Multivariate network meta-analysis incorporating class effects.

BACKGROUND: Network meta-analysis synthesises data from a number of clinical trials in order to assess the comparative efficacy of multiple healthcare interventions in similar patient populations. In situations where clinical trial data are heterogeneously reported i.e. data are missing for one or more outcomes of interest, synthesising such data can lead to disconnected networks of evidence, increased uncertainty, and potentially biased estimates which can have severe implications for decision-making. To overcome this issue, strength can be borrowed between outcomes of interest in multivariate network meta-analyses. Furthermore, in situations where there are relatively few trials informing each treatment comparison, there is a potential issue with the sparsity of data in the treatment networks, which can lead to substantial parameter uncertainty. A multivariate network meta-analysis approach can be further extended to borrow strength between interventions of the same class using hierarchical models. METHODS: We extend the trivariate network meta-analysis model to incorporate the exchangeability between treatment effects belonging to the same class of intervention to increase precision in treatment effect estimates. We further incorporate a missing data framework to estimate uncertainty in trials that did not report measures of variability in order to maximise the use of all available information for healthcare decision-making. The methods are applied to a motivating dataset in overactive bladder syndrome. The outcomes of interest were mean change from baseline in incontinence, voiding and urgency episodes. All models were fitted using Bayesian Markov Chain Monte Carlo (MCMC) methods in WinBUGS. RESULTS: All models (univariate, multivariate, and multivariate models incorporating class effects) produced similar point estimates for all treatment effects. Incorporating class effects in multivariate models often increased precision in treatment effect estimates. CONCLUSIONS: Multivariate network meta-analysis incorporating class effects allowed for the comparison of all interventions across all outcome measures to ameliorate the potential impact of outcome reporting bias, and further borrowed strength between interventions belonging to the same class of treatment to increase the precision in treatment effect estimates for healthcare policy and decision-making.

Invasive urodynamic testing prior to surgical treatment for stress urinary incontinence in women: cost-effectiveness and value of information analyses in the context of a mixed methods feasibility study.

BACKGROUND: INVESTIGATE-I (INVasive Evaluation before Surgical Treatment of Incontinence Gives Added Therapeutic Effect?) was a mixed methods study to assess the feasibility of a future randomised controlled trial of invasive urodynamic testing (IUT) prior to surgery for stress urinary incontinence (SUI) in women. Here we report one of the study's five components, with the specific objectives of (i) exploring the cost-effectiveness of IUT compared with clinical assessment plus non-invasive tests (henceforth described as 'IUT' and 'no IUT' respectively) in women with SUI or stress-predominant mixed urinary incontinence (MUI) prior to surgery, and (ii) determining the expected net gain (ENG) from additional research. METHODS: Study participants were women with SUI or stress-predominant MUI who had failed to respond to conservative treatments recruited from seven UK urogynaecology and female urology units. They were randomised to receive either 'IUT' or 'no IUT' before undergoing further treatment. Data from 218 women were used in the economic analysis. Cost utility, net benefit and value of information (VoI) analyses were performed within a randomised controlled pilot trial. Costs and quality-adjusted life years (QALYs) were estimated over 6 months to determine the incremental cost per QALY of 'IUT' compared to 'no IUT'. Net monetary benefit informed the VoI analysis. The VoI estimated the ENG and optimal sample size for a future definitive trial. RESULTS: At 6 months, the mean difference in total average cost was £138 (p = 0.071) in favour of 'IUT'; there was no difference in QALYs estimated from the SF-12 (difference 0.004; p = 0.425) and EQ-5D-3L (difference - 0.004; p = 0.725); therefore, the probability of IUT being cost-effective remains uncertain. The estimated ENG was positive for further research to address this uncertainty with an optimal sample size of 404 women. CONCLUSIONS: This is the largest economic evaluation of IUT. On average, up to 6 months after treatment, 'IUT' may be cost-saving compared to 'no IUT' because of the reduction in surgery following invasive investigation. However, uncertainty remains over the probability of 'IUT' being considered cost-effective, especially in the longer term. The VoI analysis indicated that further research would be of value. TRIAL REGISTRATION: ISRCTN. ISRCTN71327395. Registered 7 June 2010.

PROPEL: implementation of an evidence based pelvic floor muscle training intervention for women with pelvic organ prolapse: a realist evaluation and outcomes study protocol.

BACKGROUND: Pelvic Organ Prolapse (POP) is estimated to affect 41%-50% of women aged over 40. Findings from the multi-centre randomised controlled "Pelvic Organ Prolapse PhysiotherapY" (POPPY) trial showed that individualised pelvic floor muscle training (PFMT) was effective in reducing symptoms of prolapse, improved quality of life and showed clear potential to be cost-effective. However, provision of PFMT for prolapse continues to vary across the UK, with limited numbers of women's health physiotherapists specialising in its delivery. Implementation of this robust evidence from the POPPY trial will require attention to different models of delivery (e.g. staff skill mix) to fit with differing care environments. METHODS: A Realist Evaluation (RE) of implementation and outcomes of PFMT delivery in contrasting NHS settings will be conducted using multiple case study sites. Involving substantial local stakeholder engagement will permit a detailed exploration of how local sites make decisions on how to deliver PFMT and how these lead to service change. The RE will track how implementation is working; identify what influences outcomes; and, guided by the RE-AIM framework, will collect robust outcomes data. This will require mixed methods data collection and analysis. Qualitative data will be collected at four time-points across each site to understand local contexts and decisions regarding options for intervention delivery and to monitor implementation, uptake, adherence and outcomes. Patient outcome data will be collected at baseline, six months and one year follow-up for 120 women. Primary outcome will be the Pelvic Organ Prolapse Symptom Score (POP-SS). An economic evaluation will assess the costs and benefits associated with different delivery models taking account of further health care resource use by the women. Cost data will be combined with the primary outcome in a cost effectiveness analysis, and the EQ-5D-5L data in a cost utility analysis for each of the different models of delivery. DISCUSSION: Study of the implementation of varying models of service delivery of PFMT across contrasting sites combined with outcomes data and a cost effectiveness analysis will provide insight into the implementation and value of different models of PFMT service delivery and the cost benefits to the NHS in the longer term.

A mixed methods study to assess the feasibility of a randomised controlled trial of invasive urodynamic testing versus clinical assessment and non-invasive tests prior to surgery for stress urinary incontinence in women: the INVESTIGATE-I study.

BACKGROUND: The position of invasive urodynamic testing (IUT) in diagnostic pathways for urinary incontinence is unclear, and systematic reviews have called for further trials evaluating clinical utility. The objective of this study was to inform the decision whether to proceed to a definitive randomised trial of IUT compared to clinical assessment with non-invasive tests, prior to surgery in women with stress urinary incontinence (SUI) or stress-predominant mixed urinary incontinence (MUI). METHODS: A mixed methods study comprising a pragmatic multicentre randomised pilot trial, a qualitative face-to face interview study with patients eligible for the trial, an exploratory economic evaluation including value of information study, a survey of clinicians' views about IUT, and qualitative telephone interviews with purposively sampled survey respondents. Only the first and second of these elements are reported here. Trial participants were randomised to either clinical assessment with non-invasive tests (control arm) or clinical assessment with non-invasive tests plus IUT (intervention arm). The main outcome measures of these feasibility studies were confirmation that units can identify and recruit eligible women, acceptability of investigation strategies and data collection tools, and acquisition of outcome data to determine the sample size for a definitive trial. The primary outcome proposed for a definitive trial was ICIQ-FLUTS (total score) 6 months after surgery or the start of nonsurgical treatment. RESULTS: Of 284 eligible women, 222 (78%) were recruited, 165/219 (75%) returned questionnaires at baseline, and 125/200 returned them (63%) at follow-up. Most women underwent surgery; management plans were changed in 19 (19%) participants following IUT. Participants interviewed were positive about the trial and the associated documentation. CONCLUSIONS: All elements of a definitive trial were rehearsed. Such a trial would require between 232 and 922 participants, depending on the target difference in the primary outcome. We identified possible modifications to our protocol for application in a definitive trial including clarity over inclusion/exclusions, screening processes, reduction in secondary outcomes, and modification to patient questionnaire booklets and bladder diaries. A definitive trial of IUT versus clinical assessment prior to surgery for SUI or stress predominant MUI is feasible and remains relevant. TRIAL REGISTRATION: Current Controlled Trials: ISRCTN 71327395, registered 7 June 2010.

INVESTIGATE-I (INVasive Evaluation before Surgical Treatment of Incontinence Gives Added Therapeutic Effect?): a mixed-methods study to assess the feasibility of a future randomised controlled trial of invasive urodynamic testing prior to surgery for stress urinary incontinence in women.

BACKGROUND: The position of invasive urodynamic testing in the diagnostic pathway for urinary incontinence (UI) is unclear. Systematic reviews have called for further trials evaluating clinical utility, although a preliminary feasibility study was considered appropriate. OBJECTIVES: To inform the decision whether or not to proceed to a definitive randomised trial of invasive urodynamic testing compared with clinical assessment with non-invasive tests, prior to surgery in women with stress UI (SUI) or stress predominant mixed UI (MUI). DESIGN: A mixed-methods study comprising a pragmatic multicentre randomised pilot trial; economic evaluation; survey of clinicians' views about invasive urodynamic testing; qualitative interviews with clinicians and trial participants. SETTING: Urogynaecology, female urology and general gynaecology units in Newcastle, Leicester, Swansea, Sheffield, Northumberland, Gateshead and South Tees. PARTICIPANTS: Trial recruits were women with SUI or stress predominant MUI who were considering surgery after unsuccessful conservative treatment. Relevant clinicians completed two online surveys. Subsets of survey respondents and trial participants took part in separate qualitative interview studies. INTERVENTIONS: Pilot trial participants were randomised to undergo clinical assessment with non-invasive tests (control arm); or assessment as controls, plus invasive urodynamic testing (intervention arm). MAIN OUTCOME MEASURES: Confirmation that units can identify and recruit eligible women; acceptability of investigation strategies and data collection tools; acquisition of outcome data to determine the sample size for a definitive trial. The proposed primary outcome for the definitive trial was International Consultation on Incontinence Modular Questionnaire (ICIQ) Female Lower Urinary Tract Symptoms (ICIQ-FLUTS) (total score) 6 months after surgery or the start of non-surgical treatment; secondary outcomes included: ICIQ-FLUTS (subscales); ICIQ Urinary Incontinence Short Form; ICIQ Lower Urinary Tract Symptoms Quality of Life; Urogenital Distress Inventory; EuroQol-5D; costs, quality-adjusted life-years (QALYs) and incremental cost per QALY, Short Form 12; 3-day bladder diary. RESULTS: Of 284 eligible women, 222 (78%) were recruited; 165/219 (75%) returned questionnaires at baseline and 125/200 (63%) who were sent questionnaires at follow-up. There were few missing data items in returned questionnaires, with individual outcome scales calculable for 81%-94%. Most women underwent surgery; management plans were changed in 19 (19%) participants following invasive urodynamic testing. Participant Costs Questionnaires were returned by 53% 6 months after treatment; complete data to undertake cost-utility analysis were available in 27% (intervention) and 47% (control). While insufficient to recommend changes in practice, the results suggest further research would be valuable. All clinicians responding to the survey had access to invasive urodynamic testing, and most saw it as essential prior to surgery in women with SUI with or without other symptoms; nevertheless, 70% considered the research question underlying INVESTIGATE important and most were willing to randomise patients in a definitive trial. Participants interviewed were positive about the trial and associated documentation; the desire of some women to avoid invasive urodynamic testing contrasted with opinions expressed by clinicians through both survey and interview responses. CONCLUSIONS: All elements of a definitive trial and economic evaluation were rehearsed; several areas for protocol modification were identified. Such a trial would require to 400-900 participants, depending on the difference in primary outcome sought. FUTURE WORK: A definitive trial of invasive urodynamic testing versus clinical assessment prior to surgery for SUI or stress predominant MUI should be undertaken. TRIAL REGISTRATION: Current Controlled Trials ISRCTN71327395. FUNDING: The National Institute for Health Research Health Technology Assessment programme.

Network meta-analysis: development of a three-level hierarchical modeling approach incorporating dose-related constraints.

BACKGROUND: Network meta-analysis (NMA) is commonly used in evidence synthesis; however, in situations in which there are a large number of treatment options, which may be subdivided into classes, and relatively few trials, NMAs produce considerable uncertainty in the estimated treatment effects, and consequently, identification of the most beneficial intervention remains inconclusive. OBJECTIVE: To develop and demonstrate the use of evidence synthesis methods to evaluate extensive treatment networks with a limited number of trials, making use of classes. METHODS: Using Bayesian Markov chain Monte Carlo methods, we build on the existing work of a random effects NMA to develop a three-level hierarchical NMA model that accounts for the exchangeability between treatments within the same class as well as for the residual between-study heterogeneity. We demonstrate the application of these methods to a continuous and binary outcome, using a motivating example of overactive bladder. We illustrate methods for incorporating ordering constraints in increasing doses, model selection, and assessing inconsistency between the direct and indirect evidence. RESULTS: The methods were applied to a data set obtained from a systematic literature review of trials for overactive bladder, evaluating the mean reduction in incontinence episodes from baseline and the number of patients reporting one or more adverse events. The data set involved 72 trials comparing 34 interventions that were categorized into nine classes of interventions, including placebo. CONCLUSIONS: Bayesian three-level hierarchical NMAs have the potential to increase the precision in the effect estimates while maintaining the interpretability of the individual interventions for decision making.

The effect of vaginal closure technique on early post-operative pain following vaginal prolapse surgery: a feasibility pilot study and qualitative assessment.

BACKGROUND: Surgery for pelvic organ prolapse is a common surgical procedure. There is little research studying post-operative pain, contrasting with extensive literature on pain after childbirth trauma which shows that suture method has a significant effect on pain. We designed a protocol for a trial comparing suture techniques and post-operative pain, and conducted a pre-protocol pilot to inform our trial design. METHODS: Routine surgical data, post-operative pain scores (from 10 cm Visual Analogue Scales, VAS) and analgesic requirements were obtained from the notes of a cohort of women undergoing vaginal prolapse surgery. Median VAS scores at 4, 12 and 24 hours were compared by suture material used and method of closure (single continuous suture or interrupted sutures). The women whose data were obtained were invited to attend focus groups of up to six people in the twelve weeks following surgery. A semi structured question guide was used, and interviews were recorded, transcribed anonymously and analysed using the constant comparative method of grounded theory. Ethical approval was not sought because formal pre-protocol work is exempt for UK ethical requirements, but formal written consent on standard forms was obtained for publishing outcomes and anonymised comments from participants. RESULTS: Complete VAS data and operative details were available from 41 women. Surgery was completed using absorbable polyglactin sutures with continuous suturing (17 women) or interrupted sutures (24 women). Pain scores at 4 and 12 hours were similar; pain at 24 hours was greater in the women with interrupted sutures (median VAS 3 (range 0-8) versus 1.5 (0-8) (p = 0.0513). Analgesic needs were similar. Two focus groups (nine participants) revealed that women regarded post-operative pain as insignificant and not a topic worthy of formal research. It was apparent that the use, and especially removal of, vaginal packs was a practice associated with pain which women remembered as a significant part of their post-operative experience. CONCLUSIONS: Formal pre-protocol work is informative; we found a moderate difference in our proposed outcomes, suggesting a trial was feasible but women themselves were unconvinced of the need for formal research into pain following vaginal surgery.

Cost-effectiveness of duloxetine: the Stress Urinary Incontinence Treatment (SUIT) study.

OBJECTIVE: To assess the cost-effectiveness of duloxetine compared with conservative therapy in women with stress urinary incontinence (SUI). METHODS: Cost and outcome data were taken from the Stress Urinary Incontinence Treatment (SUIT) study, a 12-month, prospective, observational, naturalistic, multicenter, multicountry study. Costs were assessed in UK pound and outcomes in quality adjusted life years using responses to the EuroQol (EQ-5D); numbers of urine leaks were also estimated. Potential selection bias was countered using multivariate regression and propensity score analysis. RESULTS: Duloxetine alone, duloxetine in combination with conservative treatment, and conservative treatment alone were associated with roughly two fewer leaks per week compared with no treatment. Duloxetine alone and with conservative treatment for SUI were associated with incremental quality-adjusted life-years (QALYs) of about 0.03 over a year compared with no treatment or with conservative treatment alone. Conservative treatment alone did not show an effect on QALYs. None of the interventions appeared to have marked impacts on costs over a year. Depending on the form of matching, duloxetine either dominated or had an incremental cost-effectiveness ratio (ICER) below pound900 per QALY gained compared with no treatment and with conservative treatment alone. Duloxetine plus conservative therapy had an ICER below pound5500 compared with no treatment or conservative treatment alone. Duloxetine compared with duloxetine plus conservative therapy showed similar outcomes but an additional cost for the combined intervention. CONCLUSIONS: Although the limitations of the use of SUIT's observational data for this purpose need to be acknowledged, the study suggests that initiating duloxetine therapy in SUI is a cost-effective treatment alternative.

Patient characteristics impacting health state index scores, measured by the EQ-5D of females with stress urinary incontinence symptoms.

OBJECTIVE: To describe the characteristics of women seeking treatment for symptoms of stress urinary incontinence (SUI) and to investigate the association of SUI symptoms with generic health-related quality of life (HRQoL) as measured by the EuroQol (EQ-5D) instrument. METHODS: The Stress Urinary Incontinence Treatment (SUIT) study was a 12-month observational study in four European countries that evaluated the cost-effectiveness of duloxetine compared with other forms of nonsurgical intervention in the treatment of the symptoms of SUI. Four hundred thirty-one physicians observed women seeking treatment for their SUI, and recorded the care provided and the outcomes of that care at enrollment and at 3, 6 and 12 months afterward The impact of SUI on baseline HRQoL as expressed by the EQ-5D index score was assessed by linear and logistic regression. RESULTS: Three thousand seven hundred sixty-two women were enrolled into SUIT, with the largest patient group from Germany. Overall, the majority of women were postmenopausal, had a mean age of 58.0 years, were not current smokers, and tended to be overweight (mean body mass index [BMI]=27.7 kg/m2), with at least one comorbidity. The health state index scores were significantly and independently influenced by the presence of comorbidity(ies) affecting quality of life, total number of stress and urge incontinence episodes, urinary incontinence subtype, comorbidity(ies) affecting incontinence, BMI, socioeconomic status, educational status, age, and country. CONCLUSION: This article describes the characteristics of patients at the SUIT enrollment visit, and demonstrates that the number of incontinence episodes has a significant impact on the EQ-5D index score.

Objective assessment of urinary incontinence in women: comparison of the one-hour and 24-hour pad tests.

OBJECTIVE: To compare the performance and clinical usefulness of the one-hour and 24-hour pad tests in terms of the relationship with reported symptoms and urodynamic diagnosis. METHODS: 341 women aged 40 years and over reporting lower urinary tract symptoms who were recruited to a nurse led continence service, and went on to receive urodynamics, a one-hour and a 24-hour pad test and completed a urinary diary. RESULTS: For both pad tests, there was a significant difference in the amount of urine leaked between the different urodynamic diagnoses (p<0.0001, for the one-hour and p=0.001 for the 24-hour test). Women with sphincter incompetence leaked significantly more than women with detrusor instability, or those with no abnormality. There was a significant difference between the proportion of women dry on a one-hour pad test and those dry on a 24-hour pad test (26.0% versus 38.4%, difference 12.4%; CI 5.5; 19.4). There is a positive relationship between amount of urine leakage and symptom severity expressed in terms of number of incontinent episodes for both pad tests. CONCLUSION: Both pad tests bore little relationship to the underlying urodynamic diagnosis but there was a positive relationship with symptom severity. The 24-hour pad test appears to be clinically a more useful too than the one-hour test. The two types of pad test are probably assessing incontinence in different ways. We suggest that the minimum data set should include structured questions, diaries and the 24-hour pad test.