Synthetic sling or artificial urinary sphincter for men with urodynamic stress incontinence after prostate surgery: the MASTER non-inferiority RCT

Male sling

The male sling (i.e. the non-absorbable polypropylene mesh implant) is placed under the urethra to elevate it and is held in place by passing it through the obturator foramen of the pelvis bilaterally. It has a passive mode of action. The aim of the sling is to stop the loss of urine on exertion and the operation is effective immediately.

Artificial urinary sphincter

The AUS consists of an inflatable cuff that is placed around the urethra, a pressure-regulating balloon that keeps the cuff inflated and a pump that is placed in the scrotum, which the man squeezes to enable voiding. The aim is to close the urethra so that the patient is dry, except when he wants to void. Once implanted, the device is deactivated in the open position for 4–6 weeks to allow postoperative swelling to subside. The device is then activated in clinic to ensure that the man can use the device correctly.

Principal objectives

The primary objective was to compare (1) participant-reported continence 12 months after randomisation [a composite outcome derived from responses indicating any loss of urine to either of the two questions ‘How often do you leak urine?’ and ‘How much urine do you leak?’ from the validated International Consultation on Incontinence Questionnaire-Urinary Incontinence Short Form (ICIQ-UI SF)] and (2) the corresponding cost-effectiveness of a policy of primary implantation of the male sling compared with AUS, measured as the incremental cost per quality-adjusted life-year (QALY), 24 months after randomisation.

Secondary objectives

Secondary objectives compared (1) adverse events (AEs); (2) the costs of the benefits and harms; (3) the need for further treatments; (4) the differential effects on other outcomes, such as quality of life and general health; and (5) participant satisfaction with the male sling and the AUS surgeries up to 24 months after randomisation.

Qualitative component

MASTER included a comprehensive embedded qualitative component to establish the patient-perceived importance of outcomes and to explore participant and surgeon experiences of surgery and acceptable inferiority margins, and to determine reasons for failure that resulted in crossover to alternative surgery.

Closure of the non-randomised cohort

Recruitment to the NRC closed in October 2015, after 100 participants had consented, because it was adversely affecting recruitment to the RCT. 1 Those men already consented to the NRC remained in the study and were followed up as per protocol.

Extension to recruitment

Slower than anticipated recruitment meant that a 9-month recruitment extension was required to achieve the full sample size.

Safety reporting

Adverse events were either notified to the study office by the local research team or reported by the men in their follow-up questionnaires. If a SAE was suspected, it was verified by the local research team, where possible. Unrelated AEs were not recorded. The published protocol1 provides more details on relatedness, potentially expected AEs and reporting.

Sample size

Evidence suggested that 20% of men would still be incontinent 12 months after receiving an AUS, whereas 35% of men would still be incontinent after receiving a male sling. 1 The sample size calculation was undertaken by simulation. Assuming that there was no difference between the arms of the trial, 310 participants would give 90% power to show that male slings were non-inferior to AUS by a margin of ≤ 15%. To allow for 15% loss to follow-up, the sample size was increased to 360.

Statistical analysis of outcomes

Predefined statistical analyses were included in the published protocol. 1 Statistical analyses were based on all randomised men, regardless of whether or not they complied with their randomised surgery. The comparisons were between those who were randomised to receive a male sling and those who were randomised to receive an AUS.

The primary outcome was analysed using a generalised linear model, clustering by centre and with adjustment for previous radiotherapy and the 24-hour pad test weight at baseline as fixed effects. Statistical significance was at 5%, with a corresponding CI, equivalent to a one-sided test for non-inferiority at 2.5%. An intention-to-treat (ITT) analysis was performed with all of the participants remaining in their randomised group. ORs are presented for the primary outcome and the less strict definition to aid interpretation.

A subgroup analysis of the primary outcome on baseline urine leakage (≤ 250 g vs. > 250 g) was also performed. It was also planned to perform a subgroup analysis on previous prostate surgery (radical prostatectomy vs. TURP); however, given that 93% of randomised participants underwent radical prostatectomy, the groups were too unevenly balanced to make this worthwhile. It is also for this reason that the analysis does not adjust for type of prostate surgery. The subgroup analysis was at the stricter, two-sided, 1% level.

Secondary outcomes with multiple categories (e.g. satisfaction categories) were analysed using ordered logistic regression clustered by centre and with adjustment for previous radiotherapy and baseline pad weight. Secondary outcomes that are continuous outcomes and measured repeatedly (e.g. EQ-5D-3L, ICIQ-UI SF, ICIQ-MLUTS incontinence and voiding scores) were analysed using a mixed-effects repeated-measures model, with random effects for centre and participant and fixed effects for treatment, time point and the respective outcome at baseline, as well as for the baseline pad test weight and previous radiotherapy.

The primary outcome was tested in a non-inferiority framework with a non-inferiority margin of 15%, thought to be an acceptable lower effectiveness, in return for reduced AEs and easier operation of the sling. The null hypothesis was that the male sling was inferior to the AUS by at least 15% and a p-value of < 0.025 would indicate that the null hypothesis could be rejected, suggesting the male sling was non-inferior to the sphincter.

Secondary outcomes that measure quality of life, such as EQ-5D-3L and SF-12 version 2, were tested under a superiority framework. Continence-specific secondary outcomes, such as ICIQ-UI SF and ICIQ-MLUTS incontinence, voiding and sexual functioning scores, were also tested under a superiority framework.

Non-responder analysis

Descriptive data comparing the baseline characteristics of participants who did and did not respond at 12 months are displayed and the t-test (continuous outcomes) and chi-squared test (categorical outcomes) were used to estimate the statistical significance of the differences between responders and non-responders.

Missing data

Where an outcome was to be analysed using a model that adjusted for the baseline outcome, any individuals for whom this was missing at baseline had their outcome imputed by the centre mean or, in the case of the baseline 24-hour pad test, the overall mean.

Missing outcome data were imputed using a multiple imputation model (using chained equations) with randomised intervention, number of times leaked, quantity leaked, previous radiotherapy, previous prostate surgery and age used as covariates. Pattern-mixture models made the following assumptions: all missing data being continent; all missing being incontinent; all missing slings were continent and all missing AUSs were incontinent; and, finally, all missing slings were incontinent and all missing AUSs were continent.

Pre-funding application and design of the research

Prior to the initial funding application, we had the benefit of a patient advisor and an expert panel of service users who were extensively involved in all stages of the initial application, including the outline stage, to ensure that the trial addressed matters of most concern to men with incontinence. Members of this panel were interviewed on their leakage problems, expectations from surgical treatment and willingness to participate in the planned research. We also benefited from their review and approval of the proposed trial documentation [including the draft application, proposed questionnaires and instructions for the 24-hour urine leakage (pad) test]. Feedback from the expert panel helped to refine the study to understand what outcomes were important.

We also had a patient advisor grant holder who advised throughout. His thoughts were:

It is helpful to the trial for the consumer rep[resentative] to have some experience of the problems being investigated. Thank you for asking me to join the PMG [Project Management Group]. As someone with the problem of leakage and an AUS fitted for many years, I was able to help with the wording and arrangements of the questionnaires given to potential participants. Being a member of the PMG I was able from time to time to ask for clarification of the wording of some of the professional statements. Receiving the paperwork and documents of the TSC [Trial Steering Committee] is very important to know how the trial is proceeding and the problems occurring. It is important to know that the trial progress and problems are circulated to other parties to gain information or avoid duplication. It has been nice to observe and know that all professionals involved in this trial have ‘gone about’ the investigations in a keen and interested manner. From my eyes the trial has been worthwhile and has found other unknown problems.

Embedded qualitative component

MASTER incorporated a significant qualitative component, which was embedded at each stage of the study, to include patient and public involvement in the following:

• establishment of patient-perceived importance of different outcomes

• exploration of patients’ perspectives on acceptable inferiority margins

• patients’ experiences of procedures

• reasons for failure resulting in crossover to alternative surgery

• surgeons’ experiences of procedures.

The full methods, findings and evaluation of the qualitative component are reported in Chapter 6.

Oversight of the study

One of the independent members of the TSC was a patient representative. The TSC met throughout the study and reviewed all study documentation, including participant-facing documents and questionnaires that were sent to participants.

Report writing, academic paper preparation and dissemination

The patient and public involvement partner on the TSC has been actively involved in discussions of the trial results with the TSC and has been supportive of the study in report preparation.

Sensitivity analysis

Figure 4 shows the sensitivity analysis of the primary outcome. Multiple imputation and pattern-mixture modelling were used. Estimates were consistent in all but the most extreme circumstances, which are unlikely to reflect the missing data.

FIGURE 4.

Forest plot showing the sensitivity analysis of the primary outcomes on missing data.

Subgroup analysis by baseline pad weight

The men’s report of continence at 12 months was available for 312 men in the RCT: 134 men with a pad weight at baseline of ≤ 250 g and 178 men with a pad weight at baseline of > 250 g.

When the original continence definition was used, in both subgroups, the proportion of men incontinent was larger for those who received the male sling than for those who received the AUS (Table 8). When the less strict definition was used, the proportion of men incontinent in the > 250 g subgroup was larger for those who received the male sling than for those who received the AUS. In the subgroup for which the baseline pad weight was ≤ 250 g, the AUS group had a larger proportion of men who were incontinent than the male sling group.

The forest plot (Figure 5) displays this subgroup information visually and the size of the interaction effect. There is no evidence that baseline pad weight moderates the overall treatment effect.

FIGURE 5.

Forest plot showing the continence rate using the original and the less strict definitions based on baseline pad weight.

Questionnaire response over time

The 6-month questionnaire was sent post surgery and the 12- and 24-month questionnaires were sent post randomisation. Where there was a delay in surgery and the 6-month and 12-month questionnaires were due at the same time, the 12-month questionnaire was prioritised. This explains the slightly lower response rates at 6 months (which were shown in Chapter 3) (Table 9).

Summary

The principal research question being addressed in the economic analysis was the cost-effectiveness of a policy of primary implantation of the male synthetic sling compared with AUS, measured by incremental cost per QALY at 24 months post randomisation. The within-trial economic evaluation analysis follows the ITT principle and is based on data collected alongside the RCT and the health economics plan. A Markov cohort decision analysis model was developed to consider a longer time horizon to provide additional information for policy-makers. The base-case analysis assessed the costs and cost-effectiveness of the interventions compared with the perspective of the NHS in accordance with the NICE recommendations. 17 Sensitivity analyses exploring a broader perspective of cost were also conducted. Data were collected on resource use for the delivery of the intervention; primary and secondary NHS resource use over 24 months’ follow-up, including referral for additional specialist management; and broader societal resource use, which included personal costs to the participants for containment products, private medical costs and lost productivity costs, mainly lost income, over the 24 months. All costs were expressed in Great British pounds and were valued in 2017–18 prices. Costs and benefits experienced after the first year were discounted at the recommended 3.5% per annum. 17 The sensitivity analysis explored the impact of varying key assumptions on the base-case estimates of cost-effectiveness in accordance with NICE recommendations. The methods for within-trial and modelling analyses are described here.

Measuring resource use

Intervention resource use was recorded prospectively for every participant in the study (Table 13). Intervention resource was collected using an intraoperative (theatre) case report form (CRF). This included the grade and type of staff who administered the intervention, the operating urologist (and whether or not they were supervised), the anaesthetist and nurses. The operation time was calculated from the time of entry into the anaesthetic room to the time of leaving the theatre. The type of anaesthesia administered was identified as general, spinal or local. Information that was collected included the device that participants received, either AUS or sling; the use of prophylactic antibiotic; pre-treatment and post-treatment catheterisation; the use of either oral or parenteral pain relief; the use of postoperative antibiotic treatment; and the length of stay.

Further resource use data were collected using participant-reported questionnaires 6 months after surgery and at 12 and 24 months post randomisation. The questionnaires captured details of contacts with primary or community health-care and social care services, prescribed medications, the use of external sheaths or permanent catheters, details of inpatient and day case admissions, and outpatient attendances for reasons related to the surgery or leaking urine. Details of further hospitalisations and treatments for leaking urine and diagnostic tests were collected through a hospital admission CRF that was completed by the recruiting site. The costs of additional sling and AUS surgery related to urinary leakage were estimated using the cost estimates from the study because there is no specific Healthcare Resource Group (HRG) for the use of slings in men. The costs of additional surgery, such as removal and replacement of AUS and excision of slings, were estimated using national tariffs.

The total costs from the health services perspective were calculated by summing all intervention treatment and follow-up costs related to the incontinence symptoms for each participant. Participants were dropped from the complete-case analysis if one of the component costs was missing because of a non-returned questionnaire. In addition, the questionnaires captured the direct non-medical costs incurred by participants, as well as the number of days off work, to inform the gross loss of earnings attributable to the participant’s health state or contacts with care providers.

Valuing resource use

Resource use was valued by attaching unit costs derived from published sources. These included the British National Formulary,19 the Personal Social Services Research Unit’s Unit Costs of Health and Social Care 201818 and the Department of Health and Social Care’s NHS Reference Costs 2017–18. 20 Data on the price of the devices were obtained from the clinical centres . The cost of the AUS consumables consisted of the cost of the control pump, pressure-regulating balloon, cuff with Inhibizone and control system accessory kit. Health services costs were summed across categories to generate a total cost to the NHS of the intervention and the follow-up care for each trial participant. Table 13 gives a summary of the unit costs values and data sources for resource use categories.

Participant perspective costs

Participant perspective costs included out-of-pocket expenditure on incontinence containment products, such as pads and over-the-counter medications, and visits to private care providers to treat incontinence symptoms. The cost of time off work (for those in paid employment) was also included. For time off work, the appropriate national average price of 1 week of work was used (see Table 13). Data for personal incurred costs (i.e. private medical costs, over-the-counter medications and other expenses) are sourced directly from the participant questionnaires. A wider cost perspective was obtained by combining the total NHS and participant perspective costs, and is considered as a sensitivity analysis.

Estimation of quality of life

A generic instrument, the EQ-5D-3L, was used to measure quality of life. Trial participants completed the EQ-5D-3L at baseline, 6 months after their operation, and at 12 and 24 months post randomisation. Responses to the EQ-5D-3L questionnaire were translated into utility values using the UK general population tariffs, based on the time trade-off method. 25 Quality-of-life data were also collected using the SF-12 questionnaire for comparison. SF-12 data were collected at baseline and at 6, 12 and 24 months. These data were converted into a Short Form questionnaire-6 Dimensions (SF-6D) utility index using a published algorithm. 26 Respondents who died during the trial were assigned a utility of 0 at each time point following death. QALYs were derived using an area-under-the-curve approach, assuming a linear extrapolation between utility measurement time points.

Data analysis

The within-trial economic analysis was conducted following the ITT principle. Components of costs and QALYs were described with the appropriate descriptive statistics when relevant: mean and standard deviation (SD) for continuous and count outcomes, and numbers and percentages for dichotomous and categorical outcomes (e.g. numbers reporting problems on EQ-5D-3L). Analyses were conducted using Stata, version 14.1^® (StataCorp LP, College Station, TX, USA). To investigate the potential for skewed cost data (i.e. a small proportion of participants incurring very high costs), we used generalised linear models (GLMs), testing alternative model specifications for appropriate fit to the data. Cost data were analysed using GLMs, with adjustment for minimisation covariates and baseline utility score. 27 The models included a cluster effect for centre. GLMs allow correction for the potential for skewed cost data (i.e. a small proportion of participants incurring very high costs) by allowing specification of an appropriate distributional family and link function to best fit the data. Different distributional families offer alternative specifications to reflect the relationship between the mean and the variance of the estimates under consideration. 28,29 The most appropriate distributional family for each analysis was determined using a modified Park test. Appropriate link functions were identified using several tests (i.e. Pearson correlation test, Pregibon link test and a modified Hosmer and Lemeshow test). The decision on the appropriate link function to use was based on which function performed best on the most of the three tests, as confirmed by examination of the respective p-values of the tests. The process followed Glick et al. 28 and was performed in Stata 14 using the user-written ‘glmdiag’ command. The incremental treatment effect on QALYs and costs was represented by the coefficient on treatment in the respective analysis models. Marginal effects are estimated on the cost and QALY scale, and CIs were calculated using the delta method. The cost data were analysed using a gamma family with a log-link. A standard linear regression (ordinary least squares) model was applied to the analysis of incremental QALY gains.

Missing data

Missing data are a frequent problem in a cost-effectiveness analysis (CEA) within a RCT. Missing costs and health utility data were imputed at each time point using fully conditional multiple imputation by chain equations, implemented through the multiple imputation by chained equations (MICE) package (run within Stata, version 14.0; https://fmwww.bc.edu/RePEc/bocode/m; accessed 5 September 2020) under the missing at random assumption. The appropriateness of the missing at random assumption was assessed by (1) investigating the missing data patterns (monotonic vs. non-monotonic) and (2) comparing the attributes of participants with and participants without missing costs and health-related quality-of-life data at each follow-up time point. Regression models were used to generate multiple imputed Q70 data sets, in which missing values were predicted drawing on predictive covariates based on randomised treatment, type of prostate surgery (radical or TURP), whether or not they have had radiotherapy in addition to surgery, age and centre. Costs and EQ-5D-3L utility scores at each time point contributed as both predictors and imputed variables. Imputations were generated using predictive mean matching drawn from the five k-nearest-neighbours (knn  =  5); predictive mean matching preserves distribution of the data and is more robust to violations of the normality assumption. The multiple imputation was run 20 times, generating 20 complete data sets that reflected the distributions of and correlations between variables.

Incremental cost per quality-adjusted life-year gained

Imputed data were analysed using the appropriately specified GLMs to estimate the costs and QALYs in each treatment group over the 24 months. Joint distributions of costs and outcomes from the original data set were generated through non-parametric bootstrapping and changes in costs and QALYs were calculated for each sample. A total of 1000 bootstrap samples were drawn and means for both incremental costs and incremental QALYs (with associated 95% CIs) were calculated. Measures of variance for these costs and QALYs were derived using bootstrapping. From the results of the bootstrapping, cost-effectiveness acceptability curves (CEACs) were created. CEACs were used to represent whether or not the various interventions are cost-effective at various threshold values for society’s willingness to pay for an additional QALY. CEACs present results when the analysis follows a net benefit approach. This approach utilises a straightforward rearrangement of the cost-effectiveness decision rule used when calculating incremental cost-effectiveness ratios (ICERs; see Incremental cost-effectiveness analysis) to create the net monetary benefit. The ICER provides a ratio of cost per additional QALY gained. The net monetary benefit (NMB) of the interventions in question is equal to:

where λ represents a decision-maker’s willingness to pay for incontinence avoided or a QALY. If the above expression holds true, the intervention is considered cost-effective. Given that society’s willingness to pay is unknown, the NMB was calculated for a number of possible λ values, including the usual £20,000–30,000 for QALY values, which is a threshold often adopted by policy-makers in the NHS. 17

Subgroup analysis

Subgroup analyses were undertaken based on the:

1. type of prostate surgery that the men had – radical prostatectomy or transurethral resection of prostrate

2. volume of urine leaked per 24 hours at baseline – > 250 ml or ≤ 250 ml.

Resource use

Table 15 details the resource use associated with the intervention and follow-up care in slings and AUS groups.

Intervention

Among those who received the intervention, 178 (98%) participants in the sling group received the treatment that they were allocated to and two (1%) received AUS. In the AUS group, 166 (95%) received their allocated treatment and nine (5%) received slings. The intervention time for the male sling group was 96 minutes (SD 39 minutes), compared with 117 minutes (SD 50 minutes) for the AUS group. The rest of the resource use for the initial surgery was similar across both groups.

Follow-up care

The number of further operations is reported in Table 5 (see Chapter 3). The remainder of the follow-up resource utilisation, the use of catheters, sheaths, visits to primary and secondary health-care providers and use of medications, was similar across the groups and there was no statistically significant difference between the two groups (see Table 15).

Intervention and follow-up costs

The differences in costs reflect the differences reported in resource use. Details of costs are reported in Table 16. The cost of the intervention was, on average, lower for the male sling group than for the AUS group. Most of the differences in costs for the intervention components were statistically significantly lower for the male sling group than the AUS group, apart from the theatre, medications and inpatient stay. The main driver of the intervention cost savings for the male sling group was the difference in device costs. Overall, the total cost of the initial intervention was significantly lower in the male sling group than in the AUS group (–£2477, 95% CI –£2935 to –£2017). These findings are based on the complete-case data. The follow-up costs at 6, 12 and 24 months were not significantly different between the groups. However, over the 24-month follow-up, the total NHS perspective costs were lower for the male sling group than for the AUS group (–£2866, 95% CI –£3723 to –£2008).

Private health-care costs

A further analysis incorporated participants’ costs into the analysis over the 24-month period. Table 17 reports mean costs from a wider economic perspective. Like the base-case analysis, there were a lot of missing data and the reported summaries represent all available data. The main drivers of the private costs were the income lost, measured in the number of days off work, and the cost of pads. Overall, the participants in both groups had similar resource use over the follow-up period, apart from the use of pads and visits to the nurse, which were statistically significantly higher in the male sling group than in the AUS group.

Quality-of-life measures

Table 18 provides descriptive data of the mean utility scores and QALYs generated by combining utilities with the duration of follow-up. The differences are based on ordinary least squares models, with adjustment for minimisation covariates and baseline EQ-5D-3L score. The utility scores derived using the EQ-5D-3L indicate that the quality-of-life score at 6 months was the same as at baseline for the sling group and higher for the AUS group, but declined at 12 months for both groups and improved for both groups at 24 months. There was no statistical difference in the 24-month QALY scores between the two groups. The utility scores derived using the SF-6D were higher at 6 months, but reduced at 12 and 24 months for the sling group and reduced at only 24 months for the AUS group. There was no statistically significant difference in the 24-month QALY scores based on the complete data. The visual analogue scale scores were highest at 6 months and lower at 12 and 24 months in both groups.

The proportion of men who reported any health problems on the EQ-5D-3L domains is illustrated in Figures 6–9. The descriptive data are based on all available data reported at all of the time points at which data were collected. They show the percentage of men who reported a problem on each of the domains of quality of life (i.e. a ‘2’ or ‘3’).

FIGURE 6.

Percentage of men experiencing any problems, by randomised group (baseline).

FIGURE 7.

Percentage of men experiencing any problems, by randomised group (6 months).

FIGURE 8.

Percentage of men experiencing any problems, by randomised group (12 months).

FIGURE 9.

Percentage of men experiencing any problems, by randomised group (24 months).

The domain with the lowest percentage of men with problems was self-care. More men seemed to experience more problems with all of the domains at 12 and 24 months than at 6 months.

The male sling group reported statistically significantly lower (worse) EuroQol-Visual Analogue Scale (EQ-VAS) scores than the AUS group over the different time points. The EQ-VAS scores ranged from 73 to 77. They also reported lower SF-6D utility scores at all of the different time points and lower SF-6D QALYs (–3.40, 95% CI –6.60 to –0.19) than the AUS group. The figures of EQ-VAS and SF-6D scores are presented in Appendix 3, Figures 16 and 17.

Summary of the findings

For the base-case analysis, male sling group costs were lower (£2497) than those for AUS. This cost difference was mainly driven by the substantially more expensive AUS kits and the average time spent in theatre for the operation for the AUS group. Although the total costs at the various follow-up times were higher for the male sling group than for the AUS group, they were not statistically significant and did not have much of an impact on the overall total cost. On average, the male slings provided a lower QALY benefit (0.006) than AUS; however, this was not statistically significant. The very high cost difference and low QALY difference led to a high ICER of £425,870. Given that, on average, slings cost less and have fewer QALYs than AUS, the ICER indicates cost savings that will be accrued for loss of QALYs. According to Dowie et al. ,30 the case for adopting an intervention that costs less and has fewer QALYs becomes progressively stronger as the saving from the loss of a QALY increases.

Several deterministic sensitivity analyses were performed to address the areas of uncertainty in data collection. All of these analyses apart from the complete-case analysis were conducted using the imputed data set. The findings were not sensitive to the type of data used and were similar to the base-case analysis. It is reassuring that the results and conclusions of the imputed and the complete-case data did not differ. The probability of slings being cost-effective at a threshold of £20,000–30,000 in both data sets was > 95%. However, the subgroup analysis for men who had TURP showed that the slings were dominant because they had lower costs and higher QALYs than AUS. Data for the men who had a radical prostatectomy were too few to undertake any meaningful analysis. The results for the volume of urine leaked per 24 hours at baseline, > 250 ml compared with ≤ 250 ml, both reported lower ICERs than the baseline analysis.

As mentioned in Missing data, missing data are always a problem for the CEA. The results for those with complete data reported in Table 19 were derived from less than half of the participants (male sling group, n = 78; AUS, n = 75) because some participants’ data were missing at certain time points. Therefore, the base-case analysis was based on imputed data.

The estimates of incremental costs were sensitive to the method used to derive the intervention costs. Despite the fact that there is no HRG for the male slings, Office of Population Censuses and Surveys (OPCS) codes were used to map the costs to HRGs and the cost difference increased from £2497 to £3432 and the probability that slings would be considered cost-effective increased to 100%. When the lowest AUS price was used to estimate the costs of the intervention, the cost difference reduced to £1470 and the probability that slings would be considered cost-effective at the £30,000 willingness-to-pay threshold for an additional QALY reduced to 90%.

Methods

Model type

A long-term Markov cohort decision analysis model was developed in TreeAge Pro™ 2018 software (TreeAge Pro 2018, R1.0. TreeAge Software, Williamstown, MA; www.treeage.com) to represent the treatment pathway for men with SUI. The model was developed to represent progression through the pathway of care. A cohort of 1000 men, with an average age of 67 years (trial baseline population), who were suffering from USI requiring surgical intervention enter the model and receive either a sling or an AUS and follow a treatment pathway based on risk of requiring further surgery. A maximum of three surgeries was modelled.

The structure of the model was developed in collaboration with clinicians and trial collaborators. The model allows the cohort to have further surgery for USI in the following sequence: (1) AUS/slings for the sling group and AUS only for the AUS group. The treatment pathway described in Figure 12 is identical for both model groups, with the only difference being that a proportion of the sling group could receive a sling. The base-case analysis model was run for a maximum of 20 years to take into account the life expectancy of a 67-year-old male in the UK. 31 The cycle length was 1 year and a half cycle correction was applied to the model. Cost and QALYs experienced after the first year were discounted at the recommended 3.5%. 17

FIGURE 12.

Model structure.

After each surgical procedure, the cohort can enter into one of four mutually exclusive states:

1. Well post surgery. This reflects the proportion of the cohort in any given cycle who do not require any further surgery for USI or complications, or who die naturally. They may, however, still experience some USI symptoms or other complications that do not require hospitalisation. Once entering this state, the cohort remains there over successive model cycles until they experience failure or death.

2. Complications. A proportion of the cohort can enter the complications state in the model following surgery, reflecting that men may experience complication risks in the short term following SUI surgery. The complications state reflects major complications that require hospitalisation or further surgery, such as failure of the device. The duration of the complications state is 3 months, reflecting that the complications will be identified early and resolved quickly to enable the patient to return to the well state. As the cohort progresses through successive surgeries for failures, each complication state relates to the most recent surgical procedure. For example, it is assumed that complications following second surgery relate to the second surgery.

3. Failure. Progression of the cohort along the sequence of surgeries is triggered by failure, as described in Figure 12, that is the need for repeat surgery is considered an indication of failure.

4. Containment. The proportion of the cohort having progressed through all surgical options described and still experiencing symptoms indicative of treatment failure is assumed to be managed conservatively and not to receive further surgical intervention. Conservative management includes the use of containment products.

5. Dead. The modelled cohort may die at any point in the model, following general population all-cause mortality rates. No additional surgical mortality rate is included because no one died during the surgical procedures.

Cost-effectiveness analysis

The model analysis was conducted using second-order Monte Carlo (probabilistic or second order) simulation with 1000 iterations. Each iteration resampled a value from each input parameter based on the defined distributions. The model was probabilistic, with 1000 expected values of costs and QALYs generated for each treatment strategy from the sampling distributions. Cost and QALY values were combined into a single measure of efficiency and reported as incremental costs per QALY gained, commonly referred to as the ICER. ICERs can have either negative or positive signs. ICERs with negative signs are easy to interpret. They suggest that the new intervention is either dominant or is dominated. Interventions that generate cost savings and higher QALYs than a comparator offer an even stronger case for cost-effectiveness. On the other hand, interventions that are more costly and generate fewer QALYs than a comparator are said to be dominated and are not considered a cost-effective use of resources. Positive ICERs that are more effective and more costly are assessed based on the maximum acceptable willingness-to-pay threshold. Interventions reporting an ICER of £20,000 to £30,000 per QALY gained are generally considered cost-effective. Conversely, interventions that cost more and generate fewer QALYs than a comparator are said to be dominated and are not considered a cost-effective use of resources. Sometimes positive ICERs are also derived when an intervention costs less but is less effective than the comparator. In this case, low values indicate a low probability of cost-effectiveness.

Cost-effectiveness acceptability curves were generated by assessing the probabilistic net monetary benefit for each treatment strategy. The probability that each strategy is cost-effective at alternative threshold values of willingness to pay for a QALY gain was calculated using the CEACs. Similarly, scatterplots of incremental costs and incremental QALYs were obtained for the male sling compared with the AUS and report the results on the cost-effectiveness plane. Both CEACs and the scatterplots illustrate the uncertainty surrounding the optimal treatment strategy caused by the combined statistical variability in the model’s parameter estimates.

Deterministic sensitivity analysis

Although the CEACs illustrate sampling uncertainty, they do not address underlying uncertainty driven by choice of data, structural assumptions, methodological choices or heterogeneity across subgroups of the modelled cohorts. Several deterministic sensitivity analyses were conducted to address the main areas of uncertainty in the model and to test the effect of the assumptions on the cost-effectiveness conclusions. All deterministic analyses are subjected to the same sampling uncertainty as described for the base case, ensuring that all presented ICERs are fully probabilistic apart from the one-way sensitivity ICER. Analyses were conducted to explore uncertainty surrounding the choice of model utilities and costs.

The base-case analysis uses treatment-specific utilities and costs for each health state in the model. These data are advantageous in that they preserve the impact of treatment on the costs and quality of life in any given health state. However, for rare events, such as the complications in our data, they were derived from trial data with a follow-up period that registered many events. Sensitivity analysis explores the impact of using utilities and costs that were defined for each modelled health state, but not split according to randomised treatment in the trials. This increases the sample for estimating costs and utilities but removes any assumption about treatment-specific effects on these parameters for the model.

Sensitivity analysis was also undertaken to determine the impact of methodological uncertainty on the results. Long-term data on the time to failure and complications are lacking and, therefore, estimates have been developed for the model based on the probability of events. The impact of any modelling inaccuracies on cost-effectiveness is likely to be magnified over longer-term time horizons. Sensitivity analyses, therefore, shortened the time horizon to 5 and 10 years.

Sensitivity analysis was also conducted to determine the impact of methodological uncertainty surrounding the appropriate discount rate to apply to future costs and QALYs. Sensitivity analyses vary the discount rate between 0% and 6% in line with NICE best practice guidelines. 17

Cost-effectiveness results

The base-case analysis over a 20-year time horizon indicates that the sling group had lower costs (£1511, 95% CI –£4597 to £5577) but also fewer QALYs (0.133, 95% CI –0.782. to 0.488) than the AUS group. The ICER is £11,385, which means that for each QALY lost there is a cost saving of £11,385. These are reported in Table 23. The scatterplot of cost-effectiveness is reported in Figure 13. The base-case CEAC illustrating uncertainty is reported in Figure 14. The results indicate that the probability that slings would be considered cost-effective at the £30,000 willingness to pay for a QALY threshold is 42%.

FIGURE 13.

Scatterplot of incremental costs and QALYs: male sling vs. AUS base-case model analysis.

FIGURE 14.

Cost-effectiveness acceptability curves: male sling vs. AUS base-case model analysis. WTP, willingness to pay.

Sensitivity analysis

The results from all sensitivity analyses are reported in Table 23. On average, the cost difference in most of the analyses follows the same direction as the base-case analysis, although the magnitude in the differences reduces in most of the analyses. However, the results of the analyses that relate to non-treatment-specific costs and QALYs indicate that slings cost less and are more effective than AUS and, therefore, slings dominate. The results are sensitive to the probability of failure and complications as the costs of slings increase and QALYs reduce and the AUS dominates.

Interpretation of the model results

The model illustrates the long-term trade-offs between the choice of a male sling and the choice of an AUS and is populated with the data from the trial on failure and complications needing further treatment. The results are mainly driven by the higher QALYs experienced by those who enter the well state in the AUS group. The long-term results are similar to the within-trial analysis in that slings cost less but are still less effective than AUS. However, the magnitude of the ICER reduces over time, which means that the cost savings that are gained from the reduction of QALYs reduce. This also leads to a reduction in the probability that slings would be considered cost-effective at the £30,000 willingness-to-pay threshold, as illustrated in the results.

One of the main weaknesses of the model is the lack of long-term failure and complication data for the use of slings in men. The existing evidence tends to report success based on the number of pads used rather than the number of further surgeries. According to Leon et al. ,34 the major postoperative complication of the AUS in the long term is the need for explantation of the device because of erosion or infection. Studies suggest that the AUS has an average lifetime of 10 years,35,36 but there is no evidence of the average lifetime of male slings; therefore, this was not included in the model. Therefore, these results need to be interpreted with caution because the 2-year data are immature and there is a chance that all the important trade-offs, particularly complications, have not been observed in both groups. The ICER is sensitive to the long-term failure and complication values. There is currently too much uncertainty around the failure and complication rates and further follow-up is needed to reduce the uncertainty.

These results must be interpreted with caution, with the understanding that they would apply only in a situation in which relative differences in the effectiveness of the male sling compared with the AUS do not change over time. It is not clear at this stage if this assumption is valid and long-term follow-up is needed to inform this.

Methodology common to all studies

Qualitative study participants

Participant details are provided in Table 24 and demonstrate comparable groups at each stage of the investigation in terms of age distribution.

Aims

Focus groups are used to allow unstructured discussions on particular topics with a group of potential participants to help to guide the study. 42 The areas for exploration for this focus group were:

• important outcomes post surgery for the men

• factors affecting men’s expectations following surgery and satisfaction post operation

• men’s general comments and suggestions for amendments to the self-report questionnaires for inclusion as study outcomes.

Methods

Initially, patients who had not yet been invited to take part in the trial, but were otherwise eligible, were invited to take part in a focus group to ensure that the study was clear in its direction from the men’s perspectives. Twelve patients on the waiting list for surgery at Southmead Hospital (Bristol, UK) were invited and sent a participant information sheet and consent form; four men consented to take part.

Results

The men were aged 63–71 years. All were white, two were retired, one was semiretired and one was in full-time employment. They had PPI and were able to provide insights pertinent to the trial intention (Table 25).

Desired outcomes from surgery

There was a consensus between the men that the inconvenience associated with pad use was a central factor in their decision to seek treatment. Indeed, when asked to report on what the ideal outcome from surgery would be, the immediate answer from one participant was ‘not having the pads’. This was met with general agreement from the group. One of the men admitted to using up to 13 pads per day, before preferring to use a Conveen sheath^® (Coloplast Ltd, Peterborough, UK), although this was also not an ideal solution:

I’m finding that the Conveen . . . sometimes make me sore.

P2

However, for most of the participants, the general inconvenience of pad or Conveen was not a barrier to leading a full and active life:

Not at all. I’m still walking, I’m still working, I still do a lot of things. I’ve never changed my lifestyle at all.

P1

Some of the men described how they were able to manage their symptoms to minimise the impact on their daily lives. Others described how leakage was a major inconvenience:

I manage it during the daytime – where is the nearest loo? The whole day was ‘where am I going to be’?

P3

Driving is a nightmare. It’s OK when you know that you’re going because when you’re sat in the car it’s OK, but as soon as I stand . . .

P4

Although the cessation of pad use was agreed to be the ideal outcome for their forthcoming surgery, this did not extend to the complete control of leakage for all of the participants. One man deemed it an acceptable outcome to his surgery if he was able to control his leakage during the day:

Yes. If that could be strengthened a little bit longer into the afternoon. If that was like it all day, I wouldn’t bother at all.

P1

Embarrassment caused by leakage was a recurring theme from each of the participants, ranging from brief annoyance to an overwhelming impact on their lives:

For me it’s very embarrassing. I lead a very active life.

P1

I personally . . . from an embarrassment point. At very odd times I have leaked.

P2

The biggest embarrassment is if you have a partner . . . When you are in bed and it happens to me, and you end up peeing on her bum.

P2

Really bad. I haven’t a lot of confidence. I hate it, I really hate it.

P4

Embarrassment extended to the impact on relationships, self-esteem and confidence as important factors. In particular, this was mentioned as having a major effect on their sex lives, including on sexual confidence, which was also compounded by the side effects of the prostate surgery.

There was some recognition that sex was perhaps less frequent than when they were younger. Nevertheless, it was still stressed by several of the men as being a very important part of their lives and loss of full sexual function was recognised as a threat to the masculine identity:

I mean because there is the sex life business. You become insecure. That’s number one.

P4

Then I had a lot of trouble with erections. But then I was bent a bit. You can have an operation for that as well, but I didn’t think I wanted to go for that . . . I’m married, our sex life is not what it was when we were young but you know.

P4

But at the same time, you’re a man. Like for her, if you lose your breasts you don’t feel a full woman any more, you don’t feel a full man any more . . . If you’re married it’s probably fine but in my case I’m separated and if things start to come together you think ‘oh no, my God’, it’s an awful thing to have to explain . . .

P3

Factors affecting satisfaction with treatment

The expected outcome of the men’s treatment was highly influenced by the predicted prognoses of the medical professionals. After prostate surgery, when the side effects did not improve in line with their expectation of the outcome of treatment set by the medical staff, this had a large effect on their level of satisfaction:

They told me I would be dry within 3 months. Didn’t happen.

P1

Yes, I’d been told. You’d be dry within 3 months.

P2

Other factors that contributed to the level of satisfaction of treatment included the level of confidence in their surgeon and other staff. This was influenced by their personal perception of competence, but also by familiarity:

The guys here at [name of hospital] know me, they know all about it. The guys in [name of hospital] didn’t know about it.

P1

I had a lady come in one morning at half seven to do my dressing. She had gloves on and she moved the bed because she said there wasn’t room for the trolley. Then she proceeded to clean me up with the same pair of gloves.

P2

The length of stay in hospital after surgery, length of operation, scarring, risk of infection and general fear of surgery were mentioned in the course of the discussion as being important factors to the men. However, this was not universal:

I have had a lot of operations in my life . . . I’ve never been afraid of surgery – it’s always been good.

P4

Perspectives of the intended trial outcome measures

Participants in the focus group were also asked for their views on the appropriateness of the intended outcome measures. The questionnaires that they were asked to comment on were the International Consultation on Incontinence Questionnaire Lower Urinary Tract Symptoms Quality of Life (ICIQ-LUTSqol) module43 to assess the impact on quality of life; the ICIQ-MLUTS,44 which incorporates evaluation of urinary storage, and voiding and incontinence symptoms; and the ICIQ-MLUTSsex45 to assess male sexual matters associated with their LUTSs and impact on quality of life.

The focus group confirmed the importance of leakage and sexual matters as outcomes, which are well covered by the baseline questionnaire and the key outcomes that matter to individuals with PPI. The baseline questionnaire also covered the medically related outcomes well, for example the volume of leakage, urgency, nocturia and depression. However, it was not picking up on the impact of symptoms and issues, such as inconvenience to daily life, embarrassment, distress, frustration, effect on relationships with family and friends or sexual confidence. It was suggested that the inclusion of the ‘how much does this bother you?’ items from the ICIQ-MLUTS and ICIQ-MLUTSsex would provide an evaluation of the impact of these issues on patients. This was intended to achieve an evaluation of the differential effects of the operations on other outcomes, such as quality of life and general health.

Although inclusion of a condition-specific quality-of-life questionnaire, such as the ICIQ-LUTSqol, was considered advantageous, it was considered more appropriate for a later-stage study in the evaluation of these surgeries. The SF-1215 was included, recognising that it may lack sensitivity for the specific issues highlighted by this focus group but would enable the capture of high-level quality-of-life data.

Discussion

The men who took part in the focus group discussed their satisfaction with surgery as being dry, using fewer or no pads, being able to regain their social life and having improved sexual matters. It was felt by some that the use of pads was an inconvenience and that it had an impact on the activities that they felt able to do. This confirmed the importance of the areas covered by the baseline questionnaires. It also highlighted that complete cure of incontinence was not the only outcome to secure satisfaction with treatment. These findings were important to underpin the study and ensure measurement of the most appropriate indicators of outcome.

Aims

Initially, this study explored the rationale behind participants’ decisions to be randomised or not randomised to suggest possible recommendations for the improvement of recruitment strategies. Many trials struggle to recruit enough participants, threatening their overall validity. Any subjective view of the characteristics of each surgical procedure may have implications for participants’ eventual choice. The following provides an insight into how participants perceived the male sling and AUS prior to surgery. Participants’ expectations for the outcome of their surgery were also explored, as satisfaction with a procedure is known to be closely linked with met expectations. 46,47 This had the intention of informing the criterion of non-inferiority for the two surgeries.

Methods

The second subgroup were recruited between February and August 2014. Purposeful selection of participants was undertaken to represent both the randomised cohort and the NRC. All interviews were conducted before the scheduled surgery and within approximately 1 month of initial recruitment to MASTER. The foci of investigation with this group were:

• What are participants’ reasons for choosing to be randomised or to decline randomisation?

• How do participants perceive each of the two procedures?

• What are participants’ expectations for the outcomes of their surgery?

Results

Twenty-five men, aged between 51 and 76 years, from seven centres, were recruited to participate in this subgroup [10 consented to randomisation and 15 consented to the NRC (seven men in the NRC had the male sling and eight had the AUS)]. The main purpose of this substudy was to explore motivations surrounding trial participation to extend learning regarding trial conduct. Experiences of symptoms following initial prostatectomy surgery were also explored in this baseline group to examine the breadth of symptoms across the cohort (Table 26).

Perspectives on randomisation

Seven main themes were identified as prominent when considering the main reasons for men deciding whether or not to be randomised. These were (1) current lifestyle, symptoms and situation; (2) ‘doing your bit for society’; (3) ‘equipoise issues’; (4) ‘leave it to the experts’; (5) randomisation issues; (6) having a strong preference; and (7) centre inconsistencies.

Perceptions of the procedures

Every participant was asked for their opinion on each of the two procedures and had a good or reasonable grasp of the main differences between the surgeries. For this reason, the following sections present the views of the participants from all interviews, both randomised and non-randomised.

Outcome expectations

The main hopes for post-surgery outcomes were explored, with participants’ responses largely aligned with the participant reports from the initial focus group. Primary expectations were for an improvement in incontinence and for a reduction or complete cessation of pad use or other continence aids. However, when questioned further, there were many other expectations and hopes for positive changes to their lives. The following section presents the views of the participants from all interviews, randomised and non-randomised.

Discussion

Aims

The focus of investigation with this group was to explore the experiences of participants who had undergone both trial procedures. This investigation explored two distinct areas:

• surgical recovery – how participants felt during their recovery period, how long they spent in hospital, the symptoms that they experienced, the immediate perspective for those who had the sling and for those who had an AUS fitted, and their experience of the AUS being activated

• life after surgery – what differences surgery had made to their life, how successful their surgery had been and the symptoms they experienced post surgery.

Methods

The third subgroup was recruited 6–24 months after their surgical procedure to explore the post-surgical experience of men undergoing a sling or an AUS within MASTER. They were interviewed once and the interviews were conducted between January 2015 and November 2018.

Results

Having extensively investigated the men’s perspectives pre operation, we compared the experiences after surgery for both procedures and evaluated the differences in experiences between the two groups (Table 27). Fifteen men were recruited to this subgroup, aged 58–74 years: seven men who received an AUS and eight men who received a sling.

Surgical recovery

Despite some recall difficulties, most men reported that hospital stays ‘appeared’ relatively short. Few men who received the sling experienced an immediate improvement post surgery, although there was some report of slight improvement. By comparison, those who had an AUS fitted, although they had to wait 4–6 weeks after surgery for it to be activated, found it to be largely successful, with many experiencing positive results immediately after the AUS had been activated. Discomfort post surgery was experienced by all; however, there were varying degrees of discomfort, with some finding it better than expected and others experiencing swelling for a few weeks after. This resulted in some pain when trying to use their AUS once it was activated. Further information regarding wound care was highlighted as an unmet need for some men. This has potential implications for improving the patient experience. Realistic explanation of the recovery period and timelines for improvement are key to managing men’s expectations in the post operative period. Most men interviewed highlighted that any reported issues largely resolved over time.

Life after surgery

For both procedures, most men agreed that their quality of life had improved and, in line with reported expectations pre operation, complete continence was not necessarily required for a successful outcome. Inevitably, experiences varied from being straightforward and without incident to having complications along the way. It was recognised that there were variable circumstances for these experiences that were not necessarily related to the procedures.

The impact of success following surgery was farther reaching than symptom experience, with many men feeling that they were able to return to a sense of normality that reflected their life prior to their original prostate surgery. The associated emotional boost described by some provided a new lease of life that led to endorsements for both procedures. For some men, however, surgery was unsuccessful and the requirement for further intervention is discussed below.

Aims

In the previous substudies, we reported that surgery was regarded as successful by some men, but some men did not achieve the outcome that they were hoping for. During the study period, some men underwent further surgery and it was important to explore the reasons for this and the eventual outcomes.

Methods

The final subgroup were recruited from the main trial participants who had undergone an initial surgery for their PPI but required further surgery. The qualitative study team were informed of those requiring subsequent surgery and were all invited to participate to give an account of their opinions and the variability of reasons for further surgical input. The focus of enquiry with this subgroup was:

• reasons for further surgery – why patients needed a second operation, what their symptoms were after their first operation and how they felt about this

• feelings after second surgery – the results of the second surgery, their perception of how successful it had been and the impact that it had on their lives.

Results

Ten men aged 58–73 years who consented to participate in this substudy were interviewed. Three men had received an AUS as their first surgery and the other seven had received the sling; nine men went on to have an AUS fitted as their second surgery and one who had received an AUS was waiting to hear from the consultant about reoperation at the time of interview (Table 28).

Discussion

Aims

As discussed earlier, the surgeons counselling men for these procedures are influential in the decision-making process and are key to the discussions that underpin surgical options. It was, therefore, imperative that we explored the perspectives of these surgeons regarding the clinical experience of conducting these procedures, as well as their viewpoints on the relative merits or limitations of each.

Methods

Surgeons performing both procedures within MASTER were invited to participate in interviews. The experience of performing these operations and the surgical considerations involved are crucial to understanding how the findings might be taken forward by clinicians. A total of 20 surgeons were recruited. Areas for exploration included:

• surgical experiences – the perspective of the surgeons of each procedure, the differences and similarities between each, complications and duration, and if surgeons have a preference and any reasoning behind this

• clinical experiences – looking at the suitability of patients for each surgery and those patients who have previously undergone radiotherapy, the perspective of the surgeons on their patients’ preferences and meeting their expectations, complications that can occur and the surgeons’ own perspective on the procedures (Table 29).

Results

Discussion

The surgeons who were interviewed largely described increased confidence and expertise with performing AUS surgery given the longstanding history of its use and their perceived surgical competence with the procedure. Some were more familiar with the sling and reported increased use in their centres over recent years, which fosters confidence. Consensus was evident that the length of surgery is shorter for the sling, but there were mixed views on the perceived ‘simplicity’ of the procedure.

Both procedures present considerations with regard to complications. The AUS has more components that cause concern around direct complications, such as infection and erosion, whereas problems with voiding dysfunction are perceived to be a bigger issue for those undergoing sling insertion. Concern regarding potential long-term complications of the sling were expressed, particularly given the lessons learned from female mesh procedures, which may have an impact on surgical decision-making. Patient characteristics also influence surgical recommendations, with age, lifestyle and previous medical procedures guiding preferences.

It was collectively recognised that patients are seeking a return to normality in their symptom management and a reduction in product use, and not necessarily complete resolution of incontinence symptoms. This supports earlier findings from the men’s perspectives and is appropriate for the likely outcomes. The surgeons who were interviewed described discussing realistic perspectives of outcomes with patients and that complete ‘dryness’ is less likely. It was recognised that patients present with their own preferences that are shaped by varying sources of information and influence the clinical exchange when deciding the appropriate surgical directions. The perceived simplicity of adapting to life following surgery was also recognised as an influencing factor, with an assumption that men prefer the straightforward option afforded by the sling compared with the learning required to use the AUS.

These insights have relevance with regard to influencing surgical choices and beyond the study when data can influence these preferences further. It is recognised that surgeons are a key influence in clinical decision-making for men with PPI, and these decisions will be informed by the findings from MASTER.

Expectations regarding outcomes

Of interest at the outset was that complete continence, although desired, was realistically not the expected goal; this was expressed by both the men and the surgeons. Achieving a satisfactory outcome was dependent on the improvement in continence rather than the complete resolution. They hoped that surgery would allow them to regain their social and sexual lives and to be in a position where they needed to use fewer or no pads or incontinence products, as these were restrictive in their lives. Being able to return to ‘normal’ was key. The resumption of activities and engagements that had ceased owing to leakage, the improvements in confidence and self-esteem, and the reduction in negative emotions were intrinsically linked and important outcomes for men with PPI.

Expression of preferences

Preferences among both men and surgeons emerged as a noticeable finding within the qualitative research and, in the absence of a robust evidence base, the sources on which to base these preferences were explored. Participants often had preferences based on complex considerations and varying sources of evidence: expectations for outcome, desire to minimise the invasiveness of surgery, concerns around operating a device, their personal and work life circumstances, requirements for heavy lifting, hobbies such as cycling, willingness to use a cubicle instead of a urinal, the internet and anecdotal evidence. Key to influencing men’s perspectives was the opinion of their surgeon. Respecting the experts’ opinion was a driving force behind treatment preferences, which highlights the importance of the surgeons’ perspectives being grounded in the evidence where available.

Surgeon preferences were two-fold: based on technical aspects of the two procedures and clinical perceptions of outcomes. There was a technical preference towards the AUS owing to most surgeons having greater experience in fitting this device than the newer sling. Technically difficult aspects of sling surgery were discussed, but there was recognition that it was a quicker procedure that has benefits in itself, with fewer components that can cause problems such as infection and erosion. The lack of long-term data for the sling was also a consideration, particularly in view of the emerging picture in the field of female mesh surgery. 56,57 Many also perceived that their patients had an inclination towards the sling because it required fewer mechanical parts. This was not reflected in the discussions with patients, in which we saw a split, with some being attracted to the sling because of its ease of use and others liking the control of the AUS and understanding its mechanical function better.

Surgeons’ preferences were also based on clinical judgement of their patients, with many reluctant to put a sling in a man with heavy incontinence, judging that it would be more suited to a man with milder incontinence symptoms. Previous history of radiotherapy was also a consideration, along with the type of activities that the man was likely to undertake, expressing that one procedure or the other was more appropriate for different circumstances. It was highlighted by the surgeons that without the findings from MASTER there is no evidence base on which to make these recommendations and the gravity of the decisions being made cannot be overlooked.

Methodological considerations

Methodologically, the existence of preferences had an influence on the study that required further consideration. Participants reported a desire to be altruistic and support research for the benefit of others, but there existed a spectrum of willingness to consent. For those who had no surgical preference and did not feel best placed to make a decision, the process of randomisation assisted the treatment pathway; however, in cases where an individual had a clear preference or a surgeon expressed a likelihood that one procedure would be more suited to them, this was prohibitive to the randomisation process. Provision of the NRC provided a vehicle to still contribute to the research findings but also receive their surgical preference; however, this was somewhat of a distraction to the main study aims, despite its original intentions. This has implications for future research and contributes to the growing body of evidence surrounding the mechanisms of trial conduct and the importance of clarity regarding central trial processes, such as randomisation. 58,59

Procedure comparisons

The participants described some soreness and discomfort post surgery for both procedures, with some experiencing more than expected. Most were satisfied with their eventual outcome, but stressed the importance of fully understanding the recovery period pre operation. This in-depth exploration provides unique insights to inform patient information and counselling for surgery to fully inform patients of the likely recovery trajectory. It is crucial to manage the expectations of men following surgery, and the experiences described above can be used to provide more accurate information for men undergoing these procedures in the future.

Participants who had a successful surgery, whether it was a sling or an AUS procedure, found the results life-changing, with a positive effect both physically and mentally. It had a significant impact on their quality of life. It gave the participants back their independence and, although some men still sustained some leakage, it was vastly improved from their earlier symptoms. Having experienced the previous increased levels of incontinence, most were satisfied even with a small volume of leakage as it did not control their lives the way that their symptoms had prior to surgery.

For those for whom the treatment failed, it was a significant disappointment and threatened their belief in ever recovering in some cases. There were frustrations in the length of time that it took for men to receive further surgery when they discovered that their original surgery was not successful. They perceived that they were ‘back to square one’ instead of progressing along a treatment pathway and felt that their subsequent procedure should be expedited. Despite this perceived loss of time, men were satisfied generally with their final outcome following further surgery. Among the 10 men who were interviewed who required further surgery, seven had previously had a sling procedure. The main trial findings provide robust data comparisons for the success of the two procedures, but it is important to raise the problem caused by sling failures here. Within the preference findings, it was often described that the sling was viewed as a possibly ‘simpler’ treatment prior to a potentially definitive AUS and this was borne out in these interviews. This narrative will require exploration in the light of the overall study findings to ensure that the place of both treatments in the surgical armoury is evidence based and accurately communicated to inform decision-making going forward.

Reflections for future practice

The unique insights gained through these in-depth interview studies have provided first-hand accounts of the experience of these two procedures, in the words of the men themselves. The knowledge derived from these qualitative studies provides key considerations for improving future practice when in discussion with men considering PPI surgical procedures and also for the methodology of future trials.

Both surgeons and participants were clear about the likely successes of surgery overall, with symptom alleviation rather than symptom resolution being the primary expectation. This is a strong foundation and identifies that robust counselling exists within the clinical trial situation. These findings should be helpful when used in routine clinical practice by providing an opportunity to improve further the counselling and management of expectations pre operation, as guided by the evidence included in this chapter. Clarity regarding recovery periods, symptom alterations and comparisons of the two procedures, in the words of the men themselves, will promote readiness for surgery and manage expectations following surgery, which is directly related to perceived satisfaction with outcomes. Ensuring that these findings are implemented into practice is key to mobilising the knowledge gained from this specific element of the MASTER study. Alternative sources of patient counselling to the traditional patient leaflet might be considered to take advantage of the valuable findings from this study and include the words of the men in counselling formats; for example, explainer videos. One study found that showing patients a video of the procedure that they were going to have led to great patient satisfaction than giving them a leaflet. 60 This is something that could be explored further because patients may find the video easier to understand and, therefore, have a better pre-surgery education.

Second, this study has highlighted the clear preconceptions that exist surrounding these procedures that, to date, have been reliant on experiences and opinions without the high-quality data that come from a robust study, such as MASTER. These findings exist for the benefit of both the men and their surgeons. Going forward, it is imperative that the narrative that surrounds decision-making in relation to these surgical procedures reflects the evidence-base to appropriately inform patients of the considerations that apply to their personal circumstances. Within the trial confines, it was easier and necessary to advocate a position of equipoise for the purposes of study fidelity and owing to the lack of sufficient research evidence to suggest otherwise. Beyond the trial, findings can be interpreted to improve the advice provided to men considering surgery for PPI, but it is recognised that these emerging findings will have to be assimilated within the context of some entrenched beliefs. Widespread dissemination among the surgeons providing these procedures is essential and particularly relevant for those who were not part of MASTER.

Last, we identified issues around trial methodology that provide learning for future studies. Inclusion of the NRC provided an ideal vehicle for the collection of data for men who were not prepared to be randomised, which offered additional findings for consideration. However, our findings supported other in-trial findings that suggested that the NRC detracted recruitment to the main randomised trial. The NRC was, therefore, closed to avoid this. The qualitative enquiry provided invaluable insights into this element of trial methodology to enable us to respond within the trial timeframe and achieve the outcomes of the main study. Interesting findings were also highlighted about how researchers discuss trials and how the language used can infer altered messaging to participants, which is vital for the informed consent process. This contributes to the growing body of evidence regarding the process of trial recruitment and reinforces existing findings regarding the importance of focusing on getting the consent stage as good as it can be to improve trial outcomes from the outset.