Clinical and cost-effectiveness of pessary self-management versus clinic-based care for pelvic organ prolapse in women: the TOPSY RCT with process evaluation

Pelvic organ prolapse

Pelvic organ prolapse (hereafter prolapse) is the descent of some, or all, of the female pelvic organs from their usual position in the pelvis into the vagina. 1 Prolapse is common, with one UK survey identifying that 8.4% of community-dwelling women report a vaginal lump or bulge2 and up to 50% of women reported to have prolapse on examination. 3 The prevalence of prolapse increases with age and, with the UK population of older adults increasing, it is anticipated that the prevalence of prolapse will also increase. 4

Two large cohort studies identify that prolapse has a multifactorial aetiology. 5,6 Factors that predict prolapse include pregnancy, vaginal delivery, hereditary factors, ageing, menopause, and factors associated with chronically raised intra-abdominal pressure (caused by, e.g., obesity or heavy lifting). 5,7

Women report symptoms of ‘something coming down’ in their vagina, a dragging sensation in their vagina, a bulge coming down from their vagina, and pelvic and back pain. 1 The movement of the pelvic organs into the vagina can also cause urinary, bowel and sexual problems. 1 Women’s quality of life, body image and ability to function in their day-to-day life are negatively affected by their symptoms in general and the extent to which those symptoms are bothersome. 8–11

Current treatment options for pelvic organ prolapse

Women with prolapse can be treated with surgical or conservative options. Approximately 9.5% of women will undergo surgery for prolapse in their lifetime. 12–14 One study estimated the cost of any prolapse surgery in England, in 2005, as €81,030,907. 15

There are several conservative treatment options: lifestyle advice, pelvic floor muscle training, vaginal lubricants (hormonal and non-hormonal) and vaginal pessary use. Lifestyle interventions include features such as weight management or physical activity, each of which has minimal evidence of effect. 16,17 Pelvic floor muscle training has been shown to be effective across several clinical trials. 18,19 Local oestrogen as a treatment for prolapse may need further evidence to guide practice. 20

Few clinical trials have assessed the effectiveness of vaginal pessaries. 1 However, two-thirds of women will opt to try a vaginal pessary when it is offered. 21 The largest UK-based pessary study reported that 86% of women who successfully retain their pessary at 4 weeks will continue to use a pessary at 5 years. 22 However, other studies have reported lower continuation rates, for example 62.1% among women aged 65–74 years and 37.8% among those aged ≥ 75 at 5 years. 23 Reasons for pessary discontinuation include development of complications, dislike of pessary changes (insertion or removal) and inconvenience of attending appointments. 24–27

Care pathways for women who use a vaginal pessary as treatment for prolapse

Vaginal pessaries are a common, and recommended, treatment option in the UK NHS. 3,28 National Institute for Health and Care Excellence (NICE) guidance recommends that a woman be offered 6-monthly appointments if she is at risk of complications that may make it ‘difficult for them to manage their ongoing pessary care’ (recommendation 1.7.9). 3 The NICE guidance does not offer clear recommendations on self-management; however, joint guidance from the United Kingdom Continence Society and Pelvic, Obstetric and Gynaecological Physiotherapy recommends that women are offered self-management where that option is available. 29

A UK multiprofessional survey found that only 17% of clinicians offered their patients the option of self-managing their pessary, with most pessary care being delivered in clinics. 30 This is a difference in practice from North America, where self-management is offered more routinely. 31 Clinic-based care therefore seems to be the most common way of delivering pessary care in the UK, but alternative models of care delivery do exist.

Evidence for the effectiveness of self-management of vaginal pessaries for prolapse

There are no published trials to date comparing the effectiveness of pessary self-management with that of clinic-based care. 32 One small (n = 88), non-randomised study assessed self-management of vaginal pessaries;33 this study reported gains from self-management, in that women reported higher levels of convenience, ability to access help, support and comfort than those attending clinic. 33 A few observational studies have addressed various features of self-management. For example, Manonai et al. 34 undertook a chart review and found that self-management was a strong predictor of pessary continuation for women in Thailand at 3 years. The observational studies suggest that self-management may be a viable treatment option, but self-management effectiveness has not been tested.

Proposed mechanism of action for why self-management might improve quality of life for women with prolapse

The treatment of prolapse with self-care pessary (TOPSY) trial was developed prior to the publication of the 2021 complex interventions framework35 and therefore draws on the 2008 framework36 and the 2015 process evaluation guidance. 37 We will therefore refer to the mechanism of action of the intervention as opposed to the programme theory.

The mechanism of action of the TOPSY self-management intervention is based within self-efficacy theory38 and self-management theory. 39 Self-efficacy has been argued to be the mechanism through which self-management achieves its goals. 39,40 Self-efficacy focuses on an individual’s beliefs in their abilities to achieve goals. 38 Previous service evaluation has suggested that women who self-manage may feel more in control of managing their prolapse than those who receive clinic-based care. 33 Self-management theory posits that three tasks need to be achieved for individuals to self-manage: medical management of the condition, role management and emotional management. 39 Based within these theoretical constructs it was therefore hypothesised that self-management of a vaginal pessary will lead to improved quality of life for women with prolapse because support at service and professional levels, receipt of information and self-management support will lead to women becoming more confident (self-efficacious) about their pessary management; will improve their understanding of and confidence in their role to self-manage; and will enhance their emotional capacity and confidence to cope with their pessary such that their condition-specific quality of life will be improved more than that of those who receive clinic-based care. The details of the intervention are presented in Chapter 2.

Rationale for the research

It is unclear if self-managing a vaginal pessary, rather than receiving care in a clinic, would improve women’s quality of life. Data from non-pelvic health and pelvic health observational studies suggest that self-management interventions have the potential to improve quality of life. Thus, a robust comparison of a theoretically developed self-management intervention with clinic-based care is imperative to understand if alternative clinical models of pelvic health delivery can support better outcomes for women.

Aims and objectives

This research aims to answer the following research questions (RQs):

1. What is the clinical effectiveness and cost-effectiveness of self-management of vaginal pessaries to treat pelvic organ prolapse, compared with clinic-based pessary care, on condition-specific quality of life? (RQ1).

2. What are the barriers to and facilitators of intervention acceptability, intervention effectiveness, fidelity to delivery, and adherence for women treated with vaginal pessary and the healthcare professionals (HCPs) who treat them, and how does this differ between randomised groups? (RQ2).

The specific objectives are:

• to undertake a parallel-group, multicentre, individual randomised controlled trial to test for the superiority of pessary self-management compared with clinic-based pessary care in terms of women’s condition-specific quality of life

• to undertake an internal pilot study to ensure that the trial can recruit, randomise and retain sufficient numbers of participants while delivering the intervention as planned

• to undertake a process evaluation in parallel with the trial to maximise recruitment; assess eligible but non-randomised women; understand women’s experience and acceptability of the intervention; assess adherence to allocated trial group; describe fidelity to intervention delivery; and identify contextual factors that may interact with intervention effectiveness

• to undertake an economic evaluation to establish whether pessary self-management is cost-effective compared with clinic-based pessary care.

The structure of the report

Chapter 2 outlines the study design. It is split into three sections outlining the three component parts of the study: the trial, the process evaluation and the economic evaluation. The pilot study was reported previously and therefore is not included in this report, but it is included as part of the Project Documentation. 41 This Project Documentation, which we refer to throughout this report, is listed in Appendix 1 and available to view on the award website41 [https://fundingawards.nihr.ac.uk/award/16/82/01 (accessed June 2022)]. Chapters 3–5 present the findings of the trial, process evaluation and cost-effectiveness analysis, respectively. Chapter 6 brings some key points from across the three components together in a synthesis, and Chapter 7 presents the discussion.

Design

The TOPSY trial was designed to compare vaginal pessary self-management with clinic-based pessary care for pelvic organ prolapse in order to assess improvement in women’s quality of life. 42 It included a multicentre superiority randomised controlled trial comparing two parallel treatment groups: pessary self-management and clinic-based pessary care (the results can be found in Chapter 3).

Recruitment to the trial was completed on 6 February 2020, before the start of the COVID-19 pandemic. Follow-up procedures amended due to the pandemic are detailed in this chapter and included in a COVID-19 annex in the funder-approved protocol.

Participants and setting

The trial recruited women who used a vaginal pessary for the management of pelvic organ prolapse from 21 UK hospital-based centres (see Appendix 3, Table 36).

Recruitment procedure

To identify potential participants, local centre staff reviewed patient notes, clinic lists and caseloads to identify women who were currently using a pessary and were suitable to be approached. In addition, women were identified at appointments when attending for pessary review (existing users) or being fitted with a pessary for the first time (new users). If a woman had previously used a pessary but was having a break in pessary use, she was classed as an existing user, as it was believed that her experience would mean she had existing knowledge of pessary use. Women who learned about the TOPSY study themselves (by the website, posters, word of mouth) could approach their centre or the trial office to enquire about participation.

Women were provided with a recruitment pack (either given in person or sent by post) containing an introductory letter, a participant information leaflet, an expression of interest form and a reply-paid envelope. Once women had made their decision regarding participation, they returned the expression of interest form to the local clinical team. On receiving a positive expression of interest form, a member of the local clinical team discussed the study further with the woman in question and screened her for eligibility.

If the woman was a new pessary user (had used a pessary for ≤ 3 months), eligibility screening included a telephone call to confirm that the pessary had been retained for at least 2 weeks. If not, and the woman remained interested in participating, eligibility was reassessed once the pessary had been retained for 2 weeks.

Consent

If a woman was eligible and willing to take part, she attended a baseline clinic appointment where she provided written, informed consent for randomisation. The trial consent form asked participants if they were willing to be contacted about taking part in interviews for the process evaluation, and if they were willing to have their self-management teaching session or 2-week telephone call digitally recorded. They were also asked if they could be contacted about future research.

Women were informed of their right to withdraw at any time from all or part of the study. Any change to women’s participation was recorded in a study change of status form. Women randomised to self-management could opt to change to clinic-based care, and their reasons for choosing to do this were recorded. A woman randomised to clinic-based care could cross over to self-management only if she received formal TOPSY training and was no longer on a regular clinic-based pathway. If a woman discontinued pessary use, she could remain in the study and continue with the data collection elements of the research.

Participant retention

Active measures to minimise loss to follow-up of participants included:

• Recording women’s e-mail addresses and mobile phone numbers at the outset, their preferred method of contact (for follow-up) and their preferred method of completing questionnaires. Questionnaires could be completed online (via an e-mail link) or on paper and returned by post.

• Any participant who did not return their questionnaires within 3 weeks was sent a maximum of three reminders, the first two of which were via the participant’s preferred method. The third reminder was a telephone call in which either the researcher reminded the participant to complete and return the questionnaire or the questionnaire was completed during the call.

Response rates to the participant-completed questionnaires were monitored closely to ensure that they remained above 80%.

Randomisation, concealment and blinding

The trial was supported by the Centre for Healthcare Randomised Trials (CHaRT), a fully registered UK Clinical Research Network clinical trials unit in the Health Services Research Unit, University of Aberdeen. CHaRT developed an internet-based data management system and remote automated computerised randomisation system for the trial.

After informed consent was obtained from a woman, the local clinical staff entered the required information into the data management system to remotely generate the participant’s group allocation. The centralised randomisation of participants after enrolment ensured allocation concealment. The randomisation system assigned women to one of the two trial groups, with an even allocation ratio and naive minimisation by age (< 65/≥ 65 years), pessary user type (new user/existing user) and centre.

Due to the nature of the trial interventions, the trial group to which women were allocated could not be masked from the participants or the centre staff who provided treatment and assessed outcomes after randomisation, and therefore blinding was not possible.

Intervention

Data collection, management and storage

Data were gathered from participants and other sources throughout the trial. An overview of the trial data collected is presented in Table 1. Introductory information and instructions in the questionnaire booklet were drawn from previous trials led by the applicant team. 47 All data collection instruments are available on the project website as part of the Project Documentation. 41

All participants were given an individual trial identification number, which was used on all trial paperwork. After a participant consented, her demographic and medical history data were collected. Table 1 shows the primary and secondary outcome measures collected at each trial time point, and more detailed information is given in Outcome measures.

At 18 months after randomisation, participants in both groups attended a clinic appointment that included an examination of vaginal tissues. All follow-ups were completed by September 2021.

Participants in both trial groups were asked to complete questionnaires at baseline and at 6, 12 and 18 months. Participants who opted to complete the questionnaires on paper posted these back to the TOPSY office where they were checked for completeness and data were entered into the data management system. Women were sent a £10 voucher with their 18-month questionnaire (whether they returned it or not). For those participants who opted to complete questionnaires online, no further data entry was required but a check for completeness was carried out. If any data were missing, checks were undertaken to see if there was any supporting evidence of what the missing data should be. Self-evident correction was made only if there was evidence to allow this. Detailed information on permitted self-evident corrections was documented in the TOPSY data entry guidelines. If a large part of the questionnaire was missing, attempts were made to contact the participant to obtain the information.

Data return rates were continually monitored by the central team for completeness and timeliness of all data returned. The frequency with which those randomised to self-management reverted to clinic-based care was reviewed at every team meeting (usually fortnightly) and reported to the Data Monitoring and Ethics Committee (DMEC).

Outcome measures

Sample size

The aim was to recruit sufficient participants to detect a 20-point difference between groups in the primary outcome measure, namely PFIQ-7 score at 18 months. The potential range of scores on the PFIQ-7 is 0–300, and, in the absence of robust data on minimal clinically important difference, the clinicians in the research team and the wider Trial Steering Committee (TSC) considered 20 points to be an important clinical difference. A sample size of 330 women (165 per group) was required to provide 90% power to detect a difference of 20 points in the PFIQ-7 score at 18 months, assuming a standard deviation (SD) of 50, which was based on previous studies,55,56 two‐sided alpha of 0.05, and 20% loss to follow-up.

COVID-19: changes to follow-up assessment

This section describes the changes to study processes that were implemented from 21 April 2020 due to the COVID-19 pandemic. As recruitment was complete in February 2020, only follow-up processes were amended.

Statistical analysis

Study analyses were conducted in accordance with a prespecified statistical analysis plan (Project Documentation is available on the project website)41 using Stata version 16 (StataCorp LP, College Station, TX, USA). The primary outcome measure (PFIQ-7) and all secondary outcome measures were presented as summaries of descriptive statistics at each time point, and comparisons between the groups were analysed using general linear models. All analyses were adjusted for minimisation covariates (age, pessary user type and centre) and for baseline scores where applicable. The models used to analyse the continuous outcomes were repeated measures mixed models with a compound symmetry covariance matrix and centre fitted as a random effect. Estimates of treatment effect size were expressed as the fixed-effect solutions in the mixed models and odds ratios in the ordinal regression models. For all estimates, 95% confidence intervals (CIs) were calculated and reported.

Planned sensitivity analyses were carried out on the primary outcome measure to investigate the impact of missing data under various assumptions. The first sensitivity analysis was a complete-case analysis that used only cases for which follow-up data were available at the primary end point (18 months). The remaining sensitivity analyses used pattern mixture modelling by increasing and decreasing the imputed PFIQ-7 values in the initial sensitivity analysis by 20 points, equivalent to the minimal clinically important difference. These adjustments were then repeated in one group only and repeated again by applying the adjustments in the other group only. We used 20 points on the PFIQ-7 score as this was the clinically important difference initially assumed and hence a meaningful systematic difference to test in the sensitivity analyses.

A further set of planned sensitivity analyses of the primary outcome measure were conducted to examine crossover, adherence to treatment and the inclusion of previous hysterectomy as a covariate. In addition, a sensitivity analysis of the repeated measures mixed model specification was carried out, applying the constrained longitudinal model57 with the baseline value in the outcome vector, an approach suggested for extension to randomised studies. 58 Finally, a planned sensitivity analysis was carried out to incorporate mode of data collection, which shifted more to electronic submission as a consequence of the COVID-19 pandemic, as it was recognised that the results could be biased if collection method was not addressed in the analysis. 59

Research question for the process evaluation

The process evaluation answers the following research question: what are the barriers to and facilitators of intervention acceptability, intervention effectiveness, fidelity to delivery and adherence among women treated with vaginal pessary and the HCPs who treat them, and how does this differ between trial groups?

The process evaluation objectives are:

1. to undertake an internal pilot study to ensure that the trial could recruit, randomise and retain sufficient numbers of participants while delivering the intervention as planned

2. to undertake a process evaluation in parallel with the trial to maximise recruitment; assess eligible but non-randomised women; understand women’s experience and acceptability of the intervention; assess adherence to allocated trial group; describe fidelity to intervention delivery; and identify contextual factors that may interact with intervention effectiveness.

Study design

The process evaluation is a mixed-methods study that was nested within, and operationalised concurrently with, the trial. 43 The process evaluation samples, methods of data collection, including those that were part of the internal pilot study, and analysis are outlined briefly below. Further details of the design are in the published protocol. 41 Table 2 outlines the links between the purposes laid out in the objectives and the methods used.

Study methods for recruitment, consent and data collection

The sampling, recruitment and data collection for each method are outlined below. The participant information leaflets and consent forms for each element are part of the Project Documentation on the project website. 41

Data analysis

Recordings of recruitment sessions, teaching sessions and telephone calls and interviews with women were transcribed verbatim and imported into NVivo software (QSR International, Warrington, UK). Each data source was analysed individually in the first instance to reach separate conclusions, and the findings were then synthesised across data sources. Quantitative checklist and coded self-management appointment and follow-up recording data were transferred to Statistical Product and Service Solutions (IBM SPSS Statistics v26, Armonk, NY, USA) for analysis. All analysis was undertaken by the process evaluation subgroup of grant‐holders, which included PPI and clinical representation, with the rest of the team blinded to the analysis.

Perspective

A health sector payer (NHS) perspective was taken for the cost–utility analysis.

Time horizon and discounting

The primary economic analysis compared the costs and benefits of each group over the first 18 months after randomisation. A secondary analysis over a 5-year time horizon was performed using modelling beyond the trial data collection period. A 5-year horizon was chosen as it was assumed that the conditions and characteristics of patients will be broadly the same across the period while still being relevant to NHS funding cycles.

A discount rate of 3.5% was applied to all costs and outcomes over 1 year as recommended by NICE. 63

Health outcomes

Data about health-related quality of life for use in the cost–utility analysis were collected using the EQ-5D-5L. 49 The EQ-5D-5L is a generic measure of health-related quality of life with five domains: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. For each domain, respondents can select one of five levels ranging from ‘no problems’ to ‘unable or extreme’ based on their health today (the full version of the questionnaire is available as part of the Project Documentation). The raw scores from responses to the EQ-5D-5L domains can be used to generate health state utility values that are used to calculate QALYs. The QALY can be described as 1 year in full health and along with costs forms the basis of this economic evaluation. The utility values were calculated using the procedure recommended by NICE using the crosswalk from the UK EQ-5D-3L tariff. 64

The five questions are followed by EQ-VAS, a visual analogue scale on which respondents are asked to rate their health today between 0 (worst imaginable health) and 100 (best imaginable health). The EQ-5D-5L was completed at baseline and at 6, 12 and 18 months post randomisation alongside the primary outcome and other secondary measures reported in Chapter 2.

Resource use and costs

The intervention for the self-management group was additional training from a specialist nurse at a hospital clinic during the first appointment. This consisted of approximately 30 minutes more than clinic-based care with a specialist nurse, physiotherapist or consultant (see Chapter 2 for full details of the intervention). For the women in the self-management group, the regular follow-up clinic appointments were then scheduled for the 18-month time point only.

Healthcare resource use was collected from the clinic visit [case report form (CRF) 07] and the telephone support (CRF08) CRFs and from a participant-completed Resource Use Questionnaire (RUQ) designed for this study (all are part of the Project Documentation available on the project website). Data on outpatient clinic appointments related to pessary management were captured on the clinic CRF, and telephone support appointments were recorded on a telephone log CRF.

The RUQ consisted of six questions related to the use of secondary care services, primary care services and medications (prolapse-related treatments) and for any personal out-of-pocket expenses resulting from experiencing prolapse or having a pessary. For primary care services participants were asked to record the number of GP appointments in person and home visits, nurse appointments in person and home visits, district nurse home visits, community physiotherapy appointments and community dietitian appointments. For secondary care services participants were asked to record the number of outpatient appointments with a doctor, outpatient appointments with a nurse, attendances at accident and emergency (A&E), and inpatient stays including the number of nights. For both primary and secondary care resource use participants were asked to record this in terms of appointments for prolapse-related reasons and any other health-related reason.

The RUQ was completed by participants at the 6-, 12- and 18-month follow-ups; they were asked to report all resource use over the period since the previous questionnaire. Given the long period between follow-up questionnaires, an aide memoire was given to the participants so that they could note down any appointments or medication during the intervening period that could then be transferred to the main questionnaire.

The unit costs attached to each item of resource use are presented in Table 3. Unit costs were identified using Unit Costs of Health and Social Care for staff and British National Formulary for prescribed medication. 65,66 All costs are in Great British pounds (GBP) in 2019/20 prices. To calculate an A&E unit cost, we used the weighted average of all acute outpatient appointments as described in Unit Costs of Health and Social Care (page 87). 65 For hospital episodes, we used the average cost per non-elective inpatient stay (short stays), which is based on national data and described in Unit Costs of Health and Social Care (page 87). 65 Outpatient doctor appointments were costed based on consultant grade. Outpatient nurse appointments were costed based on 1-hour contact with a band 7 nurse. Outpatient physiotherapist appointments were costed based on a 1-hour appointment with a band 6 physiotherapist. In-person appointments with GPs were costed based on a 9-minute contact time for each appointment. Community nurse appointments were costed as 15-minute appointments. Costing of GP and nurse (band 7) home appointments assumes 1 hour of patient contact, which includes travel to the patient’s home. District nurse at home appointments costing assumes 1 hour with a band 6 nurse. Physiotherapy local clinic visits were costed assuming 1 hour of patient contact with a band 7 community (advanced) physiotherapist. Appointments with a dietitian were costed using a band 7 hospital-based dietitian. Clinic visits assume 45 minutes of patient contact, and telephone support calls assume 10 minutes of patient contact with a band 6 specialist nurse. The costing of the initial training appointment was based on patient-level trial data that were costed using Unit Costs of Health and Social Care depending on the grade of the HCP who provided the training. The statistical analysis accounted for the uncertainty in the unit costs by drawing Monte Carlo samples from normal distributions shown in Table 3.

Data analysis

Modelling

Decision-analytic modelling was undertaken to extrapolate costs and outcomes beyond the follow-up period of the trial to investigate the potential for cost-effectiveness to deviate from the baseline results under a 5-year horizon. The model simulates progression over time given the baseline analysis. Essentially, the observed data in the first 18 months are extended in time for an extra 5 years with uncertainty allowed. The model was developed using recommended methods. 72

A Markov decision model, referred to as the TOPSY model, with a monthly cycle is used to evaluate the effects of the intervention on costs, any QALY gains and cost-effectiveness over the 5-year horizon. The model is run as both a cohort and a Monte Carlo simulation. 73 The decision model comprises two groups, with one group for each intervention evaluated (intervention vs. clinic-based care). Each group is structured as a Markov model built around health states to which healthcare cost and QALY data collected as part of the trial are linked. The model structure is shown in Figure 1.

FIGURE 1.

State diagram of the TOPSY trial decision-analytic model.

The health states in the TOPSY model simulate the type of patient encountered in the trial. The health states can be related to light or heavy resource use and a small associated change in quality of life. For example, the good health state is associated with low resource use and high quality of life, whereas the poor health state is associated with high resource use and low quality of life. Differences between the quality of life in these states were very small, given the results of the main economic analysis that showed minimal differences between the trial groups. The distribution of total resource use in the two trial groups at each follow-up was examined to inform the level of resource use at each state and the transition probabilities over time. Patients can remain in the same health state throughout the 5-year period or move between states, as shown in the diagram. Patients can change states at the beginning of each month depending on the model parameters. To reflect the fact that patients at the start of a clinical trial should be identical, and to isolate the impact of randomisation, all patients start in the moderate state in both the clinic-based care and self-management groups.

The model parameters were derived from the trial data: (1) transition probabilities between states (depending on observed resource use changes over time), (2) treatment effects of the intervention, (3) quality of life and (4) healthcare costs (see Appendix 4, Table 37). The key transition probability parameters are manually varied to examine the impact on cost-effectiveness shown using an INMB tornado diagram that reports the range of INMBs generated for each parameter’s uncertainty range. We did not manually vary other parameters as transitional probabilities had the most impact and we wanted to show a visual representation of these using the tornado diagram. Probabilistic sensitivity analysis is employed to account for uncertainty across all model parameters, which includes 10,000 Monte Carlo draws of values from cost and patient utility distributions (see Appendix 4). 73

Missing data

Six participants did not complete the baseline questionnaire: two in the self-management group and four in the clinic-based care group. The number of missing responses in the PFIQ-7 questionnaire at each time point is shown in Table 4. A valid total PFIQ-7 score could not be generated for three participants at baseline, three at 6 months, four at 12 months and four at 18 months, because there were more than four missing items on at least one subscale (i.e. more missing items than non-missing items).

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).

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 (Figure 6), 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.

FIGURE 6.

Forest plot of interaction effects.

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).

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.

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.

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.

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.

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.

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).

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).

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).

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).

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

Post hoc analysis: disruption to treatment due to COVID-19 pandemic

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.

A higher proportion of participants in the clinic-based care group reported being moderately worried about not having their pessary changed.

Intervention acceptability

Intervention acceptability was explored for self-management and clinic-based care. Self-management was an acceptable intervention, as was clinic-based care. However, the underlying reasons for acceptability differed.

Fidelity to delivery

Adherence to trial group

Avoiding surgery was a considerable motivator for women to choose a pessary as a treatment option. This also affected their adherence to pessary use more broadly and was not particular to women in one specific trial group, but rather cut across women in both trial groups. Analysis of the interview data did not show that avoidance of surgery was a facilitator of trial group adherence depending on randomisation, apart from for one woman.

The majority of women who participated in the interview component were on treatment and adhered to their trial group allocation over the duration of the trial. However, there was variation within and between the trial groups, with some women fully adherent, some partially adherent and some non-adherent across both groups. For clinic-based women full adherence was defined as attending all clinic appointments (as far as was possible within COVID-19 restrictions) and not manipulating their pessary themselves between these clinic visits. Five women from the interview sample were deemed fully adherent to their trial group over the trial duration. In comparison, nine self-management women from the interview sample were fully adherent to their trial group. Full adherence for self-management group women was determined as independently changing the pessary over the 18 months of the trial and contacting the care team if necessary.

In addition, a number of women were partially adherent to their trial group allocation. Five women from the clinic-based group fell into this category. These women continued to attend all clinic appointments, but they also removed, cleaned and reinserted their pessary themselves between these appointments. Only one woman randomised to the self-management group was partially adherent. In this instance, the woman independently self-managed her pessary for 12 months, at which point she developed bothersome complications and changed to a surgical pathway, discontinuing her pessary use.

Only three women from the interview sample for whom we had a complete data set were non-adherent to their assigned trial group. Two self-management women changed to clinic-based care following their self-management teaching appointment as they were unable to remove their pessary themselves. One woman randomised to clinic-based care did not attend any clinic appointments and independently self-managed her pessary throughout her trial participation. She did complete her questionnaires at all time points throughout the trial, which provided safety data, and her care was handed back to the local care team.

Factors that affect adherence to trial group

Contextual factors were identified that supported trial group adherence; some were specific to one trial group and others were applicable to both groups (Figure 7). Those applicable to both groups were good general health, a supportive network, an absence of complications and treatment pathway preference.

FIGURE 7.

Contextual factors supporting women’s adherence to trial group.

Good general health

Good general health acted as a facilitator of trial group adherence for both clinic-based care and self-management and, conversely, poor health was a barrier. Good general health allowed women to self-manage without being restricted by other health concerns. Being in good general health was also a facilitator of adherence in the clinic-based group, as it meant that women were able to attend clinic appointments for pessary changes. In both trial groups, good general health also aided women to stay physically active.

A supportive network

Having a supportive social network was a facilitator of trial group adherence. For women in the clinic-based trial group, a supportive social network often took the form of a spouse accompanying or driving them to clinic appointments. Some women were accompanied by other family members, most notably their daughters, to clinic appointments. This was even more prevalent following the pandemic as many women tried to avoid using public transport. However, support was not limited to transport and clinic appointments. Women in both trial groups valued being able to talk to family members and spouses about their prolapse and felt emotionally supported in their journey. This emotional support allowed women to feel comfortable with the pessary and care pathway they were randomised to for the trial, and beyond.

Absence of side effects or complications of using a pessary

Not experiencing complications was a strong facilitator of women in the self-management group continuing to self-manage for the trial duration. Women randomised to self-management who experienced some complications consulted with their local care team as per the self-management instructions. Complications led to pessary discontinuation for one woman in the self-management group and one woman in the clinic-based group from the interview sample. This shows the impact the absence of side effects had on pessary continuation in both trial groups. Women in the clinic-based group did not differ from women in the self-management group in whether they sought medical help for complications when these occurred. One group difference in relation to complications, highlighted by the interview data, was that women in the self-management group would try to address mild side effects, such as discharge, on their own by removing the pessary and reinserting it once symptoms calmed down, before they contacted their care team:

I didn’t have a discharge as such that was bothering me, but I sort of got … because you get a little bit of discharge anyway, because at the end of the day you’ve got a foreign body in there, haven’t you, so obviously you do get tiny little bits. But at one point it seemed to increase so I took it out and cleaned it and popped it back in again, and then that seemed to help that little bit of increased discharge at the time.

Hazel, self-management, randomised

Preference for a care pathway

A preference for self-management was a facilitator of adherence among women randomised to the self-management group, as they received the treatment they desired, and the same was true for clinic-based care women. Preferring self-management but being randomised to the clinic-based group acted as a barrier to trial group adherence. The divergence from clinic-based trial group adherence because of a preference for self-management ranged from mild to high. Women who diverged mildly from adherence continued to attend clinic appointments but also removed and reinserted their pessary on their own between these appointments. By contrast, high non-adherence encompassed forgoing all clinic-appointments and fully self-managing the pessary. Women randomised to self-management who preferred clinic-based care sometimes reverted to clinic-based care following the teaching appointment.

All non-randomised women had a strong preference for their current care pathway and, with the exception of one woman, none deviated from the care pathway they were on over the 18-month period:

I felt so much, kind of relieved that this one was a lot better, and I was coping with it, I sort of didn’t want to upset the apple cart in any way.

Fleur, clinic-based care, non-randomised

Wouldn’t want to do it myself, I wouldn’t have the confidence to put me own pessary in and out.

Holly, clinic-based care, non-randomised

In addition to these contextual factors facilitating trial group adherence irrespective of trial group assignment, there were group-specific factors that supported adherence. Self-efficacy for self-management positively contributed to women’s adherence to self-management. Self-efficacy differed between the groups, with women in the clinic-based group needing more reassurance, as detailed below. In the self-management group, self-efficacy featured in the actions women took to solve problems while self-managing their pessary. High to moderate self-efficacy for self-management acted as a facilitator of self-management. High self-efficacy was demonstrated in women who adapted the teaching to suit themselves, such as finding a comfortable position in which to remove/reinsert the pessary:

I find the position that I’m doing it in now is more comfortable than I was doing it initially, so I think the more I’ve done it, the more confident I’ve got with it.

Margarite, self-management, randomised

In addition, high self-efficacy was displayed in women who took the initiative to address mild side effects, such as discharge or the pessary causing discomfort, before they contacted their local care team. Women who stated that they felt comfortable not being seen by a HCP unless they felt that there were problems that needed to be addressed (such as a change in pessary size) or developing complications also exhibited high levels of self-efficacy.

Women in the clinic-based group, having not received formal instructions in self-management, resorted to different approaches to working out how best to remove and reinsert their pessary in instances where they wanted to do so. These women’s accounts suggested that they lacked self-efficacy to change their own pessary. One woman reasoned that if the pessary needs to be squeezed going in, then the same should be true when removing it:

The very, very first time you go they [the nurses] show you that it bends. And she [the nurse] says, oh, you need to bend it to put it in and then it just springs open. So I just logically thought, well, if you have to bend it to put it in you’ve obviously got to try and bend it to bring it totally out.

Willow, clinic-based care, randomised

Another woman appeared to have received pessary self-management instructions before participating in the trial and continued to self-manage throughout the entire trial without attending clinic appointments.

Two contextual factors specifically supported trial group adherence to the clinic-based group: having a ‘day out’ and needing reassurance.

• A ‘day out’

For some women, having to attend a clinic-appointment provided the opportunity to leave their house with a clear destination and purpose. Sometimes this was the only ‘outing’ women had and they welcomed the opportunity to see someone who was not part of their immediate social network. In these cases, women preferred clinic-based care and adhered to clinic-based trial group assignment. The social interaction provided by a clinic appointment was often intertwined with a woman’s lack of self-efficacy with regard to pessary management and a more sedentary and sedate lifestyle.

• Needing reassurance

The need for the reassurance from a HCP provided during, and as a result of, routine clinic appointments was a facilitator of adherence to clinic-based care. Women found that the contact with an authority figure in the form of a trained HCP, whether a consultant or a nurse, and the visual inspection of their vaginal tissues was comforting and alleviated their anxieties about the pessary potentially harming their vagina. This need for reassurance demonstrated a lack of self-efficacy and a comfort with a more paternalistic model of healthcare delivery in the clinic-based care group:

So I was quite happy to go to the hospital because they just had a look to make sure everything was all right and I found that really quite reassuring.

Viola, clinic-based care, randomised

In summary, trial group adherence varied between and within the groups. Several contextual factors acted as facilitators of or barriers to trial group adherence.

Intervention effectiveness and quality of life

There was variation in intervention effectiveness, with multiple contextual factors affecting this, as demonstrated by the quality-of-life outcomes across both trial groups. There was variation in each group, with women exhibiting good, moderate and poor quality of life. Good quality of life for women in the self-management group was demonstrated by symptom control, elimination of painful pessary changes in clinic and being able to move house without the additional burden of having to worry about immediately having to arrange pessary care in the new location. Symptom control and an improvement in symptoms or sex life, as well as the alleviation of concerns about their pessary due to the routine input from a trained HCP, all contributed to good quality of life for women in the clinic-based group. Where some prolapse symptoms persisted despite the pessary, women in both trial groups displayed moderate quality of life. However, for women in the self-management group, being able to self-manage the pessary contributed to an overall increase in well-being.

Quality of life did not improve for all women, and sometimes it worsened. This was evident in both trial groups. Where women experienced decreased quality of life, this commonly resulted in pessary discontinuation irrespective of trial group assignment.

In fact, I think it’s probably better now the pessary is out.

Viola, clinic-based care, randomised

There was no obvious group difference, in terms of qualitative comparison, for the primary outcome of quality of life.

Contextual factors that influenced intervention effectiveness

There were contextual factors that influenced effectiveness as conceptualised by quality of life, with some factors common to both groups, some unique to self-management and some unique to clinic-based care (Figure 8).

FIGURE 8.

Contextual factors affecting quality of life.

Factors that influenced quality of life and were common to the groups were the effect of the pessary, the absence of severe complications and confidence in the HCP.

• The effect of the pessary

The pessary itself was effective for symptom control and therefore improved women’s quality of life regardless of their trial group assignment. The impact of the pessary itself meant that it was sometimes difficult to distinguish between the effect of the pessary and the effect of the trial intervention on quality of life.

• Absence of severe complications

Complications such as non-menstrual bleeding, bothersome smell or discharge, recurrent urinary tract infections or an embedded pessary had a negative impact on a woman’s quality of life, both physically and emotionally. In the case of non-menstrual bleeding this was a distressing experience for a woman, especially as women are routinely informed that non-menstrual bleeding can be a sign of other, more serious, health issues. The possible worries and discomfort caused by vaginal bleeding impacted on a woman’s emotional well-being and quality of life. Similarly, urinary tract infection symptoms were unpleasant and could limit physical activity and social interaction as women may experience discomfort while urinating and need to urinate more frequently. Discharge or a bothersome smell equally could limit a woman’s physical activity and social interactions if she felt unclean, smelly and uncomfortable, again negatively impacting on her quality of life. In the very rare scenario of an embedded pessary, severe pain and other physical symptoms considerably affected a woman’s quality of life and caused emotional distress on top of the physical pain.

• Confidence in the healthcare professional

All women participating in the interviews expressed confidence in their HCP, asserting that this gave them peace of mind regardless of the trial group they had been randomised to. Women in the self-management group said that they were confident that they would receive help if they needed it by calling the HCP on the telephone number provided. This allowed them to confidently self-manage their pessary without unnecessary worries, consequently enhancing their quality of life. Women in the clinic-based group stated that they felt confident that they received good-quality care from their HCP and that seeing the HCP routinely alleviated their concerns about possible internal abrasions, which they would not be able to see themselves, again improving their quality of life by enhancing their mental well-being.

Impact of COVID-19 pandemic on care receipt

The COVID-19 pandemic affected several patients in the TOPSY trial. For women randomised to clinic-based care, and non-randomised women on a clinic-based care pathway, the COVID-19 pandemic sometimes resulted in the cancellation of at least one clinic appointment. Depending on centre protocols, some women received a telephone call from a member of their care team to ensure patient safety and check that no complications had arisen that required the woman to be seen urgently in clinic. Receiving a telephone call from a member of their healthcare team was appreciated by, and often reassured, women.

However, not all participating centres took this approach, leaving some women without clinical contact until their centres had reopened. Some women felt concerned about not having been seen in clinic once they had reached 12 months since they last had their pessary changed. Despite reassurances from clinical staff, some women questioned the safety of keeping the pessary in situ for longer than 6 months considering that they would usually have a clinic appointment at these time points. In very few instances, women who received clinic-based care as part of the TOPSY trial resorted to removing, cleaning and inserting their pessary themselves without having been formally instructed in pessary self-management.

One woman randomised to clinic-based care removed and reinserted the pessary during the pandemic as her appointment had been cancelled due to the restrictions, and she had felt that she needed to clean the pessary because it was approaching 6 months since this had last been done. Before randomisation she had been asked by her HCP to try to remove and reinsert her pessary. One non-randomised woman inserted the pessary on her own after it accidentally came out when she was removing a tampon. This woman called her local HCP and received coaching over the telephone so she could reinsert the pessary on her own. Both these women did not feel comfortable self-managing their pessary despite being able to do so, and they continued to receive clinic-based care. Their experiences of self-managing the pessary differ from those of women who received formal instructions in clinic by a qualified HCP and who demonstrated self-efficacy in self-management. This highlights the moderating effect of self-efficacy on a woman’s proficiency in self-managing her vaginal pessary.

Women in the self-management group did not experience this level of care disruption. Several women commented that being able to self-manage the pessary during national lockdown had been an additional benefit. Receiving replacement pessaries had not been a problem during the pandemic for the majority of women randomised to the self-management group. This was because women either received the pessaries in the post or already had been given the required pessaries during the teaching appointment.

Within-trial cost–utility analysis

Intervention cost

The cost of the intervention, that is the initial self-management training given to women, consisted of an additional 30-minute appointment with a specialist nurse (band 5, 6 or 7), physiotherapist (band 7 or 8) or consultant-level doctor. The majority of appointments involved a specialist nurse, with only 14% of appointments undertaken by a consultant. This is reflected in the mean cost of the appointments, which was estimated to be £29.90. To calculate the intervention cost, we examined 156 initial appointments in the self-management group that were available in the data. The values ranged from £20 to £59.50 per appointment, and this formed the basis for the distribution of values defined as the intervention cost shown in Table 3.

Resource use data

Resource use assessment was completed for 310 participants at 6 months, 298 participants at 12 months and 297 participants at 18 months. Healthcare resource use is reported by the participants by trial group and separately for their prolapse and other health reasons (reported in Table 27 over the 18-month period). It is clear from the raw data that clinic-based care patients had more contacts with healthcare services over the 18-month period. This was the case across most categories in Table 27 for both prolapse or other reasons. One exception is district nurse home visits, although this was driven by a single patient in the self-management group (who had reverted to clinic-based care), with 30 home visits due to prolapse and 16 visits for other reasons, 89% of total visits in this group. Another exception is hospital bed-days due to prolapse; however, careful examination of this variable showed that there might have been a measurement error, with only six of these nights correctly reported. To minimise measurement error, the analysis is based on hospital episodes rather than reported bed-days. This decision makes our estimated probability of cost-effectiveness of self-management conservative given that the clinic-based care group reported more total bed-days.

The unit costs detailed in Table 3 were applied to the healthcare resource use to estimate the mean cost per patient by trial group (Table 28). The initial training appointment was applied only to the self-management group. The costs of medications prescribed for prolapse-related conditions were calculated based on information reported in the RUQ. The difference in resource use between the trial groups is driven mostly by outpatient and GP surgery appointments. The heaviest resource use was reported by two self-management patients who had reverted to clinic-based care early in the trial. For the calculation of prescribed medication costs, we excluded medications for long-term conditions such as diabetes. We focused on medications related to prolapse, infection, bowels, bladder and gynaecological conditions. These kinds of medications were reported by 110 patients (46% self-management), with average spending of £49 in the self-management group compared with £67 in clinic-based care out of the patients who reported medications or appliances. Only one patient reported the use of appliances, which were urinary catheters; this patient was in the clinic-based care group. Twenty-five per cent of drug use in the clinic-based care group was related to antibiotics, compared with 20% in the self-management group. The most commonly reported drug in both groups was prescribed oestrogen (systemic and vaginal), at 68% in self-management and 53% in clinic-based care. The proportions of vaginal oestrogen prescriptions were similar in the groups (at 6 months: self-management group, n = 9; clinic-based care, n = 17; at 12 months: self-management group, n = 15 and clinic-based care, n = 22; at 18 months: self-management group, n = 14 and clinic-based care, n = 17). Twelve per cent of clinic-based care patients reported drug use for constipation as opposed to 5% in self-management. A Cramér’s V statistic of 0.17 shows a small association between the reported types of drug use and the two trial groups, which implies significant differences between the groups in terms of the types of drugs they consumed during this trial.

Estimation of incremental net benefit

The primary analysis was based on ITT group allocation and excluded participants with missing data; this resulted in a sample of 264 patients (47% in the self-management group), with 58 excluded because they had missing data on the EQ-5D-5L. No participants were excluded because of complete missingness in the resource use data. Some of the questions had non-response, which we assumed meant one visit if non-response was to the number of visits after a positive response or zero resource use if there was no response at all. This assumption was applied to nine self-management and 15 clinic-based participants whose non-responses we had imputed. In total, < 2% of the resource use questions across all participants were imputed in this way. When we excluded these patients (not shown), our results remained very similar to what is shown below.

The incremental cost and incremental effectiveness (QALYs) of self-management compared with clinic-based care are presented in Table 29 along with the ICER and INMB.

As can be seen in Table 29, clinic-based care is dominated by self-management, and therefore the negative ICER is not shown as it is not possible to interpret it. This means that self-management was less costly than clinic-based care and was not less effective in terms of the number of QALYs gained from treatment. The INMB was estimated at £564.32 at a £20,000 willingness-to-pay threshold per QALY gained value. The INMB was calculated by translating both effectiveness and cost into a monetary valuation that depends on policy-makers’ willingness to pay per QALY gained; therefore, the INMB should be thought as value added rather than in simple cost terms. This is the baseline value added per patient by self-management at a £20,000 willingness to pay per QALY gained value. When the INMB is positive, the intervention is cost-effective when compared to the alternative. In other words, this result means that the cost to derive the benefit from self-management is lower than the maximum amount that the decision-maker would be willing to pay for this benefit.

This analysis is based on the costing of individual hospital episodes rather than the costing of individual nights in hospital as a more conservative estimate in terms of self-management cost-effectiveness (see Table 27 for a breakdown of hospital bed-days). A large number of nights were reported for non-prolapse-related reasons and therefore it was decided that to include the cost per hospital night would dominate the overall costing and not reflect the impact of pessary and prolapse required treatment options.

Probability of cost-effectiveness

The probability of cost-effectiveness can be described as the probability that an individual (random) patient will have a positive individual INMB. Essentially, this will tell us how likely it is that the baseline result (INMB = £564.32) applies to the average patient in the population, or simply how likely it is for self-management to be a cost-effective intervention when compared with clinic-based care based on the available evidence.

To calculate this probability, non-parametric bootstrapping methods were used to estimate the distribution of incremental costs and effects associated with self-management compared with clinic-based care. The results of the 10,000-bootstrap resample are shown in Table 30.

According to these results, the probability of cost-effectiveness of self-management is 80.81% at a willingness-to-pay threshold of £20,000 per QALY gained. The visual representation of these results is shown in Figure 9. We randomly selected 5% of the bootstrapped values to make the figure legible, and so the figure shows 500 points (dots) rather than the actual 10,000.

FIGURE 9.

Incremental cost-effectiveness scatterplot of self-management compared with clinic-based care for 10,000 sampled individuals (5% of values shown).

Secondary analysis with multiple imputation

To maximise the data, the analysis was repeated using the multiple imputation methods described in Chapter 2 (data analysis) to estimate EQ-5D-5L values for participants who had completed a baseline EQ-5D-5L but had missing data at a follow-up time point. This increased the sample size to 320 participants who had baseline EQ-5D-5L data [158 (49%) self-management]. Table 31 presents the mean costs and QALYs for the self-management and clinic-based care groups with multiple imputation.

Multiple imputation regression estimates of the impact of self-management on costs and effectiveness, taking into account baseline utility, are shown in Table 32. Essentially, this is the estimated marginal effect of the intervention. The methods used to undertake the estimation are described elsewhere. 77,78

To validate the results presented in Table 32, which is based on a parametric regression methodology, the probability of cost-effectiveness using a non-parametric method was calculated. A total of 1000 bootstrap samples was drawn from each of the 100 imputed data sets, creating 100,000 samples, and the difference in net benefit was estimated between the trial groups in each bootstrap sample (at £20,000 per QALY gained). The proportion of bootstrap samples in which the net benefit is positive represents the probability that the treatment is cost-effective for each multiply imputed data set. This probability is then averaged across all multiply imputed data sets and is shown in Table 33. 70

The intervention appears cost-effective when compared with clinic-based care given the estimated negative incremental cost, which implies that self-management is a cost-saving intervention. This result is similar to the previous result where missing values were not imputed. However, the imputation, together with the regression (or bootstrap) approach that took into account starting utility in both groups, reduced the probability of cost-effectiveness by approximately 9%.

Decision-analytic modelling results

A model was constructed as described in Chapter 2 to extend the main analysis for a further 5-year period. The results of the baseline analysis are shown in Table 34.

Self-management remains a cost-effective intervention when compared with clinic-based care 5 years after the initial trial period. The modelling results are consistent with the main analysis.

The cost-effectiveness acceptability curve (Figure 10) is based on 10,000 Monte Carlo probabilistic sensitivity analysis samples. This graph summarises the impact of uncertainty on the results by showing the probability of cost-effectiveness across a range of willingness to pay per QALY gain values for the two strategies. At a willingness-to-pay threshold of £20,000, the probability of self-management being a cost-effectiveness intervention is 69.74%, reflecting the probability of self-management remaining cost-effective for 5 years after the end of the trial.

FIGURE 10.

Cost-effectiveness acceptability curve at 5 years post trial.

The corresponding cost-effectiveness scatterplot showing the visual representation of the probabilistic sensitivity analysis can be seen in Figure 11.

FIGURE 11.

Incremental cost-effectiveness scatterplot of self-management compared with clinic-based care at 5 years post trial.

A deterministic sensitivity analysis is presented that examines how uncertainty changes outputs from the base-case values using the best estimate for each parameter considered (Figure 12). This focuses on transition probability parameters, which were derived from observed patient transitions in level of resource use between 12 and 18 months. The sensitivity analysis is presented in the form of a tornado diagram reporting the range of INMBs generated for each parameter’s uncertainty range. The top bar represents the probability of a self-management patient transitioning from moderate to poor over time, which has a base case of 5% per month and an uncertainty range of 1–9%. The blue portion of the bar represents the low part of the uncertainty range (from 1% to 5%), while the red portion of the bar represents the high part of the uncertainty range (from 5% to 9%). Therefore, decreasing the parameter increases the INMB calculation, which makes sense because decreasing this parameter implies that self-management patients are less likely to transition to the poor state over time. The second bar represents the probability of a self-management patient transitioning from poor to moderate over time, which has a base case of 3% per month and an uncertainty range of 0–10%. Similarly, the blue portion of the bar represents the low part of the uncertainty range while the red portion of the bar represents the high part of the uncertainty range. In this case, increasing the parameter increases the INMB calculation, which again makes sense because increasing this parameter implies that self-management patients are more likely to improve from poor to moderate.

FIGURE 12.

INMB tornado diagram. CBC, clinic-based care; SM, self-management.

This sensitivity analysis suggests that varying these parameters has a moderate impact on the INMB, which remains positive for most values in the uncertainty range. The probabilities of self-management patients changing between moderate and poor states appear to have the biggest impact. This means that cost-effectiveness depends on the intervention’s impact on patients’ health status, which is then associated with higher or lower resource use. This impact mostly relates to the ability of self-management to keep patients from transitioning into a poor state rather than keeping them in a good or moderate state.

Strengths

The key strengths of this study are its uniqueness in the field, the robust implementation and the holistic perspectives gained from using a mixed-methods design. To the best of our knowledge, no other trials worldwide have compared self-management with clinic-based care. There are studies suggesting that self-management supports pessary continuation33,34 and an ongoing trial comparing self-management for two different pessary types,32 but nothing that compares the process of offering self-management as an alternative service structure. TOPSY therefore offers a unique perspective for clinical practice.

The TOPSY trial was large enough to have detected a meaningful difference in quality of life had one existed. The trial, process evaluation and cost-effectiveness analysis were undertaken using robust methods that were implemented with scientific rigor, and the response rates at all time points were > 87%. The core findings of the study were consistent across methods and when sensitivity analyses were applied. The sample was drawn from geographically spread and diverse locations across the UK. The appointed staff team were the same throughout the study, offering consistency in implementation. The mixed-methods design ensured a holistic perspective on self-management in a way that supported understanding of the trial findings. Overall, the study design and its implementation were strengths that support confidence in the validity of the findings.

Limitations

Limitations focus on the skew in the baseline data for the PFIQ-7, the influence of the global pandemic, the potential effects of crossover and the lack of ethnic diversity in the recruited sample.

A total of 32.9% of the recruited sample scored zero at baseline on the primary outcome measure (PFIQ-7), which suggests that these women had the best possible prolapse-specific quality of life, the implication being that are that there was no room to improve that outcome by implementing self-management. It raises the question of whether the population from whom the sample was recruited was the ‘right’ population and/or whether the chosen outcome measure was the best measure. In terms of population, TOPSY was a pragmatic trial and therefore efforts were made to ensure that the inclusion criteria were as wide as possible and exclusion criteria were minimised. In practice, clinical collaborators confirmed that the recruited sample were representative of those to whom they would offer self-management. They would, for example, still offer an individual self-management to a woman whose prolapse-specific quality of life scored zero. Thus, although the skew to zero may have made it less likely that a difference would be found, the population was the correct population. Previous self-management studies have used disease-specific quality of life to measure outcome. There are many outcome measures for pelvic floor dysfunction symptoms,82 with the PFIQ being one commonly used to measure prolapse-specific quality of life in pessary trials. 1 It is possible that the measure used was not sensitive enough to identify changes in prolapse-specific quality of life over and above the effects caused of the pessary itself.

The global COVID-19 pandemic had the potential to limit the study. However, TOPSY had completed recruitment before the first UK lockdown. Although some clinic appointments were cancelled, and some were held later than planned, sensitivity analysis revealed that this did not affect the findings. Qualitative data suggested that some women in the clinic-based care group removed and replaced their pessary themselves because their appointment was not available. However, women also did this sometimes pre-pandemic. Therefore, although the pandemic had the potential to alter the intervention, it did not seem to do so.

Thirty-four women (20.1%) crossed over from self-management to clinic-based care, and nearly 40% of those randomised to clinic-based care inserted their pessary themselves at least once over the 18 months’ follow-up. Both these actions could potentially dilute the prolapse-specific quality of life effect of self-management in an ITT analysis. TOPSY was a pragmatic trial, and therefore crossovers between treatments occurred as they would in routine pessary management. As a result, this is an important part of the assessment of effectiveness.

The recruited sample in the TOPSY trial had minimal ethnic diversity (see Table 5). This will in part have been because the self-management information was only available in English. In hindsight, funds could have been requested to enable materials to be translated into the other common languages spoken in the UK. The implications of the underrepresentation of women from different ethnic groups is that the findings may not be generalisable beyond women from white ethnic backgrounds. Future studies of pessary self-management should undertake translation of materials to allow the inclusion of women who represent the wide variety of ethnic groups in UK society.

Self-management and quality of life

There was no evidence that self-management improved or worsened prolapse-specific quality of life (measured using the PFIQ-7) in comparison with clinic-based care at 18 months. This finding held true at the 6- and 12-month time points; when the primary outcome was examined under different assumptions; when COVID-19 alterations to clinic-based care pathways were considered in the analysis; and for different subgroups. Generic health-related quality of life (measured using the EQ-5D-5L) also demonstrated no difference between the groups at any time point. A range of quality-of-life outcomes were articulated by randomised women in both groups throughout the interviews. However, women who self-managed appeared to have the potential to improve their quality of life to a greater extent than women in the clinic-based care group because they managed the pessary in ways that suited their lifestyle.

One possible explanation for this finding of no difference is that there was a skew to zero (best quality-of-life score on the PFIQ-7) in the primary outcome data at baseline. This makes it difficult to detect improvement, which could be due to the primary outcome measure not capturing the full range of the intended quality-of-life constructs when used in this population. 83 However, in a post hoc sensitivity analysis, all participants with a baseline PFIQ-7 score of zero were removed and the analysis of the primary outcome was rerun. The outcome remained the same, with no group difference in PFIQ-7 score. Another possible explanation could be that the intervention effect is diluted by women crossing over from self-management to clinic-based care. However, although not technically crossing over as they were not given formal self-management teaching, women in the clinic-based care group did at times remove their pessary themselves. Guidance was followed for a treatment policy strategy that ignored these intercurrent events in the primary analysis by following the ITT principle. 84

Two small observational studies have been identified that focus on pessary self-management. One offers a non-randomised comparison of self-management of vaginal pessary with clinic-based care33 but did not measure quality of life. The authors did report higher levels of pessary changes being comfortable and higher convenience for the self-management group. The second study is a retrospective chart review of 289 women, which identified self-management as a strong predictor of continuation but did not measure quality of life. 34 A 2020 Cochrane review of pessary effectiveness identified uncertainty about there being a difference in prolapse-specific quality of life when pessary was compared with pelvic floor muscle training. 1 However, prolapse-specific quality of life was improved when pessary was added to pelvic floor muscle training compared with pelvic floor muscle training alone (moderate certainty of evidence). Some previous reviews of self-management interventions in other long-term condition contexts have found improvements in quality of life, albeit with low-quality evidence,85,86 and others have not,87 again with low-quality evidence. Therefore, variance in the effect on disease-specific quality of life has been reported for both pessary use generally and self-management programmes more widely, and the TOPSY trial adds further data to these debates about self-management’s links to quality of life.

Some self-management interventions may act to improve quality of life by improving symptoms, for example decreasing exacerbation days for chronic obstructive pulmonary disease,85 whereas in this trial the effective treatment component for symptoms (the pessary) was delivered prior to recruitment. The tool selected to measure prolapse-specific quality of life (PFIQ-7) is a measure of how much ‘symptoms affect the quality of life of women with pelvic floor disorders’. 88 The TOPSY trial aimed to evaluate a self-management versus clinic-based care model that may not have impacted on the symptoms enough to be detected by the primary outcome measure, which focused on quality of life through symptoms. However, generic quality of life also did not differ in ways that might have been expected from a self-management programme that targeted self-efficacy. It may be that more sensitive tools for measuring quality of life factors that are important to women need to be developed for pelvic floor dysfunction studies.

Overall, we can conclude that self-management of a vaginal pessary did not improve disease-specific or generic quality of life more than clinic-based care.

Clinical implications of self-management

A key finding was that women in the clinic-based care group experienced proportionally more complications than women in the self-management group. This finding held true when analysis excluded women who had missed appointments due to COVID-19; was not caused by differences in pessary material (as pessary materials were similar between the groups); was not linked to different proportions of additional appointments between the groups; and was not linked to differences in prescriptions for local vaginal oestrogen (which were similar between the groups). There are few other studies that report on the different rates of complications between women who self-manage and those who do not. NICE guidance identifies a lack of data on pessary complications. 28 One small observational study of 100 women with prolapse identified self-management as a means of reducing adverse events, with 16% of self-managing women experiencing adverse events compared with 62% of non-self-managing women;89 and another small observational study (n = 74 participants) notes low complication rates of 6.8% in women who self-manage but did not offer a comparison group. 90 Complication rates of vaginal pessaries have been reported to be > 50% in some studies91,92 and are reported to be markedly higher than those experienced by women who are treated using pelvic floor muscle training. 1 Women in the TOPSY study, regardless of group allocation, and evidence from other studies link complications to pessary discontinuation. 34,91

Fewer complications in the self-management group may be explained by these women having greater confidence in their abilities to manage their pessary (remove and insert it) and manage any problems associated to it. Women having the confidence and ability to manage the pessary and pessary problems is coherent with the self-management theory on which the self-management intervention was based. 39 One small randomised study identified that vaginal pessary complications were lower when pessary changes occur every 3 months, as opposed to every 6 months, but the sample was small (n = 60) and the difference was not statistically significant. 93 Complications have also been linked to the duration the pessary is in situ. 94 Self-efficacy may support women in changing their pessary more frequently and using it for less time (e.g. only when they felt they needed to), which may be one possible mechanism through which complication rates are lowered. Further research is needed to understand the mechanisms through which self-management reduces complication rates.

Women in the TOPSY study reported that removing their pessary was more problematic than inserting it because once it was inside the vagina the pessary was more difficult to manipulate. Discontinuation of pessary use in women who self-manage has been linked to difficulties inserting or removing the pessary. 28 Most self-management training for pessaries (including our own) places greater emphasis on insertion than removal. The HCPs delivering the intervention instructed women to try different positions that would allow for a more comfortable insertion or removal of the pessary; however, despite these instructions, the findings from women in TOPSY suggest that further emphasis is needed on removal during self-management training. The role of pessary removal devices in increasing patient confidence and reducing discomfort when removing the pessary is an area for future research.

Clinically, self-management of vaginal pessaries seems to have a valid place in reducing the complications women experience. Providing further information in self-management training about pessary removal may be helpful. Each of these actions may support a reduction in pessary discontinuation; however, this assertion needs tested in future studies.

The cost-effectiveness of self-management

The results suggest that self-management is cost-effective at a WTP threshold of £20,000 per QALY gained. Given the economic analyses presented, a recommendation can be made that self-management should become an established NHS intervention as it is very likely to be cost-effective. Decision-analytic modelling supports this result and suggests that the intervention remains a cost-effective option for the health service for longer durations than the actual trial. The nature of this intervention meant that there was a rather small impact on quality of life, and the difference between the groups was not statistically significant. One reason for this may be that the EQ-5D-5L was not sensitive enough to capture differences between the groups. Therefore, our results were mostly driven by differences in resource use, which were clearly in favour of self-management in all of the examined cost categories. This picture remained consistent across all methods and data examined either fully completed or with imputed values. There is a degree of uncertainty in our results, and this is captured by the probability of cost-effectiveness and the incremental cost-effectiveness scatterplots (see Figures 9 and 11). The probability of cost-effectiveness results show that self-management is likely to be a cost-effective alternative to clinic-based care. The scatterplots can also be interpreted as the distribution of individual patients who received the intervention compared with clinic-based care. These show that most individual (random) patients will have a positive individual incremental net benefit if they receive this intervention. It can also be seen as the proportion of all patients in the population who have positive individual incremental net benefits when practising self-management rather than receiving clinic-based care. Results across all methodologies presented suggest that self-management was successful in reducing contacts with the health service without compromising patient quality of life. This was the case for both clinic visits and overall contact with primary and secondary care services. The observed lower resource use in the self-management group could have been because self-management patients either felt more confident to deal with their medical issues on their own or had less need for medical care than patients in clinic-based care. Our results suggest that some self-managing women gained in quality of life and at the same time had less need for health services than those in clinic-based care. The data and methods employed were not designed to give us more details of the characteristics of these patients who gained the most from this intervention. Further research is needed to establish the mechanism by which self-management reduces demand for health services and the type of patient who will gain the most from self-management. In the meantime, however, we can be relatively confident that pessary self-management is a cost-effective option for the majority of patients being treated for pelvic organ prolapse.

Self-efficacy as a mechanism of action

The proposed mechanism of action of the self-management intervention was that self-management support (implemented as a teaching session, a follow-up telephone call, a leaflet and as-required telephone support) would improve women’s self-efficacy more than clinic-based care. However, general self-efficacy did not differ between the groups at 18 months. Self-efficacy in elements of pessary self-management specifically did differ between the groups, with women in the self-management group having more confidence in their ability to remove and insert their pessary as well as manage pessary problems and manage their own lifestyle to include pessary care. There is little available global literature to support or dispute these findings in the context of pelvic floor dysfunction. One study of 60 participants compared women’s understanding of pessary care when they were given an educational brochure alongside verbal instruction to support understanding compared with verbal instruction alone. 95 Those who received the education brochure were found to be more confident in self-management 1 week and 3 months after teaching was delivered. Murray et al. ’s95 study included only women who self-managed, and therefore no comparison with clinic-based care was available. It may be that condition-specific elements of self-efficacy are improved through self-management, programmes but overall self-efficacy may be less likely to be influenced.

Self-management theory purports that self-efficacy is a mediator for quality of life. 39 It has already been discussed that the links between self-management and quality of life are uncertain. The same is true for the links between self-management and self-efficacy. For example, a Cochrane review of self-management programmes for people with stroke reported that self-management did improve self-efficacy (low-quality evidence),86 but digital interventions for chronic obstructive pulmonary disease management did not improve self-efficacy (very uncertain evidence). 96 Based on the TOPSY findings and findings from these other studies in other clinical contexts, the links between self-efficacy and quality of life seem uncertain.

Effects of the COVID-19 pandemic on the treatment of prolapse with self-care pessary study

Both the TOPSY trial and process evaluation had completed recruitment by the time the first national lockdown occurred. Delivery of the self-management intervention was also completed, apart from one woman whose teaching was delayed (at her request) due to COVID-19. Therefore, effects on the study were limited to the delivery of clinic-based care, the final 18-month clinic appointment for both groups and participant follow-up.

The data presented in Chapters 3 and 5 show that some clinic-based care appointments were cancelled, with most of these appointments replaced by telephone calls. Data from the process evaluation (see Chapter 4) show that some women in the clinic-based care group did remove their pessary themselves while awaiting an appointment. However, the effect of this on the trial outcome overall was likely to be minimal as, even without missed appointments, some women removed their pessary themselves. The primary outcome findings when those women who had missed appointments were removed was the same as those of the main ITT analysis. The economic analysis took into account that some appointments were cancelled or changed from in-person to over the telephone due to COVID-19. These changes were small and, given that COVID-19 affected both groups in a similar manner, should not have an impact on the cost-effectiveness results.

Women in the clinic-based care group also reported being moderately worried about their pessary not being changed because of COVID-19 restrictions, whereas self-management women reported less worry. Self-management was, therefore, a useful care pathway for women during the unanticipated pandemic.

Therefore, although there were some unanticipated changes to care pathways due to the pandemic, it is believed that these did not unduly influence the findings beyond the influence of individual and service variance that would be normally found with a pragmatic trial design.

What is self-management and how does it fit with patient-centred care?

Self-management support is an important component of person-centred care. The four principles of person-centred care are:

1. affording people dignity, compassion and respect

2. offering coordinated care, support or treatment

3. offering personalised care, support or treatment

4. supporting people to recognise and develop their own strengths and abilities to enable them to live an independent and fulfilling life (see De longh 2015 under Further reading).

How do we support women to self-manage their pessary?

As part of a self-management programme, women have three tasks to achieve and six skills to learn in order to enable the self-efficacy that is needed to self-manage. The three tasks are:

• Medical management of the condition. For pessary self-management this is women’s medical management of prolapse using a pessary.

• Role management. For pessary self-management this is maintaining, changing or creating new behaviours such as altering exercise activities, or removing the pessary (if required) for sexual intercourse.

• Emotional management. For pessary self-management, this is learning to manage the emotions of having prolapse and using a pessary and is a part of the work required to manage the prolapse (e.g. coping with fear of the prolapse getting worse with age or worries about being able to successfully remove and replace the pessary).

For TOPSY the logic of how pessary self-management will improve quality of life is as follows: via information received, teaching and support, women will become more confident (self-efficacious) about the medical management of their prolapse using a vaginal pessary, they will understand their role in relation to self-managing their prolapse and have confidence in their ability to do so, and they will have the emotional capacity and confidence to cope with the prolapse and pessary such that their quality of life will be improved more than those who are followed up as per clinic-based pessary care.

The six skills are:

• problem-solving, for example women knowing how to troubleshoot pessary problems such as discomfort or discharge

• decision-making, for example women knowing when to remove and insert the pessary and knowing to call the advice line if they experience non-menstrual bleeding

• resource utilisation, for example making sure women know what resources are available for example by providing more than one source of information such as a telephone number, information leaflet and online video

• the formation of a patient–provider partnership role, for example the health professional role changing to that of a teacher and health partner rather than just a health care provider so that women can feel a partner in their care

• action planning, for example a woman deciding to remove the pessary to have sex

• self-tailoring, for example the woman using the pessary in a way that works for her.

Section 3 in the guide outlines how we are asking those delivering the self-management intervention to support women in gaining the skills to be competent in the tasks.

Why do this study now?

Self-management has been researched in other conditions. In patients with chronic obstructive pulmonary disease, self-management interventions have been proven to improve their quality of life, reduce hospital admissions and improve the symptom of shortness of breath. Further research is ongoing in patients after acute brain injury, cardiac event and long-term management of diabetes (www.health.org.uk). This evidence supports the idea that self-management can improve quality of life.

A quality improvement project assessing the impact of the introduction of a self-management pathway for women using pessaries for vaginal prolapse in the UK reported that women in the self-management pathway reported higher levels of convenience, ability to access help, support and comfort than women attending clinics for pessary changes. This supports the idea that the gains seen through self-management in other conditions might be possible for women who use a pessary for prolapse, but we need robust evidence before this can be implemented in practice.

The TOPSY study aims to evaluate self-management in a randomised control trial to provide high-quality evidence on the effect of self-management on the quality of life of women using pessaries.

Appointment to teach self-management (teaching appointment)

The following should be discussed with the woman prior to teaching self-management technique:

Follow-up phone call

The purpose of the follow-up phone call is to find out how the woman is getting on with self-managing her pessary and to check that she has removed it and replaced it at least once.

• Contact the woman by telephone approximately 2 weeks after the appointment to teach self-management.

• Check that the woman has managed to remove, clean and replace their pessary at least once since her appointment and record this in the telephone call CRF.

• Check that the woman did not have any issues removing or replacing her pessary.

• Check that the woman does not have any complications, for example problems opening her bowels or passing urine, bleeding, pain or discomfort.

• Check whether the woman has any questions at this point.

• Ensure the woman has the appropriate clinic telephone number to call in case of any problems.

• Remind the woman to remove, clean and reinsert the pessary at least every 6 months.

Helpline

The purpose of the centre specific advice helpline is to offer a forum for self-managing women to report any concerns they have. All women will have the self-management information leaflet that tells them the symptoms that they ought to report.

If the woman contacts the helpline with any problems, document the problem they are reporting and either provide reassurance over the phone if clinically appropriate or arrange an appointment for her to attend and be assessed within a suitable timeframe depending on the issue as per standard clinical care.