Open urethroplasty versus endoscopic urethrotomy for recurrent urethral stricture in men: the OPEN RCT

Urethral stricture disease

The male urethra is an epithelium-lined tube enveloped by smooth muscle erectile tissue (corpus spongiosum) and an outer layer of striated muscle (bulbospongiosus). It runs from the bladder outlet (neck), through the prostate gland, perineum and the ventral (under) aspect of the penis, opening at the fossa navicularis at the penile tip (external urethral meatus). It is anatomically divided (from proximal to distal) into prostatic, membranous, bulbous and penile segments (Figure 1).

FIGURE 1.

Pictorial representation of male urethra. This figure was published in Surgery, Vol. 35, Watkin and Patel,4 The diagnosis and management of acquired urethral stricture disease, pp. 313–23, © Elsevier 2017. Reproduced with permission.

This figure was published in Surgery, Vol. 35, Watkin and Patel,4 The diagnosis and management of acquired urethral stricture disease, pp. 313–23, © Elsevier 2017. Reproduced with permission.

It functions as an active conduit to void urine and thereby empty the urinary bladder when desired. 5 It is also the conduit for ejaculation and emission of semen, which enters the urethra from ejaculatory ducts running through the prostate.

Urethral stricture is caused by annular scar formation in the epithelium and underlying smooth muscle layers, resulting in narrowing of the lumen, loss of propulsive muscle action and restriction of urine flow. Published estimates of the prevalence of urethral stricture disease among adult men derived from hospital activity registries in the USA include 0.6%6 and 0.9%. 7 Urethral stricture disease (OPCS Classification of Interventions and Procedures version 3 – character primary diagnosis code N35) resulted in > 17,000 admissions of men to NHS hospitals in England during 2016–17. 8 Men notice a reduction in the strength of urinary stream, prolonged voiding time and slow emptying of the urethra, often with dribbling of urine after micturition. When these symptoms become intolerable men seek medical help and are generally assessed by a urology specialist. Stricture site and length are evaluated by performing either a telescopic (urethroscopy) or radiological (urethrography) examination of the urethra or both (Figure 2). Severity of urinary obstruction is assessed by measurement of maximum urinary flow rate (Q_max) and residual urine after voiding (Figure 3). Most strictures are located in the bulbar segment of the urethra between the pelvic floor and penoscrotal junction. The cause of stricture formation is most often unknown, but it may result from previous injury or infection (urethritis).

FIGURE 2.

Ascending urethrogram showing 2-cm bulbar urethral stricture. Courtesy of Nick Watkin, St George’s University Hospitals NHS Foundation Trust.

FIGURE 3.

Uroflowmetry record of man with bulbar urethral stricture. The flow is slow up to a maximum of 6 ml/second (normal ≥ 15 ml/second) and prolonged. Courtesy of Alison Bray, Newcastle upon Tyne Hospitals NHS Foundation Trust.

Treatment options for bulbar urethral stricture

Men with bulbar urethral stricture have two management options: (1) endoluminal treatment, in which the stricture is disrupted from within the urethral lumen (urethrotomy or dilatation); or (2) open surgery (urethroplasty), in which the stricture is approached through an incision in the perineal skin and excised or bridged by interposition of a healthy tissue graft. Unless there are complicating factors, men with a first stricture occurrence generally choose to undergo an endoscopic urethrotomy, as it is a straightforward procedure and can be carried out under the same anaesthetic following diagnostic urethroscopy. A review of case series suggested that following this first urethrotomy the rate of recurrence at 2 years was 40% and 60% at 8 years. 9 For those men whose stricture recurs, the choice is between further urethrotomy (or simple dilatation for short flimsy strictures) and open urethroplasty. Urethroplasty is a more complex procedure, but case series suggest a lower stricture recurrence rate of around 15% at 2 years and 20% at 6 years. 10–12 Management discussions between urologists and men with recurrent bulbar stricture centre around the choice between these two strategies: (1) planning and carrying out urethroplasty surgery with a view to longer-term freedom from recurrent stricture, or (2) further urethrotomy repeated as needed at stricture recurrence. The relative efficiency of these two strategies for the management of recurrent bulbar urethral stricture in men is the focus of this trial.

Interventions under study

The experimental intervention strategy is open urethroplasty. 13 The bulbar urethra is approached through a longitudinal incision in the perineal skin behind the undersurface of the scrotum. The bulbar urethra is dissected free and the strictured segment incised in either the dorsal or ventral surface. The area of the stricture is then either removed and the cut ends rejoined (anastomotic urethroplasty) or augmented by insertion of a tissue graft. 14 The tissue graft (typical dimensions 5 × 2 cm) is harvested from the mucosa of the mouth. The urethra is then retubularised and the repaired area protected by passage of a urethral catheter to drain urine. Urethroplasty requires a general anaesthetic, takes 2–3 hours to perform and is followed by a 2-week period of urethral catheterisation postoperatively. Median hospital stay in the UK NHS is 2 days. 8

The control intervention strategy is endoscopic urethrotomy (also known as direct visual internal urethrotomy, internal urethrotomy or optical urethrotomy). 15 This procedure involves passing an endoscope under direct vision down to the distal end of the stricture. 16 The narrowed area is then progressively incised from distal to proximal with a straight longitudinal cut, either using a steel blade (‘cold knife’) mounted on the endoscope17 or under endoscopic control with a diathermy needle or laser fibre (‘hot knife’). Once the lumen has been incised sufficiently to restore a normal calibre (approximately 6 mm diameter) the endoscope is withdrawn and a urethral catheter passed to tamponade bleeding from the cut area and provide reliable urine drainage during the postoperative period. If the stricture is short and flimsy, simple dilatation with the endoscope or graduated dilators may suffice without incising the urethral mucosa and underlying corpus spongiosum. Endoscopic urethrotomy requires a general anaesthetic and takes approximately 45 minutes to perform. The median hospital stay in the UK NHS is 1 day. 8 The urethral catheter is typically removed 24–48 hours postoperatively. Men having repeated urethrotomy may be offered training for a programme of intermittent self-dilatation, using an appropriately sized soft plastic catheter. This adjunctive intervention appears to lessen the risk of recurrence over the subsequent 12 months. 18

Both procedures have a similar spectrum of complications, predominantly bleeding from the wound site and urinary infection. If an oral mucosal graft is used to augment the urethra, then there are specific complications related to the donor site in the mouth.

Current management guidance

Evidence for effectiveness of optical urethrotomy and open urethroplasty for treatment of bulbar urethral stricture has been summarised in four systematic reviews. 13,15,19,20 The Cochrane review,19 up to June 2012, found no randomised trials comparing outcomes after urethrotomy and urethroplasty and could not make any conclusion regarding comparative effectiveness. An updated Cochrane search performed to August 2017 found no more recent relevant completed randomised controlled trials (RCTs) or any other RCTs in progress. Reviews underpinning published guidelines, sponsored jointly by the Société Internationale d’Urologié, the International Consultation on Urologic Disease13,15 and by the American Urological Association,20 included non-randomised studies, predominantly of retrospective cohort design. Guideline formulation subsequent to these reviews was predominantly based on expert opinion and panel consensus. Guidance from the Société Internationale d’Urologié15 suggests that optical urethrotomy is an appropriate management option for initial treatment of bulbar urethral stricture, with an overall long-term success rate of about 50%. The guidance also states that the first recurrence of bulbar stricture could be treated with urethrotomy as long as no adverse factors, such as recurrence at < 3 months post surgery, are present. Success rates in terms of up to 2 years freedom from recurrence in the case series varied from 53% to 95%. Buckley et al. 15 recommended that urethrotomy should not be used for second or subsequent recurrences. Urethroplasty with scar excision and primary mucosal anastomosis or use of an interposition graft of oral mucosa was recommended for complex or recurrent bulbar stricture. 13 The more recent systematic review and guideline from the American Urological Association also did not find any robust evidence on which to base clinical practice recommendations. 20 Wessells et al. 20 suggested that both optical urethrotomy and urethroplasty were appropriate initial procedures for men with bulbar urethral stricture. For those men with recurrent stricture, they considered that only urethroplasty would potentially give long-term symptom control. Further recurrence after repeat urethrotomy was regarded as highly likely and therefore should be considered only as a palliative management option. There is little published qualitative research on patient experience of living with and managing the condition. Patients do appear most concerned regarding the impact of their urinary symptoms, particularly difficulty voiding and post-micturition dribbling, on their daily activities. 21 Agreement between patients and specialist clinicians regarding ranking of importance of particular symptoms is poor. 21

Evidence regarding choice of urethrotomy and urethroplasty

Guidance from two professional urologist organisations, the Société Internationale d’Urologie and the American Urological Association, is clear that open urethroplasty is the recommended option for men with recurrent bulbar stricture. However, this recommendation is based on low-level evidence and expert opinion. Both guideline panels were predominantly made up of urologists specialising in urethroplasty, without general physician or patient representation. Registry studies suggest that optical urethrotomy is more frequently used than urethroplasty for recurrent strictures. 22 Case series from a number of health-care systems also suggest that repeated urethrotomy is preferred in practice for recurrent bulbar strictures,23–26 with urethroplasty reserved for complex strictures. Reasons for the more frequent choice of urethrotomy for treatment of recurrent bulbar stricture may include patient and clinician preference, restricted availability of urethroplasty, health service organisational issues and lower cost to the patient and health-care provider. Information regarding management of men with urethral stricture disease and costs in the UK, derived from NHS England hospital activity and tariff data8 from April 2016 to March 2017, showed a total cost of £17.8M, comprising 742 urethroplasty procedures [OPCS Classification of Interventions and Procedures version 4 (OPCS-4) M73.6], with a mean tariff cost of £4157 (Healthcare Resource Group code LB29A), 5074 endoscopic urethrotomy procedures (OPCS-4 M76.3) and 5838 male urethral dilatation procedures (OPCS-4 M76.4/M79.2), both with a mean tariff cost of £1468 (Healthcare Resource Group code LB55A). A number of studies have made cost comparisons between the two procedures accounting for procedure costs derived from personally funded health-care provision and risk of recurrence derived from cohort studies. A UK-based cost comparison suggested that a strategy of initial urethrotomy followed by urethroplasty at first recurrence was least costly. 27 US decision-analytic models, using cost minimisation28 and cost-effectiveness29 methodology, suggest that initial urethrotomy (providing a success rate of > 40%) followed by urethroplasty on recurrence is the most efficient treatment strategy for men with bulbar urethral stricture, although utility weights were not used and parameter estimates were from non-comparative studies.

Primary objectives

• Determine the relative impact on symptoms over 24 months.

• Determine the incremental cost per quality-adjusted life-year (QALY) over 24 months.

Secondary objectives

• Determine the relative rate of need for reintervention.

• Determine the relative change in Q_max at 24 months.

• Establish the safety profile of each procedure.

• Model the incremental cost per QALY of the most effective treatment over 10 years.

• Qualitatively assess the views of men with urethral stricture and their clinicians regarding the disease, the available treatment interventions and participation in the trial.

• Determine the factors that men trade-off in deciding between the two strategies.

Sites

From 18 February 2013 to 5 March 2015 we progressively established 53 research sites (38 recruited at least one participant), all being NHS secondary care providers affiliated to the National Institute for Health Research (NIHR) Clinical Research Network in England and equivalent organisations in Scotland and Wales, which agreed to host the study locally. We initially concentrated on sites that could undertake both urethrotomy and urethroplasty. Subsequently, we set up sites that offered only urethrotomy. Those sites required delineation of routine referral pathways to a previously established trial site that offered urethroplasty (Table 1).

Additionally, four participant identification centres were opened between August and September 2014 to support recruitment to the trial.

Trial management

The central trial office was established at the Newcastle Clinical Trials Unit (NCTU), Newcastle University, Newcastle upon Tyne, UK. The NCTU was responsible for obtaining approvals, trial registration, trial management and organising the collection of outcome measures. The health economic evaluation and qualitative research teams were also based in Newcastle. The randomisation service, database construction and management, and statistical analysis were based in the Centre for Healthcare Randomised Trials (CHaRT) at the University of Aberdeen, Aberdeen, UK.

Participants

Adult men with a history of urethral stricture disease having previously undergone at least one intervention for bulbar urethral stricture were identified at the time of clinic presentation with recurrent symptoms and from health-care records at each site. Clinicians were encouraged, through the NIHR Clinical Research Network, local and national meetings, and relevant professional organisations, including the British Association of Urological Surgeons (BAUS) and the British Association of Genito-urethral Surgeons, to introduce the study to men under their care. Patients were approached and introduced to the study by clinical staff at site. If men were interested, an eligibility check was carried out by local research staff at each site according to the following criteria:

Consent procedures

Men who were eligible and in provisional agreement for participation were seen by local research staff and given trial information. Trial eligibility was checked for each potential subject using information from the prospective participant and from his clinical record. Sites kept a screening log documenting non-identifiable information and reasons for non-participation. The right to refuse to participate without giving reasons was respected.

Following a period of at least 48 hours post provision of trial information, interested and eligible patients were then contacted using their preferred means of communication (surface mail, e-mail, telephone or text) and reviewed by local research staff to explain fully the purpose and conduct of the study, including the need for randomisation to allocate them to either procedure. If they agreed to take part they gave written consent to be randomised to either open urethroplasty or endoscopic urethrotomy by signing and dating the study consent form, which was witnessed and dated by a member of the research team with documented, delegated responsibility to do so. The timing of randomisation was usually at the time of trial consent, provided that there was agreement between patient and clinician that intervention for recurrent bulbar stricture was required. Standard local arrangements concerning preoperative assessment, hospital admission, consent for surgery, conduct of surgery and after care continued unaffected by study participation. Men eligible for the study who were not willing to consider randomisation were asked to consent to being approached by a qualitative researcher, regarding participation in an interview-based study. Urologists participating in trial recruitment were also asked to give expressions of interest in participating in this qualitative study. The recruitment timetable is shown in Figure 4.

FIGURE 4.

Chart showing flow of participants through the trial. AE, adverse event; CRF, case report form; PROM, patient-reported outcome measure; USS-PROM, urethral stricture surgery – patient-reported outcome measure.

Randomisation

Progress on study

The schedule of events for trial participants is shown in Table 1. Baseline data were collected by site research staff just prior to randomisation to minimise any biases that might result from knowledge of the allocated intervention. We aimed for participants to undergo their allocated procedure as soon as possible and preferably no longer than 12 weeks after randomisation, subject to participant and clinician preference and health-care provider service constraints. At 3, 12 and 24 months post intervention, research staff at site contacted participants for a follow-up to complete case report forms (CRFs), face to face or by telephone, with supplementation by health-care record review.

Participant expenses

Expenses incurred by participants as a result of extra attendances outside standard local NHS care were reimbursed. Participants were given £25 thank-you gift vouchers to cover any unforeseen additional costs at the time of randomisation, 24 months after randomisation and on receipt of a completed end-of-trial questionnaire in December 2017.

Withdrawal

Consented participants remained on study unless they withdrew their consent or trial staff deemed that further participation by an individual was not appropriate. Participants who declined their allocated procedure after randomisation were kept on study, including those men who underwent the alternative procedure and those men who did not undergo an intervention. The reason for withdrawal was recorded if the participant agreed.

Patient and public involvement

Prior to funding application, we discussed the rationale of the trial with a patient with urethral stricture who had experienced both urethrotomy and urethroplasty. He agreed to be a co-investigator and reviewed the funding application from a patient perspective. Subsequently, he reviewed and commented on trial documents relevant to patients prior to submission for ethics approval. He contributed to discussions at trial management meetings, particularly at the start of recruitment and at the end of the study. As part of governance of the trial, we recruited a member of the public through the Northern Regional Public Involvement in Medical Research Network to act as a lay member of the Trial Steering Committee (TSC).He attended and contributed to TSC meetings throughout the trial, offering very helpful lay input into trial oversight and management. Both individuals critically appraised the final report from a lay perspective.

Primary effectiveness outcome

The primary outcome for the OPEN trial was the area under the curve (AUC) of the urinary voiding symptom score component of the USS-PROM, repeatedly measured over 24 months following randomisation. The voiding symptom measure comprised six questions each scored from 0 (symptom not present) to 4 (symptom present all of the time), giving a range of total score of 0–24. The completed questionnaires were returned to the trial office at NCTU for data entry.

Secondary effectiveness outcomes

Harms

Harms arising from trial participation, principally related to the procedure undergone, were collected through CRFs completed by research staff at site at the time of, and shortly after, the trial intervention, at 3, 12 and 24 months after intervention, and at the time of, and shortly after, reintervention. They were categorised as being expected adverse events (AEs), as listed in the trial protocol, or unexpected AEs. The trial office were notified of any AE deemed serious [i.e. a serious adverse event (SAE)] via an e-mail alert originating from the CHaRT database, triggered by data entry at site. The SAEs were then adjudicated by the chief investigator in discussion with other members of the trial team and, when required, by the site principal investigator (PI), to decide if the SAE was indeed serious and whether it was related or unrelated to participation in the OPEN trial.

Primary cost-effectiveness outcome

The primary outcome of the health economic evaluation was cost–utility, assessed as the incremental cost per QALY at 24 months after randomisation. Health-care costs were calculated from hospital visits recorded by research staff at site on CRFs completed at 3 and 24 months, and from use of primary care services as reported by participants on a patient’s cost questionnaire completed at 6 and 12 months after allocated intervention and at 18 and 24 months after randomisation. Patient costs were collected through a questionnaire completed at 6 months after allocated intervention. Responses to the EQ-5D-5L health status questionnaire were transformed using UK population tariffs39 to produce a health state utility score for each participant.

Secondary cost-effectiveness outcome

To complement the primary cost–utility analysis, the health economic evaluation included a time trade-off (TTO) experiment to better understand choices made by men with urethral stricture and assess the likely short-term disutility of the procedures that would not be measured at the EQ-5D-5L time points, and also a Markov model to project observed differences in outcome and their impact on cost-effectiveness over a 10-year time horizon. The health economic evaluation carried out as part of the OPEN trial is fully described in Chapter 5.

Summary

Data from trial CRFs were entered by research staff at site through a password-protected portal onto the internet-based data management system developed, set up, hosted and maintained by CHaRT. Participant-completed questionnaires from catheter removal post intervention onwards were collated at the central trial office and entered into the trial database by NCTU staff, whereas patient-reported outcomes at baseline and just prior to intervention were entered at site by research staff. Staff in the NCTU trial office in Newcastle worked closely with the local research teams at site and the trial team at CHaRT in Aberdeen to ensure that data were complete and accurate. We made concerted efforts to chase and complete any missing data entries through contact with the sites. Participants’ details were stored securely in the CHaRT study database under the guidelines of the Data Protection Act 1998. 40 Participants were allocated an individual specific trial number and all data, other than personal data, were identified only by this unique study number. Data collected during the course of the research were kept strictly confidential and accessed only by members of the trial team.

Trial events

The schedule of events for the OPEN trial is shown in Table 1.

General

Study design was pragmatic, in that, apart from randomised allocation of intervention and outcome data collection, standard care pathways for each procedure at individual sites were followed. These included type of anaesthesia, use and regimen of antibiotic prophylaxis, surgical instrumentation, closure or non-closure of oral mucosal graft donor site, duration of postoperative urethral catheterisation and clinical follow-up schedule, including use of investigations to detect stricture recurrence. The interventions were funded by the NHS in accordance with local contracting mechanisms. The NHS excess treatment costs were approved by the sponsor (Newcastle upon Tyne Hospitals NHS Foundation Trust, Freeman Hospital, Newcastle upon Tyne, UK) and the host NHS organisation at each participating site.

Urethroplasty (experimental)

For urethroplasty, men were positioned supine on the operating table with hips and knees held in an abducted and flexed position by suitable leg supports. A longitudinal skin incision was made in the perineum above the anus and towards the base of the scrotum. The bulbar urethra was localised and mobilised from its attachments. The stricture segment was incised longitudinally on the dorsal or ventral surface according to surgeon preference. The surgeon then decided if the stricture could be excised with or without transection of the corpus spongiosum and a primary anastomosis made (typically, proximal strictures < 2 cm in length with limited fibrosis) or if an oral mucosal graft should be placed without stricture excision in an augmented repair (typically, fibrotic strictures ≥ 2 cm in length). The oral mucosal graft was harvested from the inner cheek, defatted and sutured to the cut urethral edges, then stabilised against the corpora cavernosa if dorsally placed or the corpus spongiosum if ventrally placed. A 16-French silicone Foley urethral catheter was then placed and left in situ for free drainage of urine for a postoperative period of approximately 2 weeks. Once recovered, the patient was discharged home to return at a planned later date for a urethrogram to check that there was no leakage from the area of repair, catheter removal and trial of voiding. The median stay from NHS England hospital activity data at the time of the trial was 2 days. 8 Clinical follow-up was usually by wound and symptom review at 3 months and urinary flow rate measured at 3 and 24 months post intervention.

Endoscopic urethrotomy (control)

For urethrotomy, men were positioned supine on the operating table with legs supported in an abducted and flexed position. The endoscope (Sachse urethrotome) was passed retrogradely through the urethral lumen under direct vision until the distal end of the stricture segment was encountered. A fine-calibre guidewire was placed through the stricture into the bladder to aid incision planning. The stricture was progressively incised longitudinally under vision using the steel blade (‘cold knife’) mounted on the endoscope until healthy mucosa signalling the proximal end of the stricture was reached. Alternatively, a diathermy needle mounted on the endoscope or a laser fibre passed through the endoscope was used to make the incision (‘hot knife’). For short flimsy strictures, dilatation with the endoscope or graduated dilators could be used rather than formal urethrotomy. A 16-French silicone Foley catheter was then placed for free drainage of urine during the postoperative period. It was typically removed at 24–48 hours postoperatively, with a trial of voiding prior to discharge or following hospital reattendance after discharge. Follow-up was by outpatient review and urinary flow measurement 3 and 24 months postoperatively. According to patient and clinician decision, a standardised programme of intermittent self-dilatation could be initiated 1 week after catheter removal, as this can delay time to recurrence. 18

Delivery of interventions

All procedures were carried out by accredited consultant urologists or senior trainees in urology. Competency in performing endoscopic urethrotomy is a mandatory component of training as urologists and is regularly performed by approximately 95% of practising urologists. 24 Urethroplasty is a specialised technique requiring extra training. Specialist surgeons with recognised expertise working in specific UK centres were identified through BAUS and acted as PIs for the trial at these sites. The precise technique of urethroplasty used for each participant was decided by the operating surgeon. Details of each intervention were recorded by research staff at site on a CRF.

Sample size calculation

We aimed to detect at least a 0.1 (10%) difference in the AUC of the voiding symptom score calculated from the plotting of repeated measurements over 24 months for both trial groups and using a 0–1 utility scale. This conservatively assumed a standard deviation (SD) score of < 0.33 based on the finding of a SD of 0.15 in a previous shorter-term study among men undergoing bulbar urethroplasty. 30 For 90% power this would require 500 men to be randomised. Following the trial feasibility phase, it was clear that this sample size was not achievable in a fundable time frame. We therefore recalculated the SD of the symptom score in August 2014, using trial data collected from the first 69 men randomised who had completed at least one score following intervention (220 measurements in total), while maintaining blinding of allocated intervention. This recalculation gave a SD of 0.165, which reduced to 0.15 when adjusted for baseline score and trial site. Using these data we updated our sample size estimate assuming a reduced SD of ≤ 0.21. This indicated a requirement of 170 men to be randomised to have 90% power for detection of a 10% difference at a two-sided 5% level. We inflated the figure to 210 to allow an up to 19% loss to follow-up rate. When interpreting actual trial data we found results more straightforward to consider by scaling the AUC to the USS-PROM scale minimum and maximum over 24 months, in which a score of 0 indicated a complete lack of symptoms over the trial and a score of 24 indicated full symptoms throughout the trial. Rescaling the AUC did not change the assumptions of the sample calculation.

Statistical analysis

Qualitative substudy

Qualitative work was undertaken to establish factors determining willingness of patients and support of their clinicians to consider participation. Timely and successful completion of the planned qualitative study first established that the aims of the trial were important to men eligible to participate, given the troublesome and chronic nature of their symptoms, and to both general and specialist clinicians; these findings reinforced the rationale and need for the trial. As part of this work, we found that men eligible for inclusion were most likely to be willing to participate when their symptoms had first recurred and this was the point at which they expressed most uncertainty as to which option would be best for them as individuals. Both general and specialist clinicians were also very supportive of the aims of the trial given the uncertainty of guidance on best treatment, but expressed concerns regarding delivery of balanced information to men eligible for participation. To assist men eligible for participation in making a decision about participation, appropriate written guidance and an example video were provided, supported by personal contact from the trial team.

Trial progress and monitoring

The study initially set out to progressively build to a target of 500 participants over 24 months. This included an initial 12-month feasibility study, during which recruitment and patient adherence to the intervention were evaluated. Feasibility of recruitment was analysed after 9 months of active recruitment (trial month 12) and reported in August 2014 to the TSC and the funder, with an additional safety report reviewed by the Data Monitoring Committee (DMC). Recruitment continued to be monitored by the Trial Management Group (TMG) through returns to the randomisation website. The funder and TSC requested a recovery plan to mitigate the slow recruitment and ensure that the trial was completed over a fundable period of time. The recovery plan, which principally involved a re-estimation of required sample size using early observed data to 210 randomised participants, opening of additional sites and the recruitment window increasing from 24 months to 35 months, was submitted and approved by the TSC and funder in December 2014.

Sources of bias

To allow randomisation, both the eligible participant and the responsible clinician needed to be sufficiently uncertain whether the experimental or control strategy was best for management of the individual’s recurrent urethral stricture. Given the lack of high-level evidence as to which was the more effective intervention, trial information was provided illustrating the uncertainty and the need for a definitive trial. This aimed to ensure that any selection bias in terms of differing characteristics of men with recurrent bulbar urethral stricture willing to be randomised compared with those men who were eligible but not willing to participate was minimised. As far as possible and within the limits of data protection legislation, we recorded reasons for declining randomisation, but patients were free to decline participation and randomisation without giving a reason.

Trial literature for men eligible for participation or who were participating in the trial included the following: the OPEN trial participant information sheet and consent form v1.5 (main trial), the OPEN trial participant information sheet and consent form interview study v1.3, the OPEN trial website synopsis v1.1, the OPEN trial patient end of study questionnaire letter v1.0, the OPEN trial patient questionnaire letter v1.2, the OPEN trial patient invitation letter main study v1.0, the OPEN trial contact card v1.0 and participant flowchart v1.0. These documents are available at www.journalslibrary.nihr.ac.uk/programmes/hta/105723#/ (accessed 26 June 2019).

Primary analysis

In the primary complete-case ITT analysis we included 159 (72%) of the randomised participants using the criteria of having completed primary outcome questionnaires on a minimum of three occasions: at baseline, during the first year after intervention and at 18–24 months after randomisation. This was 94% of the pre-set target of 170 participants providing data on the primary outcome in our revised sample size calculation. The total of 159 men comprised 69 men (63% of those men randomised) allocated to the urethroplasty group and 90 men (81% of those men randomised) allocated to urethrotomy. A total of 25 men allocated to urethroplasty had urethrotomy and a further 15 men had no intervention. From the urethrotomy group, 11 men had urethroplasty and 8 men did not have any intervention. Accordingly, 19 men (17%) allocated to urethrotomy did not receive urethrotomy and 40 men (37%) allocated to urethroplasty did not receive urethroplasty. All participants who ‘crossed over’ or did not have any intervention but who remained in active follow-up and provided sufficient completed trial questionnaires were analysed according to allocated group (complete-case ITT analysis). Reasons for exclusion from the primary analysis are detailed in Table 6.

Of 222 randomised participants, 109 were allocated to the urethroplasty group and 113 allocated to the urethrotomy group. Two randomised participants, one from each group, were excluded soon after randomisation because further assessment showed them to be ineligible and they were classified as post-randomisation withdrawals, excluding them from the primary analysis. One man (allocated to urethroplasty) was unfit for prolonged anaesthesia and one (allocated to urethrotomy) had a stricture that, following urethrography, was deemed unsuitable for endoscopic management. Accordingly, at baseline, there were 108 participants in the urethroplasty group and 112 in the urethrotomy group. Baseline characteristics were well balanced between the two groups (Tables 7 and 8). Twenty-two participants allocated to urethroplasty underwent urethrotomy and 15 participants did not have either procedure. In the urethrotomy group, 11 men underwent urethroplasty and eight men no procedure. These protocol deviations and their reasons are given in Tables 9 and 10. The median time between randomisations and interventions was 92 days, compared with 47.5 days for urethrotomy (Table 11). The rate of trial questionnaire return by each group for each time point is given in Table 12. The attrition rate for inclusion in the primary analysis was 30 of 109 (27%) men in the urethroplasty group, 23 of 113 (20%) men in the urethrotomy group and 53 of 222 (24%) men in the randomised population as a whole.

Outcomes

Primary

Area under the curve for USS-PROM total voiding score

Figure 7 shows the mean total USS-PROM voiding score for each randomised group by time point. The mean AUC of multiple (at least three) voiding score measurements on a scale from 0 (no symptoms) to 24 (worst symptoms) over the 24 months after randomisation was 7.4 (SD 3.8) in the urethroplasty group and 7.8 (SD 4.2) in the urethrotomy group (a mean difference of –0.36, 95% CI –1.74 to 1.02; p = 0.60 in favour of the urethroplasty group). To assess whether or not the primary outcome was biased by the exclusion of participants with fewer than the required three measures, we undertook as sensitivity analyses an ITT calculation using the predefined methods of data imputation to replace missing values and a per-protocol analysis using data only from participants eligible for the primary analysis who underwent the intervention to which they were randomly allocated. The results of these three analyses are presented in Table 13. The estimate of the primary outcome was robust to sensitivity analyses using pattern-mixture models for missing data for all but unrealistic scenarios (e.g. missing observations in one arm of the trial staying at baseline levels for all 24 months, whereas in the other arm all missing observations were consistent with observed scores).

FIGURE 7.

Mean (standard error of the mean) for total voiding score at each nominal time point. AI, after intervention; AR, after randomisation.

TABLE 13 - Sensitivity analyses for primary result

Analysis	AUC (primary outcome) in each sensitivity analysis
	Urethroplasty (N = 108), mean (SD); n	Urethrotomy (N = 112), mean (SD); n	Effect size (95% CI); p-value
Complete-case ITT analysis (primary analysis)	7.4 (3.8); 69	7.8 (4.2); 90	–0.36 (–1.74 to 1.02); 0.60
Sensitivity analysis with multiple imputation (ITT)	7.7 (3.9); 93	7.9 (4.2); 104	–0.33 (–1.74 to 1.09); 0.64
Per protocol (underwent allocated intervention)	6.8 (2.9); 54	7.8 (4.2); 80	–1.02 (–2.12 to 0.07); 0.07

Subgroup analyses

We identified number of previous interventions, length of stricture, time since last intervention prior to randomisation and age as risk factors with potential to increase the risk of failure to control symptoms. Figure 8 presents subgroup analyses for the AUC of the voiding score. Table 14 presents the interaction term for each subgroup. The results show no evidence that the treatment effect of the different subgroups varies significantly from the main effect.

FIGURE 8.

Subgroup analyses for the USS-PROM voiding score AUC.

TABLE 14 - Interaction terms for the subgroups

Interaction term	Effect size (95% CI); p-value
Overall	–0.36 (–1.74 to 1.02); 0.60
Time since last intervention	–0.36 (–3.19 to 2.46); 0.79
Length of stricture	2.54 (–1.42 to 6.50); 0.20
Number of previous interventions	0.56 (–2.89 to 4.02); 0.74
Age	–0.25 (–2.58 to 2.08); 0.83

Secondary outcomes

Area under the curve for impact of urinary symptoms on daily activities: assessing condition-specific quality of life

Sufficient data were available to calculate this outcome for 69 participants in the urethroplasty group and 90 participants in the urethrotomy group. The urethroplasty group had a mean score at 24 months of 1.1 (SD 0.8), compared with 1.0 (SD 0.7) in the urethrotomy group. The mean difference between treatments adjusting for minimisation variables was 0.06 (95% CI –0.19 to 0.30).

Area under the curve for satisfaction with sexual function

Sufficient data were available to calculate this outcome for 63 participants in the urethroplasty group and 87 participants in the urethrotomy group. The urethroplasty group had a mean AUC of 2.9 (SD 1.2), compared with a mean of 2.5 (SD 1.2) in the urethrotomy group. The mean difference between treatments adjusting for minimisation variables was 0.35 (95% CI –0.06 to 0.75).

Reintervention

In total, 44 participants had at least one reintervention and there were 52 reinterventions overall (Table 15). During the trial observation period for each participant (up to 4 years), 15 (16%) men in the urethroplasty group required a reintervention at a median of 474 (IQR 399–577) days after initial surgery, compared with 29 (28%) men at a median of 308 (IQR 211–448) days for men allocated to the urethrotomy group. The hazard ratio for time until first reintervention was 0.52 (95% CI 0.31 to 0.89; p = 0.02) (Figure 9), representing a 48% lower risk of reintervention with urethroplasty. Calculation including multiple events gave a similar hazard ratio of 0.49 (95% CI 0.30 to 0.82). A secondary analysis involving only men who underwent the allocated intervention (per protocol) showed a hazard ratio for time to reintervention of 0.28 (95% CI 0.15 to 0.55; p < 0.001) (see Figure 9). Subgroup analyses showed no evidence of heterogeneity around the ITT treatment effect (Figure 10). For men undergoing urethrotomy, commencement of intermittent self-dilatation did not lengthen the time to recurrence. Median time until first reintervention was 384 (IQR 214–555) days for 21 participants who did not commence dilatation and 308 (IQR 171–399) days for seven participants who did.

TABLE 15 - Number of reinterventions according to randomised group

	Intervention, n (%)
	Urethroplasty	Urethrotomy
Total participants randomised	108	112
Participants with data available	93	104
Number of reinterventions received
0	78 (83.9)	75 (72.1)
1	13 (14.0)	26 (25.0)
2	1 (1.1)	2 (1.9)
3	1 (1.1)	0 (0)
4	0 (0)	1 (1.0)

FIGURE 9.

Hazard curves for reintervention by randomised or treatment received group up to 4 years after initial intervention. (a) Analysis of participants who had surgery according to their randomised allocation (modified ITT); and (b) restricted to men who underwent procedure allocated at randomisation (per protocol). Reprinted from European Urology, vol. 78, Goulao B, Carnell S, Shen S, MacLennan G, Norrie J, Cook J, et al. , Surgical treatment for recurrent bulbar urethral stricture: a randomised open-label superiority trial of open urethroplasty versus endoscopic urethrotomy (the OPEN trial), pp. 572–80,45 Copyright 2020, with permission from Elsevier.

Reprinted from European Urology, vol. 78, Goulao B, Carnell S, Shen S, MacLennan G, Norrie J, Cook J, et al. , Surgical treatment for recurrent bulbar urethral stricture: a randomised open-label superiority trial of open urethroplasty versus endoscopic urethrotomy (the OPEN trial), pp. 572–80,45 Copyright 2020, with permission from Elsevier.

FIGURE 10.

Subgroup analyses for time to reintervention. a, Upper limit of the CI truncated to fit the graph (95% CI 0.19 to 28.4).

Improvement in maximum flow rate

A successful outcome in terms of Q_max was defined as an increase of ≥ 10 ml/second from baseline at 3 months, and between 12 and 24 months. Data allowing calculation of this outcome were available from 64 men in the urethroplasty group and from 68 men in the urethrotomy group at 3 months. Participants in the urethroplasty group had twice the odds of experiencing an improvement ≥ 10ml/second in their maximum flow rate at 3 months compared with participants in the urethrotomy arm. Data were available from 44 participants in the urethroplasty group and from 63 participants in the urethrotomy group at 12 or 24 months. At these time points, participants in the urethroplasty arm had 2.64 times greater odds of experiencing an improvement of ≥ 10ml/second in their maximum flow rate (Table 16).

TABLE 16 - Maximum flow rate (≥ 10 ml/second) increase from baseline

Odds ratio adjusted for minimisation variables.

Improvement defined as an increase in the flow rate of ≥ 10 ml/second.

Maximum flow rate	Urethroplasty (N = 93)	Number of participants providing data	Urethrotomy (N = 104)	Number of participants providing data	Odds ratio (95% CI); p-valuea (urethroplasty vs. urethrotomy)
Flow rate baseline, mean (SD)	13.2 (15.5)	75	10.5 (6.7)	89
Flow rate at 3 months, mean (SD)	22.3 (11.4)	64	17.6 (11.1)	68
Flow rate at 12 months, mean (SD)	22.3 (12.2)	41	16.8 (11.5)	55
Flow rate at 24 months, mean (SD)	21.7 (13.7)	29	19.1 (11.8)	49
Improved at 3 months from baseline,b n (%)					2.08 (1.05 to 4.12); 0.035
Not improved	28 (30.1)		42 (40.4)
Improved	29 (31.2)		21 (20.2)
Missing	36 (38.7)		41 (39.4)
Improved at 12 or 24 months from baseline,b n (%)					2.64 (1.14 to 6.15); 0.024
Not improved	26 (28.0)		50 (48.1)
Improved	18 (19.4)		13 (12.5)
Missing	49 (52.7)		41 (39.4)

Recurrence

Stricture recurrence was observed in 19 (20%) participants in the urethroplasty group and 39 (38%) participants in the urethrotomy group. The difference was significant in favour of urethroplasty, with a mean odds ratio of 0.43 (95% CI 0.24 to 0.76) (Table 17).

TABLE 17 - Recurrence (for those participants who had surgery)

	Intervention, n (%)		Odds ratio (95% CI); p-value (urethroplasty vs. urethrotomy)
	Urethroplasty (N = 93)	Urethrotomy (N = 104)
Any recurrence	19 (20.4)	39 (37.5)	0.43 (0.24 to 0.76); < 0.001
Reintervention	15 (16.1)	29 (27.9)
Stricture recurrence at 12 months	4 (4.3)	14 (13.5)
Stricture recurrence at 24 months	6 (6.5)	8 (7.7)
Deterioration in flow rate	5 (5.4)	7 (6.7)
Deterioration in voiding score	1 (1.1)	1 (1.0)

Urinary stream picture score

Median stream score was lower at 24 months post randomisation than at baseline in both randomised groups. A total of 24 out of 59 (41%) participants reported an improvement of at least two grades in their urinary stream score in the urethroplasty group compared with 18 out of 59 (31%) participants in the urethrotomy group. There was a significant association between self-reported urinary stream improvement at 24 months and improved Q_max at 24 months (n = 66; p = 0.02).

Health status: EuroQol-5 Dimensions

Participants’ rating of their state of health at 24 months on the EuroQol-5 Dimensions (EQ-5D) VAS from 0 (worst health imaginable) to 100 (best health imaginable) was a mean of 81.9 (SD 19.5) in the urethroplasty group and a mean of 79.4 (SD 17.5) in the urethrotomy group.

Absolute changes in outcomes between baseline and 24 months

Changes in the absolute values of participant-reported outcomes between baseline and 24 months were similar between the urethroplasty and urethrotomy groups (Table 18).

TABLE 18 - Changes in participant-reported outcomes measured on the trial questionnaire at baseline and 24 months

P25, 25th percentile; P75, 75th percentile.

USS-PROM domain	Intervention
	Urethroplasty (N = 108)				Urethrotomy (N = 112)
	Baseline (USS-PROM score)	n	24 months (USS-PROM score)	n	Baseline (USS-PROM score)	n	24 months (USS-PROM score)	n
Total voiding score (0 = no symptoms to 24 = symptoms all the time), mean (SD)	13.4 (4.5)	104	6.0 (5.5)	58	13.2 (4.7)	109	6.4 (5.3)	59
Delay before starting to urinate (0–4), median (P25–P75)	2.0 (1.0–3.0)	105	0.0 (0.0–1.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–2.0)	59
Poor strength of urinary stream (0–4), median (P25–P75)	3.0 (3.0–4.0)	105	1.0 (0.0–2.0)	58	3.0 (3.0–4.0)	109	1.0 (0.0–2.0)	59
Having to strain before urinating (0–4), median (P25–P75)	2.0 (1.0–3.0)	105	0.0 (0.0–1.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–1.0)	59
Intermittent urinary stream (0–4), median (P25–P75)	2.0 (1.0–3.0)	104	1.0 (0.0–1.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–1.0)	59
Feeling of incomplete bladder emptying (0–4), median (P25–P75)	2.0 (1.0–3.0)	105	1.0 (0.0–2.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–2.0)	59
Post-micturition dribbling (0–4), median (P25–P75)	2.0 (1.0–3.0)	105	1.0 (0.0–2.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–2.0)	59
Impact of urinary symptoms (0 = no impact to 3 = a lot of impact), median (P25–P75)	2.0 (1.0–3.0)	107	0.0 (0.0–1.0)	58	2.0 (1.0–3.0)	110	1.0 (0.0–1.0)	59
Satisfaction with sexual function (0 = very satisfied to 4 = very dissatisfied), median (P25–P75)	3.0 (2.0–4.0)	97	1.0 (0.0–3.0)	53	3.0 (2.0–4.0)	100	1.0 (0.0–2.0)	56
Urine stream strength picture (1 = weak to 5 = strong), median (P25–P75)	3.0 (3.0–4.0)	103	2.0 (1.0–3.0)	58	3.0 (3.0–4.0)	108	2.0 (2.0–3.0)	58
Overall health state (0 = worst health imaginable to 100 = best health imaginable), mean (SD)	72 (20)	104	81.9 (19.5)	58	77 (17)	105	79.4 (17.5)	59

Adverse events

At least one AE was reported for 89 of the trial participants (81 participants had one AE, seven participants had two AEs and one participant had three AEs). The reported AEs and SAEs are tabulated according to allocated group and treatment received (Tables 19–22). Postoperative complications categorised by the Clavien–Dindo grade are also tabulated by allocated group and intervention received (see Table 21).

Mouth pain was the most common perioperative AE and was more frequent in the urethroplasty group, according to both randomised allocation and treatment received.

Serious adverse events

During the trial 22 participants were reported to have experienced at least one SAE (17 participants suffered one SAE, three participants suffered two SAEs and two participants suffered three SAEs).

Interviews with men eligible for participation

Interviews with trial-eligible men were designed to explore their priorities in seeking surgical intervention and deciding whether or not to participate in the OPEN trial (interview schedules can be found in Appendices 2 and 3). To be eligible for the OPEN trial, and therefore for the qualitative study, men had to have a diagnosed urethral stricture and be able to undergo both endoscopic management by urethrotomy and reconstructive surgery by urethroplasty. Participating sites were encouraged to ask men if they would be willing to participate in an interview, regardless of their decision to participate in the trial. Men were asked to sign a consent form, separate from trial participation, and the site research nurse then passed the men’s contact details onto the qualitative team. All men who were willing to be interviewed were then contacted by the researcher (PW) and offered an interview either face to face or by telephone. For face-to-face interviews the researcher either travelled to the participant’s home or preferred location or offered them travel expenses to come to where the qualitative team was based. It was anticipated that interviews with men with bulbar urethral stricture may raise potentially sensitive issues, such as erectile function and urination problems. The researcher (who was male) was provided with guidance and support in how to prepare for, and manage, potentially sensitive topic areas.

Interviews with clinicians

Interviews with urologists were designed to explore clinicians’ opinions regarding the management of urethral stricture, their sense of equipoise between the trial groups and the acceptability of approaching men about being randomised (interview schedule can be found in Appendix 4). In discussion with the chief investigator and trial team, it was decided that qualitative interviews would be conducted with a broad range of urological practitioners. Criterion sampling was used to include both specialist and general urologists, including those urologists who were not recruiting to the OPEN trial. General urologists are those who perform urethrotomy but would have to refer men elsewhere to receive an urethroplasty procedure. Specialist urologists are those who can perform both procedures. To be eligible for an interview, urologists needed to be currently practising and routinely treating men with urethral stricture, but were not necessarily contributing to the OPEN trial recruitment. The sampling and recruitment of clinicians for interviewing involved a mixture of snowballing and contacting practitioners from a BAUS database. Surgeons were contacted by the researcher either by telephone or by e-mail. Potential interviewees were sent a consent form and information about qualitative substudy. All those clinicians approached were reassured of their anonymity and that their interview data would be managed independently of the OPEN trial team. When a clinician was willing to be interviewed, the researcher arranged for the consent form to be signed and the interview to be conducted either face to face or by telephone. All interviews were audio-recorded, except for one that the clinician requested not be. In this case, a full description of the interview was written by the researcher and discussed with the qualitative research team before informing the analysis.

Data management and analysis

Qualitative analysis is an interpretive process that benefits from interdisciplinary working. Throughout the qualitative substudy, results and findings were discussed with the CI and presented to PIs at the OPEN trial investigators meeting in Newcastle upon Tyne (August 2013). All interviews were audio-recorded, transcribed verbatim, checked and edited to ensure respondents’ anonymity. Analysis was conducted in accordance with the standard procedures of rigorous qualitative analysis (Rapley60) and aimed to identify, explore and refine emergent patterns. Procedures from first-generation grounded theory – coding, constant comparison, memoing61 – were used, alongside tables,62 diagrams and mapping. 63 Proprietary software (NVivo version 10; QSR International, Warrington, UK) was used to support the management and retrieval of data. 64 Data collection and analysis occurred concurrently, so that issues raised in earlier phases of fieldwork were explored in subsequent ones. Data collection and analysis were also supported by discussion in trial meetings and regular qualitative data clinic sessions, which included health professionals and social scientists from different clinical and academic backgrounds.

Interviews with men

The interviews with men revealed a number of common themes in the experience of living with symptoms of a stricture and seeking medical treatment. The men’s accounts revealed some of the practical issues that they faced on a daily basis. Understanding their experiences provides an insight into their help-seeking and priorities in treatment decision-making.

Clinician interviews

To place these interviews in context, it is important to note that there are two types of clinicians participating in the trial: specialist urologists and general urologists. Specialist urological surgeons are able to offer both of the OPEN trial treatment options as they are trained to carry out urethroplasty, whereas general urologists are able to deliver only urethrotomy and have to refer patients to a specialist if they feel urethroplasty is required. Regardless of their participation in the trial, the clinicians who were interviewed were generally supportive of OPEN trial, with one exception. The exception, a general urologist who asked not to be audio-recorded, said that they felt that the OPEN trial was a promotion of urethroplasty and therefore not to the benefit of general urologists. All other clinicians supported the OPEN trial and felt that there existed a genuine uncertainty in the management of urethral stricture patients:

We know that urethroplasty lasts for quite a while but we don’t know really how they compare in the same patients or group, because classically they’ve been different lots of patients.

Specialist urologist 3

A less experienced specialist had decided not to participate in the OPEN trial because it would mean reducing the opportunity to practise urethroplasty and potentially having their ‘learning curve’ contained within the study results:

The last thing I want, from a personal perspective, is to see a great little stricture and think ‘perfect, that will do nicely’ only to have him randomised to a [urethrotomy] – do you see what I mean? . . . Obviously we all support research but I don’t think [we are a] busy centre and do I want my result – my learning curve results in a trial?

Specialist urologist 3

Despite general support for the study, when asked about their expectations or experience of recruiting to the OPEN trial, most clinicians felt that it already was, or would be, difficult. Some of the general urologists felt concerned that they would see only a small number of eligible patients at their clinic. However, the most prominent concern about recruitment was dealing with men’s preferences.

The clinical organisation of referrals

Interview data from both clinicians and men were used to map the standard care pathways for bulbar stricture patients in the UK. Figure 11 is a simplistic representation of a standard patient pathway to treatment and the two potential points of recruitment. Typically, a man will present his symptoms to a general practitioner (GP) before being referred to a local urology clinic. Most men will be initially treated at a general urology clinic and receive their first urethrotomy. If the man’s stricture symptoms return, he tends to return to the same clinic, either directly or through his GP. At this point, the man is now eligible to receive either a repeat urethrotomy or a reconstructive urethroplasty, making him potentially eligible for inclusion in the OPEN trial. In standard practice, the urologist can either recommend another repeat urethrotomy or refer the man to a specialist to discuss reconstructive options. There is no precedent for how many repeated urethrotomies a man should have had prior to referral for urethroplasty; indeed, this is part of the uncertainty about how to manage the condition. If the man is interested in urethroplasty, or the clinician feels that the patient’s case is severe enough, the clinician may recommend referral to a specialist centre to discuss urethroplasty.

FIGURE 11.

Different points of recruitment to the OPEN trial.

In practice, this means that younger patients and more complex cases tend to be referred to specialists more readily than older patients or those with relatively simple strictures:

You know the 18-, 20-, 30-year-olds, it just seems to me that consultants around the region acknowledge the fact that they’re facing a lifetime of self-catheterisation and they tend to get referred younger, whereas the people in the seventies or what have you, may well get pushed towards intermittent self-dilatation and referred to me only when they’ve had complications.

Specialist urologist 5

As outlined above, specialist urologists felt reluctant to randomise patients whose symptoms were severe, whereas general urologists felt that it was difficult to refer patients whose symptoms were manageable with a repeat urethrotomy. These differing perspectives correspond with the types of stricture patient they typically treat. General urologists see patients with moderate symptoms, whereas specialists see more complex and severe cases. In this way, the difficulty recruiters have in achieving clinical equipoise is not adequately explained as individual bias, but is partly a result of the relative positions of general and specialist urologists in the organisation and division of urethral stricture care.

The differences in clinicians’ expectations of patient preferences reflect their relative position within the overall organisation of care for urethral stricture. Intersite referrals were a key factor and underpinned the expectations of men’s willingness to be randomised. In particular, there was concern that men who were approached post referral would be expecting to discuss urethroplasty and therefore less willing to be randomised. Echoing some of the interviews with men, a specialist urologist said that speaking to men about randomisation would be difficult where the man was already anticipating discussion of urethroplasty, as ‘a lot of the men are already primed’ (specialist urologist 3). Specialists were concerned that men who had already been told by the referring clinician that they would be having urethroplasty would make recruitment harder:

The referring urologist says, ‘Oh, what you need is an urethroplasty, I’ll send you to [consultant’s name]’. And then I try and talk them back into randomisation! It’ll be more difficult for us. If they come to me [without preconceptions] it’ll be easier for me . . . what you need really [is] the guy who’s not quite sure what to do.

Specialist urologist 3

A few of the specialists said that they had spoken to or were planning to speak with referring centres and to try and involve them in the recruitment process and in keeping patients open to both procedures. One of the clinicians outlined that this was more feasible when there was already a good network of communication between urological sites. Related to this, a few urologists expressed concern about having to refer patients to consultants whom they did not know personally. One remarked that, had trial recruitment involved referring a patient to a clinician they did not know, they would not have agreed to become a recruiting centre.

When asked in the interview about the referral process, many of the specialist urologists said that men were not referred frequently or early enough. Some complained that the men who were referred to them from general clinics were severe or complex cases and felt that the men may have benefited from an urethroplasty sooner. On the other hand, some general urologists reported that the patients often wanted to be treated simply and locally if possible:

You know you’re talking about a 2.5-hour journey each way, you know. Plus, if it’s a consultation and surgery, that’s two stages. Some men will go for it and some won’t.

General urologist 2

Other general urologists admitted that they may try and ‘hold on to’ a patient if they could reasonably be treated locally. They were concerned that early referral, and therefore trial participation, might not be in the interest of their patients who wanted local care. General urologists also felt that referrals could potentially undermine recruitment. This is because patients would have to agree to potentially travel between sites before they could agree to be randomised. This would be particularly burdensome for patients when the nearest specialist was some distance away:

You tell [men] that in this arm of the trial, I mean this part, will be done locally but if you, for the reconstruction you might have to go to another hospital, and then that might influence. They might say ‘OK I don’t want to take part in the trial because I’d rather stay local’.

General urologist 4

Relatedly, another issue raised by some was the need for men to take time off work and inform friends and family. As discussed earlier, many men with urethral stricture prefer to conceal and self-manage the condition. Several of the clinicians anticipated that it might not be easy for patients to agree to potentially take time off work and add to the recovery time, as this would usually require some personal preparation.

Shaping trial recruitment and design

While the qualitative substudy was being conducted during the first 10 months of recruitment, the OPEN trial was not recruiting to target. The findings of the qualitative work suggested specific barriers to recruitment. Clearly, some men have strong preferences, and when recruitment discussion took place with them they declined randomisation. However, some men who took part in recruitment discussions were at a specific point in the trajectory of their experience of symptom recurrence, severity and help-seeking, at which point they were willing to accept randomisation. Given the differing expectations and expressions of reluctance of approaching some men, which both general and specialist urologists raised, some men potentially eligible for the trial may not have had the opportunity to discuss potential trial participation at all.

Following the report of the substudy, the qualitative team collaborated with the chief investigator and Trial Management Group to develop implementation strategies to increase both the number and location of recruitment discussions. It was agreed that the findings would be used to develop an overall approach, as well as relevant tools to help with the initiation of new sites, the training of recruiters and the maintenance of good, consistent recruitment practices. We mapped the key problems we had identified against Powell et al. ’s74 compilation of strategies for implementing clinical innovations in health and mental health care (Table 25). This helped to inform the detail, direction and content of the strategies we suggested to shape the trial recruitment and design going forward. It was important that all discussions and support tools used with clinicians reflect the types of concerns that men were likely to have, as well as the discomfort experienced by specialists and general clinicians in approaching different types of men.

Enabling more recruitment discussions

Given that recruiting clinicians can be selective about the men to whom they introduce the trial, it was felt that new and existing recruiters should be provided with support that emphasises the need to offer trial participation to all potentially eligible men. We recommended that, within the OPEN trial, support should include examples of men with different treatment preferences. This should also be sensitive to the different perspectives and biases of specialist and general urologists. Thus, support for general urologists should emphasise how it can be appropriate to refer men to a specialist, even if they are older and their symptoms are relatively minor. Conversely, support for specialist urologists should emphasise how repeat urethrotomy can also be appropriate for younger men or those men with severe or frequently recurring urethral stricture.

In part, the focus needed to be on getting clinicians to explicitly reflect on, and so challenge, their expectations and feeling of discomfort about approaching some potentially eligible men. For example, to encourage recruiters to approach more potentially eligible men, the findings were used to create several illustrative vignettes of urethral stricture patients (see Appendix 5). These reflected the types of men who were seen, their concerns, previous treatment history and severity of symptoms. Early versions of the vignettes were designed to look like GP referral letters; however, feedback from clinicians suggested that these were too lengthy and not like typical referrals. Incorporating this feedback, these were revised to be short case studies with explanations of their learning outcome for recruiting clinicians. A video was also made to introduce and promote the OPEN trial. It outlined the trial and then depicted two clinical scenarios, which simulated recruitment discussions around the available treatment options and study participation. The video was publicly available and was designed to target new and potential recruiting clinicians, as well as to be a resource for men who had been informed about the study and were considering trial participation.

Enabling different recruitment discussions

Another focus was on expanding the points at which eligible men could be initially informed about and recruited to the OPEN trial. As noted above, at the time of conducting the qualitative substudy, the majority of recruitment discussion was happening at specialist centres and was undertaken by specialist clinicians. However, findings strongly suggested that recruitment discussions at specialist centres could be too late in the treatment pathway for some men, given that they could be entering those discussions already primed with an expectation that the focus of the consultation would be on one option, namely urethroplasty. However, recruitment discussions with men eligible to participate in the trial could potentially occur at two distinct points in the standard treatment pathway, at both general centres and specialist centres.

One strategy that was suggested was to focus on changing the emphasis of discussions that men have at general sites, prior to referral to specialist sites. When possible, specialist sites and the OPEN trial team could directly engage with their colleagues at general sites to reframe elements of the discussions general clinicians have with men prior to referral. Rather than suggest that the focus of the referral would be to only discuss and then receive urethroplasty, they could outline that both options would be considered. A related strategy was also suggested, to approach men earlier in the pathway by increasing the number of general urology centres participating in the OPEN trial. This could then improve access to eligible men who are yet to develop a strong preference for either procedure. However, a disadvantage of trying to approach men at general clinics is that they would see far fewer men potentially eligible for the OPEN trial than specialist sites. This meant that more general sites would be needed to recruit the same number of participants as specialist sites, albeit with the potential for a better conversion of eligible patients accepting randomisation.

Analysis of resource use and cost

Total cost

Taking the perspective of the NHS and patients and their families, total costs combining NHS resource use costs (intervention, reintervention and health service use during follow-up) and patients’ out-of-pocket costs are presented in Table 30.

As Table 30 shows, the cost of urethroplasty was higher over the 24 months post randomisation than that of urethrotomy. However, the magnitude of the difference in cost reduced when reintervention and follow-up costs were included.

Quality-adjusted life-years

Effectiveness was measured in terms of utility derived from responses to the EQ-5D-5L collected at baseline, immediately prior to surgery, 1 week after catheter removal, at 3, 6, 9, 12 and 24 months following surgery, at 18 and 24 months after randomisation and at the end of study. Additional EQ-5D-5L questionnaires were intended to be completed at times of participants requiring reintervention; however, very few patients submitted both pre- and post-reintervention EQ-5D-5L data. Among those participants who had both pre- and post-reintervention EQ-5D-5L data, no difference was found between the allocated intervention groups or between pre and post reintervention; therefore, these EQ-5D-5L data were not used in the calculation of total QALYs. The responses to the EQ-5D-5L questionnaire were transformed using UK population tariffs39 to produce a health state utility score for each participant in each of the treatment groups using the AUC method. 80

Given the large number of time points for EQ-5D-5L data, and to align with the primary effectiveness analysis, it was decided that, to be included in the AUC analysis without imputation, the participant must have at least three EQ-5D-5L observations with one at the start of the assessment period, one at the mid-range and one at the end, and the specific requirement was dependent on the status of the participants (whether or not they had received surgery) and the type of analysis conducted, as explained below.

For those participants who did not receive an initial surgery, to be included in the AUC analysis without imputation, they must have completed EQ-5D-5L data on the three time points: baseline, 18 months and 24 months after randomisation.

For those participants who received an initial surgery, the base-case analysis examined QALYs over the period from baseline to 24 months after randomisation; therefore, the AUC analysis required complete EQ-5D-5L data at baseline and 24 months after randomisation, and at one of the data collection points of 3, 6, 9 and 12 months following surgery and 18 months following randomisation. Given the differences in the duration of the period between randomisation and undergoing an intervention between urethroplasty and urethrotomy, we also examined QALYs over the period from the time prior to surgery to 24 months post surgery, in which case the AUC analysis required complete EQ-5D-5L data at prior to surgery and 24 months after surgery, and at one of the data collection points of 3, 6, 9 and 12 months following surgery and 18 and 24 months following randomisation.

For all calculations of QALYs the first observation used (either baseline or prior to surgery, depending on the patient groups and analysis type) was set at time point zero and the date on which the EQ-5D-5L was recorded as being completed was used to calculate the number of days from the first observation. This time dimension was initially used in the calculation of QALYs rather than the nominal time point at which the EQ-5D-5L was to be completed. In an alternative analysis, the responses were rescaled to the nominal data collection points (i.e. 730 days; note that cost data did not need to be rescaled in the same way as the recall period was predefined within the data collection tools). Additionally, multiple imputation for EQ-5D-5L at all missing time points was conducted to calculate QALYs for all participants.

A summary of the results of the EQ-5D-5L for each time point by intervention, as well as the estimates of QALYs for both the base-case analysis (QALYs at 24 months after randomisation) and the sensitivity analyses described above, is provided in Table 31 and in Figure 12.

FIGURE 12.

Utility values at each time point and QALYs over the trial follow-up with corresponding number of observations.

Cost–utility analysis

Results

The base-case analysis and all sensitivity analysis results are presented in Table 32. In the base case, urethroplasty cost more (cost difference £2148, 95% CI £689 to £3606) than urethrotomy, while generating a lower point estimate of QALY gain (QALY difference –0.01, 95% CI –0.17 to 0.14). Urethroplasty was therefore dominated by urethrotomy. This suggests that urethrotomy was more cost-effective than urethroplasty in the base-case scenario. The base-case result appears to be robust, as it is seen in all of the sensitivity analyses that urethroplasty cost more and the point estimate of effect was lower. The cost-effectiveness acceptability curve (Figure 13) and incremental cost and QALY plots (Figure 14) are presented for the base case. As seen in Figure 13, the probability of urethroplasty being cost-effective increases as the threshold for society’s willingness to pay for a QALY increases; however, urethroplasty never had more than approximately a 20% chance of being considered cost-effective over a 2-year time horizon for the range of cost per QALY thresholds considered. This is supported by the incremental cost and QALY plots (see Figure 14), in which the mean incremental cost per QALY falls in the north-east quadrant of the graph, indicating urethroplasty to cost more but to be less effective using the point estimate of benefit.

FIGURE 13.

Cost-effectiveness acceptability curve (base case).

FIGURE 14.

Incremental cost and QALY plots (base case).

Methods

The Markov model consisted of three health states: (1) cured (symptom free), (2) symptomatic and (3) deceased. The deceased state was included to allow for general population mortality when examining the cohorts over a long-term horizon. In the model, every patient started off receiving one of the two trial interventions. They then moved through the care pathways over time, incurring costs and accruing QALYs. The care pathways followed described the process of care, disease incidence and progression. These features were linked in a logical way, defined by a set of mathematical relationships that dictated how and when an individual might move through the model. The structure of the model is shown in Figure 15. The model parameters were based on information derived from the trial and the distribution for each parameter was defined considering the mean, standard error and anticipated shape of the distribution. Model parameters included costs for each intervention and their follow-up; utility for cured (symptom-free) and symptomatic states; probability of surgery success for each intervention; probability of receiving treatment when symptomatic; and probability of the type of treatment procedure received conditional on the previous intervention. Survival analysis was used to generate forward transition probabilities of recurrence following each intervention. Table 33 describes all model parameters. Half-cycle corrections were applied in the model.

FIGURE 15.

Markov model structure.

The base-case analysis used parameters estimated based on information from study participants who received the intervention to which they had been allocated. Probabilistic sensitivity analysis of the base-case scenario was carried out using Monte Carlo simulation, in which model inputs for each parameter were randomly selected from predefined distributions and the results recorded. This process was repeated for 10,000 iterations to produce a distribution of results from the model.

Deterministic sensitivity analyses were conducted by changing the key parameters in set ways that were determined based on information from those men who received the same intervention, regardless of whether or not it was the intervention they were allocated.

An important model parameter is the choice of the next treatment given the previous treatment the patient had. The parameters used in the model were based on observations from the trial, in which about 70% of patients would receive urethroplasty and 30% of patients would receive urethrotomy if the last treatment was urethrotomy, and about 12% of patients would receive urethroplasty and 88% of patients would receive urethrotomy if the last treatment was urethroplasty. This suggests that a large proportion of patients switch to a treatment different from their previous one every time they have a reintervention. There is no consensus on the treatment choices in reintervention and such choices are often influenced by many non-clinical factors, such as patient choice, waiting time and travel time. Given the high uncertainty around treatment choice in reintervention, deterministic sensitivity analyses also examined the impact of changing the percentages of patients switching treatments.

The costs and effects were combined in an incremental analysis, comparing different treatment strategies, and presented as the point estimates of mean costs, mean QALYs and mean incremental cost per QALY. These results were presented graphically using the incremental cost and QALY plots and a cost-effectiveness acceptability curve.

Results

In the base-case analysis, in which model parameters were based on those men who received the intervention they were allocated by randomisation, urethroplasty is unlikely to be considered cost-effective under society’s current willingness to pay for a QALY. This is also illustrated by the probabilistic sensitivity analysis in which urethroplasty has almost no chance of being considered cost-effective at any of the threshold values for society’s willingness to pay for a QALY considered (Table 34 and Figures 16 and 17).

FIGURE 16.

Cost-effectiveness acceptability curve (Markov model – base case).

FIGURE 17.

Incremental costs and QALY plots (model base case).

In the deterministic sensitivity analysis, in which model parameters were based on those men who received the same treatment procedure regardless of their allocated intervention group, a similar trend is observed to the base-case analysis, with urethroplasty being unlikely to be cost-effective under society’s current willingness-to-pay threshold for a QALY. The same is observed in the deterministic sensitivity analyses varying the probabilities of the choices of treatments in reintervention.

Both probabilistic and deterministic sensitivity analyses show that urethroplasty is unlikely to be cost-effective over a 10-year time horizon under a number of plausible scenarios. This is mainly driven by the higher cost of urethroplasty than urethrotomy, while both of the treatment procedures produce similar QALY gains, despite those men receiving urethroplasty having a lower chance of recurrence than those men receiving urethrotomy. It could be argued that the EQ-5D-5L may not be a sensitive measure for this group of patients and would, therefore, not capture important changes in quality of life for this condition. Additionally, decrement in quality of life following the procedures were not taken into consideration in the analysis conducted. Given the likelihood of recurrence requiring repeated treatments, this decrement could make a difference in QALY gains between the two treatments. The potential way to measure this decrement was explored using the TTO exercise.

Development of time trade-off materials

Prior to the TTO exercise being conducted among trial participants, extensive development and piloting of materials and processes were carried out. The TTO materials developed include a decision board, which was used as a visual aid for participants to compare health states, and written health state profiles (see Appendix 7), which described the likely side effects of the two procedures (categorised into mild, moderate and severe). These profiles were printed on A6-sized card for use with the decision board. An anchor health state was designed for the chained TTO, as an intermediate state that was worse than the six health profiles, but not as severe as death, and described a scenario containing severe pain with usual functioning impaired, but maintaining basic self-caring activities possible. A set of practice profiles was also designed based on a selection of EQ-5D-5L profiles to allow participants to become familiar with the process prior to valuing the health states relating to the study. The time horizon for the profiles to be evaluated was chosen as 14 days, and this was a balance between providing a clinically accurate prediction of the likely duration of side effects and providing participants with a clear scenario.

Three rounds of piloting were conducted with different groups of participants before finalising the materials and process: round 1 conducted with 17 male and female volunteers from Newcastle University; round 2 conducted with 15 male staff members at a participating (hospital) study site; and round 3 conducted with nine men eligible for the open trial. Piloting interviews consisted of mock TTO interviews, but also encouraged feedback on the process and materials. Feedback was collated and used iteratively to refine processes for subsequent pilots. Health-care professionals involved in the trial were consulted throughout the development process to ensure that the TTO materials were clinically accurate and understandable to a lay audience. A patient and public involvement representative was also consulted as part of the piloting process.

The TTO participants were recruited from those men who were eligible for the OPEN trial. At the point of consent to the main trial, participants were asked to indicate willingness to be contacted about a further interview study. Those men who expressed interest were posted an information pack regarding TTO, containing a response slip and post-paid envelope. On receipt of an affirmative response slip, a researcher contacted respondents to answer any further questions or arrange a time and place of the participant’s choosing. Interviews were most frequently conducted in the participant’s home and written consent was taken prior to commencing the interview.

Following the TTO interview, participants were asked to provide feedback on the interview process. Participants were asked to rate on a scale of 1 to 5 (with 1 being not difficult at all and 5 being very difficult) how difficult they found the TTO process and asked if they had any comments relating to the process. This was recorded by the researcher as free text. In addition, the interviewer recorded anything about the conduct and process of the interview that they felt to be important.

Data analysis

Stata was used to analyse data. Descriptive statistics were produced for demographic details of the sample and Tobit regressions of reported TTO values were performed, controlling for sociodemographic characteristics. The level of usual physical activity and place of residence (urban or rural) were included as variables in the analyses, as it was considered possible that these would have an impact on how participants valued health states in which their mobility was affected. An additional control variable was generated based on the consistency between utility values derived and how each participant ranked those profiles prior to the TTO exercise in order to indicate data quality. Predicted values of health state utility were estimated for each health state and compared using unpaired t-tests. Qualitative data collected were also analysed to understand the feasibility of conducting a TTO alongside a surgical RCT.

Results

Forty participants were recruited to take part in the TTO. Two participants had missing or invalid essential data and were excluded. Of the 38 participants included in the analysis, the sociodemographic details were as follows: mean age of the sample was 38 years; 84% were married; 41% had a household income > £36,400; just over half (55%) of the sample were employed and 29% were retired. The majority (71%) of participants described living in an urban area. Twenty-nine per cent of the sample reported high levels of physical activity, 47% reported medium levels and 24% reported low levels of activity.

The estimated utility values for the two types of surgery using the two TTO methods are presented in Table 35, together with significance of the difference in post-surgery utilities between the two surgeries. The majority of the urethroplasty health states had lower utility scores than the urethrotomy health states, this was significant for the severe scenario irrespective of the TTO method used. For both types of surgery, mean utility values decreased as health states became more severely impaired, which suggests face validity of the profiles.

The median difficulty for the TTO exercise reported by participants was 2 out of 5, which indicated that the TTO exercise was considered by most participants as reasonably easy to complete. Additional comments covered participants’ opinions on the difficulty of the interview process, as well as elaborating on their individual decision-making process. Some described a desire to avoid particular symptoms, particularly pain and anxiety. Some participants related previous life experience as informing their perception of the impact of certain symptoms; for example, one participant stated that having experienced frequent catheterisation, this was something he no longer worried about. In general, the interviewers felt that participants had understood the process reasonably well.

Discussion

Using a TTO nested in a clinical RCT allowed estimation of the short-term impact on HRQoL from side effects arising immediately following the procedures being compared.

Lower utility values were observed for urethroplasty-related health states, implying that there was a larger reduction in HRQoL immediately following the urethroplasty procedure, although the differences were seen to be statistically significant only for the severe impairment health states. Although this may be due to insufficient sample size, when taken in the context of the qualitative feedback regarding decision-making, this could be because participants were particularly averse to one or more of the symptoms described in the severe urethroplasty health state.

Only differences between utilities for each intervention were examined statistically, a more meaningful comparison would be to examine whether or not these differences were relevant from a clinical perspective. Further understanding of the short-term impact of the different treatments would offer valuable information from a policy-making perspective, as well as improving patient information on treatment choices. Combining these data with other information, such as incidence of side effects and recurrence rates, may help to support shared decision-making between patients and clinicians.

No attempt was formally made to incorporate the results of the exploratory analysis of the TTO exercise into the economic valuation. However, if the results were taken at face value and knowing something about the number of reoperations in both arms of the trial, then we might expect a more marked difference in QALYs between the two treatment groups, given that patients in the urethrotomy group are more likely to require repeated treatments than those patients in the urethroplasty group. Such marked difference is unlikely to come from the impact of the treatment itself if the findings were robust in terms of little difference found in utility between the two treatment strategies, except in the severe state which may account for only a small percentage of the patient population. Although qualitative comments were generally brief and did not constitute a valid or robust qualitative interview, they were useful to add to our understanding of people’s experience of taking part in the interviews and highlight aspects which were deemed to be difficult and could be revised for future studies. The aim of this research was to ask people to compare two states of health for varying lengths of time. People’s decision-making process appeared to be influenced by multiple factors, including their own experience of health and illness, their current health and their family situation. These data are useful in developing future TTO exercises and, in particular, will help research teams to refine their information materials, such as introductory scripts.

Conclusion

The TTO exercise has proved to be a feasible method of collecting data on short-term changes in HRQoL within a clinical trial. Although the study finding is preliminary, it suggests that both urethroplasty and urethrotomy have a negative impact on HRQoL immediately following the procedures. The likelihood of men with recurrent urethral stricture requiring repeated treatments may indicate that this important but short-lived decrement to HRQoL immediately following the procedures should be incorporated into the calculation of QALYs associated with each intervention in the longer term. Future research should aim to elicit the utilities associated with this decrement with a large sample size.

The conventional method

Respondents will be offered a choice between two alternative ‘lives’. Life A, containing the less desirable health state (temporary health; h_i), is measured relative to life B, which contains the best health state (perfect health; 1). The respondents will be shown on the decision board one of the health state profiles describing the health consequences post treatment for the length of time (t) as life A and the perfect health profile as life B. The time in life A will be fixed at t (14 days), whereas there is a moving slider for the length of time (between 0 and t) spent in life B. Both lives are set at the maximum length t initially. Respondents will be asked to find a point on the board with the slider pointing at a time (x) for life B, at which they are indifferent to the length of time spent in life A (t) and life B (x). This is achieved by asking respondents a series of questions for each profile while moving the slider for life B. ‘Assuming you are either in life A or life B for the specified number of days shown on the board, after which time you will die painlessly, which life would you prefer?’ If the participant has a preference for life B, we will reduce the time (x) spent in life B, and if the respondent prefers life A, we will increase the time (x) spent in life B. The above question will be repeated until a point of indifference is reached, by confirming with the respondent that ‘you consider spending t days in life A is equivalent to spending x days in life B’, then we will take a note for the length of time the respondent settles on (x). We will do the same for all six health profiles to be valued in a random order.

For analysis, the value of each health state (h_i) is then calculated as:

The chained method

This method requires two stages of comparison. In the first stage, we will ask the respondents to compare each health state profile (h_i) as life A with an anchor state (h_j) as life B. As with the conventional method, the number of days (t) in life A will be fixed at 14 days and the number of days (x) in life B will also be initially set at 14 days. Respondents will be asked a similar question to that in the conventional method, but instead of death following the length of time spent in the health states, they would return to full health: ‘Assuming you are either in life A or life B for the specified number of days shown on the board, after which time you will return to full health, which life would you prefer?’ If the participant has a preference for life B, we will increase the time (x) spent in life B; and if the respondent prefers life A, we will reduce the time (x) spent in life B. The above question will be repeated until a point of indifference is reached, by confirming with the respondent that ‘you consider spending t days in life A is equivalent to spending x days in life B’, then we will take a note for the length of time the respondent settles on (x). We will do the same for all six health profiles to be valued in a random order.

The second stage is then carried out to compare the anchor state profile with the perfect health profile. In this stage, participants will choose between the anchor state as life A for 14 (t) days and perfect health as life B for (y) days conducted using the conventional TTO approach. We choose to leave the value as 14 days to be consistent with the rest of the valuations, despite the length of life being relatively short for use in a conventional TTO method.

For analysis, the value of the anchor state (h_j) is calculated as:

The utility of each health profile is then calculated as:

Control intervention urethrotomy health state profiles

Experimental intervention urethroplasty health state profiles