher proportion of participants in the clinic-based care group reported being moderately worried about not having their pessary changed.
