Tony Fouweather

Background: There are limited options currently recommended in National Institute for Health and Care Excellence guidelines for the treatment of bipolar depression. Pramipexole has been shown to improve mood symptoms in two small pilot studies in such patients. Objectives: Primary: to evaluate the clinical effectiveness of pramipexole versus placebo alongside routine mood-stabilising medications over 12 weeks in patients with treatment-resistant bipolar depression. Secondary: evaluate the impact of pramipexole on mood and anxiety, psychosocial function, cost-effectiveness, and safety and tolerability over 48 weeks. Design: Multicentre, randomised, placebo-controlled trial of pramipexole versus placebo in addition to standard-of-care mood stabilisers. Clinicians, researchers and participants were blinded throughout the duration of the study. Pre-randomisation stage (to adjust antipsychotics or commence mood stabilisers where required) before randomisation. Weekly online assessments of mood and anxiety from randomisation to week 52, with psychosocial function, quality of life and healthcare resource utilisation assessments conducted at regular intervals. Setting: Twenty-one National Health Service trusts and Health Boards across England and Scotland. Participants: Patients aged 18 years and over with a diagnosis of treatment-resistant bipolar depression currently under secondary care mental health services. Aim to randomise 290 participants. Interventions: Pramipexole or matched placebo orally once daily, titrated from 0.25 mg to maximum of 2.5 mg (salt weight) depending on efficacy and tolerability. Main outcome measures: Depression – Quick Inventory for Depressive Symptomology; anxiety – Generalised Anxiety Disorder-7-item scale; psychosocial functioning – Work and Social Adjustment Scale; hypomania/mania – Altman Self-rating Scale of Mania; tolerability – Treatment Satisfaction Questionnaire for Medication; well-being and quality of life – EuroQol-5 Dimensions, five-level version, ICEpop CAPability measure for Adults and Oxford CAPabilities questionnaire-Mental Health tools. Results: Thirty-nine participants randomised (18 to pramipexole and 21 to placebo) with 36 providing data for the primary analysis. Pramipexole led to greater reductions in depressive symptoms at 12 weeks compared to placebo [4.4 (4.8) vs. 2.1 (5.1)]: a medium-sized (d = −0.72) but not statistically significant difference (95% confidence interval −0.4 to 6.3; p = 0.087). There were some statistically significant positive effects of pramipexole on secondary outcomes (reduction in depressive symptoms at 36 weeks, response and remission rates at trial exit, psychosocial function). Pramipexole was associated with an increased rate of hypomania/manic symptoms, but this appeared to be reduced by coadministration with an antipsychotic. General tolerability of pramipexole was good. There were significant annual gains in health-related quality of life and capability-well-being and tendency towards reduced health and social care costs. Limitations: Small sample size and variable follow-up period due to recruitment during COVID-19 pandemic and the trial closing early. Participants limited to those in secondary care mental health services. All assessments only available in English. Conclusions: No change in clinical practice can be recommended as there was not a significant difference between pramipexole and placebo on the primary efficacy outcome measure. However, there was evidence of positive effects of pramipexole on mood, psychosocial function and quality of life. Future work: Replication in a larger population and research to investigate the impact of coadministration of antipsychotics alongside pramipexole. Trial registration: This trial is registered as ISRCTN72151939 and EudraCT 2018-2869-18. Funding: This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 16/154/01) and is published in full in Health Technology Assessment; Vol. 29, No. 21. See the NIHR Funding and Awards website for further award information.

Background: Options for ‘treatment-resistant bipolar depression’ (TRBD) are limited. Two small, short-term, trials of pramipexole suggest it might be an option. Aims: To evaluate the clinical effectiveness and safety of pramipexole in the management of TRBD. Methods: A multi-centre randomised, double-blind controlled trial including participants ⩾18 years old with TRBD (failure to respond, tolerate or clinical contraindication/patient refusal of ⩾2 of quetiapine, olanzapine, lamotrigine or lurasidone) randomised 1:1 to pramipexole (max 2.5 mg/day salt weight) or placebo added to ongoing mood stabiliser ( n = 39). Primary outcome: Quick Inventory of Depressive Symptoms, Self-rated (QIDS-SR) at 12 weeks. Up to 48 weeks follow-up. Results: Pramipexole ( n = 18) was associated with a greater reduction in QIDS-SR score at 12 weeks versus placebo ( n = 21, 4.4 (4.8) vs 2.1 (5.1)): a medium sized ( d = −0.72) but not statistically significant difference (95% CI: −0.4 to 6.3, p = 0.087). Similarly, there was a non-significant approximate 2-point ( d = −0.76) improvement in pleasure at 6 weeks (95% CI: −0.11 to 4.20). Significant advantages of pramipexole on QIDS-SR score (6.28 points: 95% CI: 1.85–10.71) and psychosocial function (5.36 points: 95% CI: 0.38–10.35) were seen at 36 weeks post-randomisation, and on the response (46% vs 6%; p = 0.026) and remission (31% vs 0%; p = 0.030) rates at trial exit (48 weeks or last available data after 16 weeks for those affected by the early study closure). Hypomania ratings were significantly higher at 12 weeks. Otherwise, pramipexole was well tolerated. Conclusions: Clinically large, but statistically non-significant, effects of pramipexole on depression at 12 weeks, with significant longer-term benefits on mood and function were observed. Pramipexole use was complicated by dose titration and increased hypomanic symptoms. The small sample size limits interpretation. Furthermore, larger randomised placebo-controlled trials are warranted.

Background: The indications for septoplasty are practice-based, rather than evidence-based. In addition, internationally accepted guidelines for the management of nasal obstruction associated with nasal septal deviation are lacking. Objective: The objective was to determine the clinical effectiveness and cost-effectiveness of septoplasty, with or without turbinate reduction, compared with medical management, in the management of nasal obstruction associated with a deviated nasal septum. Design: This was a multicentre randomised controlled trial comparing septoplasty, with or without turbinate reduction, with defined medical management; it incorporated a mixed-methods process evaluation and an economic evaluation. Setting: The trial was set in 17 NHS secondary care hospitals in the UK. Participants: A total of 378 eligible participants aged > 18 years were recruited. Interventions: Participants were randomised on a 1: 1 basis and stratified by baseline severity and gender to either (1) septoplasty, with or without turbinate surgery (n = 188) or (2) medical management with intranasal steroid spray and saline spray (n = 190). Main outcome measures: The primary outcome was the Sino-nasal Outcome Test-22 items score at 6 months (patient-reported outcome). The secondary outcomes were as follows: patient-reported outcomes – Nasal Obstruction Symptom Evaluation score at 6 and 12 months, Sino-nasal Outcome Test-22 items subscales at 12 months, Double Ordinal Airway Subjective Scale at 6 and 12 months, the Short Form questionnaire-36 items and costs; objective measurements – peak nasal inspiratory flow and rhinospirometry. The number of adverse events experienced was also recorded. A within-trial economic evaluation from an NHS and Personal Social Services perspective estimated the incremental cost per (1) improvement (of ≥ 9 points) in Sino-nasal Outcome Test-22 items score, (2) adverse event avoided and (3) quality-adjusted life-year gained at 12 months. An economic model estimated the incremental cost per quality-adjusted life-year gained at 24 and 36 months. A mixed-methods process evaluation was undertaken to understand/address recruitment issues and examine the acceptability of trial processes and treatment arms. Results: At the 6-month time point, 307 participants provided primary outcome data (septoplasty, n = 152; medical management, n = 155). An intention-to-treat analysis revealed a greater and more sustained improvement in the primary outcome measure in the surgical arm. The 6-month mean Sino-nasal Outcome Test-22 items scores were −20.0 points lower (better) for participants randomised to septoplasty than for those randomised to medical management [the score for the septoplasty arm was 19.9 and the score for the medical management arm was 39.5 (95% confidence interval −23.6 to −16.4; p < 0.0001)]. This was confirmed by sensitivity analyses and through the analysis of secondary outcomes. Outcomes were statistically significantly related to baseline severity, but not to gender or turbinate reduction. In the surgical and medical management arms, 132 and 95 adverse events occurred, respectively; 14 serious adverse events occurred in the surgical arm and nine in the medical management arm. On average, septoplasty was more costly and more effective in improving Sino-nasal Outcome Test-22 items scores and quality-adjusted life-years than medical management, but incurred a larger number of adverse events. Septoplasty had a 15% probability of being considered cost-effective at 12 months at a £20,000 willingness-to-pay threshold for an additional quality-adjusted life-year. This probability increased to 99% and 100% at 24 and 36 months, respectively. Limitations: COVID-19 had an impact on participant-facing data collection from March 2020. Conclusions: Septoplasty, with or without turbinate reduction, is more effective than medical management with a nasal steroid and saline spray. Baseline severity predicts the degree of improvement in symptoms. Septoplasty has a low probability of cost-effectiveness at 12 months, but may be considered cost-effective at 24 months. Future work should focus on developing a septoplasty patient decision aid. Trial registration: This trial is registered as ISRCTN16168569 and EudraCT 2017-000893-12. Funding: This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 14/226/07) and is published in full in Health Technology Assessment; Vol. 28, No. 10. See the NIHR Funding and Awards website for further award information.

Background: The place of tonsillectomy in the management of sore throat in adults remains uncertain. Objectives: To establish the clinical effectiveness and cost-effectiveness of tonsillectomy, compared with conservative management, for tonsillitis in adults, and to evaluate the impact of alternative sore throat patient pathways. Design: This was a multicentre, randomised controlled trial comparing tonsillectomy with conservative management. The trial included a qualitative process evaluation and an economic evaluation. Setting: The study took place at 27 NHS secondary care hospitals in Great Britain. Participants: A total of 453 eligible participants with recurrent sore throats were recruited to the main trial. Interventions: Patients were randomised on a 1 : 1 basis between tonsil dissection and conservative management (i.e. deferred surgery) using a variable block-stratified design, stratified by (1) centre and (2) severity. Main outcome measures: The primary outcome measure was the total number of sore throat days over 24 months following randomisation. The secondary outcome measures were the number of sore throat episodes and five characteristics from Sore Throat Alert Return, describing severity of the sore throat, use of medications, time away from usual activities and the Short Form questionnaire-12 items. Additional secondary outcomes were the Tonsil Outcome Inventory-14 total and subscales and Short Form questionnaire-12 items 6 monthly. Evaluation of the impact of alternative sore throat patient pathways by observation and statistical modelling of outcomes against baseline severity, as assessed by Tonsil Outcome Inventory-14 score at recruitment. The incremental cost per sore throat day avoided, the incremental cost per quality-adjusted life-year gained based on responses to the Short Form questionnaire-12 items and the incremental net benefit based on costs and responses to a contingent valuation exercise. A qualitative process evaluation examined acceptability of trial processes and ramdomised arms. Results: There was a median of 27 (interquartile range 12–52) sore throats over the 24-month follow-up. A smaller number of sore throats was reported in the tonsillectomy arm [median 23 (interquartile range 11–46)] than in the conservative management arm [median 30 (interquartile range 14–65)]. On an intention-to-treat basis, there were fewer sore throats in the tonsillectomy arm (incident rate ratio 0.53, 95% confidence interval 0.43 to 0.65). Sensitivity analyses confirmed this, as did the secondary outcomes. There were 52 episodes of post-operative haemorrhage reported in 231 participants undergoing tonsillectomy (22.5%). There were 47 re-admissions following tonsillectomy (20.3%), 35 relating to haemorrhage. On average, tonsillectomy was more costly and more effective in terms of both sore throat days avoided and quality-adjusted life-years gained. Tonsillectomy had a 100% probability of being considered cost-effective if the threshold for an additional quality-adjusted life year was £20,000. Tonsillectomy had a 69% probability of having a higher net benefit than conservative management. Trial processes were deemed to be acceptable. Patients who received surgery were unanimous in reporting to be happy to have received it. Limitations: The decliners who provided data tended to have higher Tonsillectomy Outcome Inventory-14 scores than those willing to be randomised implying that patients with a higher burden of tonsillitis symptoms may have declined entry into the trial. Conclusions: The tonsillectomy arm had fewer sore throat days over 24 months than the conservative management arm, and had a high probability of being considered cost-effective over the ranges considered. Further work should focus on when tonsillectomy should be offered. National Trial of Tonsillectomy IN Adults has assessed the effectiveness of tonsillectomy when offered for the current UK threshold of disease burden. Further research is required to define the minimum disease burden at which tonsillectomy becomes clinically effective and cost-effective. Trial registration: This trial is registered as ISRCTN55284102. Funding: This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 12/146/06) and is published in full in Health Technology Assessment; Vol. 27, No. 31. See the NIHR Funding and Awards website for further award information.

Objective To assess the use of proton pump inhibitors (PPIs) to treat persistent throat symptoms. Design Pragmatic, double blind, placebo controlled, randomised trial. Setting Eight ear, nose, and throat outpatient clinics, United Kingdom. Participants 346 patients aged 18 years or older with persistent throat symptoms who were randomised according to recruiting centre and baseline severity of symptoms (mild or severe): 172 to lansoprazole and 174 to placebo. Intervention Random blinded allocation (1:1) to either 30 mg lansoprazole twice daily or matched placebo twice daily for 16 weeks. Main outcome measures Primary outcome was symptomatic response at 16 weeks measured using the total reflux symptom index (RSI) score. Secondary outcomes included symptom response at 12 months, quality of life, and throat appearances. Results Of 1427 patients initially screened for eligibility, 346 were recruited. The mean age of the study sample was 52.2 (SD 13.7) years, 196 (57%) were women, and 162 (47%) had severe symptoms at presentation; these characteristics were balanced across treatment arms. The primary analysis was performed on 220 patients who completed the primary outcome measure within a window of 14-20 weeks. Mean RSI scores were similar between treatment arms at baseline: lansoprazole 22.0 (95% confidence interval 20.4 to 23.6) and placebo 21.7 (20.5 to 23.0). Improvements (reduction in RSI score) were observed in both groups—score at 16 weeks: lansoprazole 17.4 (15.5 to19.4) and placebo 15.6 (13.8 to 17.3). No statistically significant difference was found between the treatment arms: estimated difference 1.9 points (95% confidence interval −0.3 to 4.2 points; P=0.096) adjusted for site and baseline symptom severity. Lansoprazole showed no benefits over placebo for any secondary outcome measure, including RSI scores at 12 months: lansoprazole 16.0 (13.6 to 18.4) and placebo 13.6 (11.7 to 15.5): estimated difference 2.4 points (−0.6 to 5.4 points). Conclusions No evidence was found of benefit from PPI treatment in patients with persistent throat symptoms. RSI scores were similar between the lansoprazole and placebo groups after 16 weeks of treatment and at the 12 month follow-up. Trial registration ISRCTN Registry ISRCTN38578686 and EudraCT 2013-004249-17.

Background: Persistent throat symptoms are commonly attributed to ‘laryngopharyngeal reflux’. Despite a limited evidence base, these symptoms are increasingly being treated in primary care with proton pump inhibitors. Objective: To assess the value of proton pump inhibitor therapy in patients with persistent throat symptoms. Design: This was a double-blind, placebo-controlled, randomised Phase III trial. Setting: This was a multicentre UK trial in eight UK ear, nose and throat departments. Participants: A total of 346 participants aged ≥ 18 years with persistent throat symptoms and a Reflux Symptom Index score of ≥ 10, exclusive of the dyspepsia item, were recruited. Intervention: Random allocation (1 : 1 ratio) to either 30 mg of lansoprazole twice daily or matched placebo for 16 weeks. Main outcome measure: Symptomatic response (i.e. total Reflux Symptom Index score after 16 weeks of therapy). Results: A total of 1427 patients were screened and 346 were randomised. The mean age was 52 years (standard deviation 13.7 years, range 20–84 years); 150 (43%) participants were male and 196 (57%) were female; 184 (53%) participants had a mild Reflux Symptom Index minus the heartburn/dyspepsia item and 162 (47%) had a severe Reflux Symptom Index minus the heartburn/dyspepsia item. A total of 172 patients were randomised to lansoprazole and 174 were randomised to placebo. Main outcomes: A total of 267 participants completed the primary end-point visit (lansoprazole, n = 127; placebo, n = 140), of whom 220 did so between 14 and 20 weeks post randomisation (‘compliant’ group); 102 received lansoprazole and 118 received placebo. The mean Reflux Symptom Index scores at baseline were similar [lansoprazole 22.0 (standard deviation 8.0), placebo 21.7 (standard deviation 7.1), overall 21.9 (standard deviation 7.5)]. The mean Reflux Symptom Index scores at 16 weeks reduced from baseline in both groups [overall 17.4 (standard deviation 9.9), lansoprazole 17.4 (standard deviation 9.9), placebo 15.6 (standard deviation 9.8)]. Lansoprazole participants had estimated Reflux Symptom Index scores at 16 weeks that were 1.9 points higher (worse) than those of placebo participants (95% confidence interval –0.3 to 4.2; p _adj = 0.096), adjusted for site and baseline severity. Secondary outcomes: Ninety-five (43%) participants achieved a Reflux Symptom Index score in the normal range (< 12) at 16 weeks: 42 (41%) in the lansoprazole group and 53 (45%) in the placebo group. A total of 226 participants completed the end-of-trial follow-up visit (lansoprazole, n = 109; placebo, n = 117), of whom 181 were ‘compliant’. The mean Reflux Symptom Index scores at 12 months reduced from baseline in both groups [lansoprazole 16.0 (standard deviation 10.8), placebo 13.6 (standard deviation 9.6), overall 14.7 (standard deviation 10.2)]. A total of 87 (48%) participants achieved a Reflux Symptom Index score in the normal range at 12 months: 33 (40%) in the lansoprazole group and 54 (55%) in the placebo group. Likewise, the Comprehensive Reflux Symptom Score and Laryngopharyngeal Reflux – Health Related Quality of Life total scores and subscales all showed very similar changes in the lansoprazole and placebo cohorts at both 16 weeks and 12 months. Limitations: Drop-out rate and compliance are issues in pragmatic clinical trials. The Trial Of Proton Pump Inhibitors in Throat Symptoms (TOPPITS) aimed to detect clinically relevant difference with 90% power. The 346 randomised participants reduced to 283 at the primary end point; 267 completed the primary outcome measure, 220 within the protocol time scale. Despite this, the powers to detect the clinically relevant difference in Reflux Symptom Index score at 16 weeks were 82% (compliant comparison) and 89% (pragmatic comparison). The lack of difference between lansoprazole and placebo is generalisable across NHS clinics. Conclusions: Participants on lansoprazole did not report significantly better outcomes than participants on placebo on any of the three patient-reported outcome tools (Reflux Symptom Index, Comprehensive Reflux Symptom Score and Laryngopharyngeal Reflux – Health Related Quality of Life). This multicentre, pragmatic, powered, definitive Phase III trial found no evidence of benefit for patients by treating persistent throat symptoms with lansoprazole. Trial registration: Current Controlled Trials ISRCTN38578686 and EudraCT number 2013-004249-17. Funding: This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 25, No. 3. See the NIHR Journals Library website for further project information.

Young people in state care, often due to abuse or neglect, have a four-fold increased risk of drug and alcohol use compared to their peers. The SOLID study aimed to investigate the feasibility of a definitive randomised controlled trial, comparing two behaviour change interventions to reduce risky substance use (illicit drugs and alcohol), and improve mental health, in young people in care. We recruited young people in care aged 12–20 years, self-reporting substance use within the previous 12 months and residing in 1 of 6 participating local authority sites in the North East of England. Participants were randomised to either i. Motivational Enhancement Therapy (MET), ii. Social Behaviour and Network Therapy (SBNT) or iii. Control (usual care). All interventions were delivered by trained drug and alcohol workers. Follow-up data were collected 12 months post recruitment. Feasibility for trial progression was compared to pre-specified stop: go criteria (recruitment of 60% of eligible participants, 80% of participants attending 60% of offered sessions and retention of 70% of participants at 12 month follow up). Of 1450 eligible participants, 860 (59%) were screened for drug and alcohol use by social workers, 211 (24.5%) met inclusion criteria for the trial and 112 young people (7.7%) consented and were randomised. Sixty of these 112 participants (54%) completed 12-month follow-up questionnaires. Only 15 out of the 76 (20%) participants allocated to an intervention arm attended any of the offered MET or SBNT sessions. By reference to pre-specified stop: go criteria it is not feasible to conduct a definitive trial for SOLID in its current format. Despite co-designing procedures with staff and young people in care, the screening, referral and treatment pathway did not work here. Future work may require dedicated clinically embedded research resource to evaluate effectiveness of new interventions in services.

Background: Looked-after children and care leavers (henceforth children in care) are young people placed under the care of local authorities, often because of a history of family abuse and/or neglect. These young people have significantly increased risk of substance use and mental health problems compared with peers. Aim: The Supporting Looked After Children and Care Leavers In Decreasing Drugs, and alcohol (SOLID) trial aimed to investigate the feasibility of a definitive randomised controlled trial comparing two behaviour change interventions to reduce risky substance use (illicit drugs and alcohol) in and improve the mental health of children in care aged 12–20 years. Methods: The study consisted of two phases: (1) a formative phase that adapted the motivational enhancement therapy and social behaviour and network therapy interventions for use with children in care and (2) a three-arm pilot randomised controlled trial (comparing motivational enhancement therapy, social behaviour and network therapy and usual care), and a linked process and economic (return on investment) evaluation. Trial feasibility was compared with prespecified STOP/GO criteria. Setting: Six local authority areas in the north-east of England. Participants: Children in care (aged 12–20 years) who screened positive for drug and/or alcohol use within the last 12 months were eligible for the trial. The formative and process evaluations included children in care, carers, social workers, and drug and alcohol workers. Outcome measures: The primary outcomes were recruitment and retention rates at 12 months’ follow-up. Baseline and 12-month follow-up questionnaires measured self-reported drug and alcohol use, mental health and health-related quality of life. The process evaluation considered acceptability and engagement with the interventions and trial procedures. Results: Formative findings (n = 65) highlighted the need for interventions to increase the emphasis on therapeutic relationships, use creative methods of engagement and support the identification of treatment goals wider than substance misuse. Within the randomised controlled trial, of 860 participants screened, 211 (24.5%) met the inclusion criteria. One hundred and twelve (53%) of the 211 eligible children were recruited and randomised. Just 15 of the 76 (20%) participants allocated to intervention attended any of the motivational enhancement therapy of social behaviour and network therapy sessions, and 60 (54%) participants completed the 12-month follow-up. The screening and recruitment of children in care required significantly more time and resource investment by researchers and children’s services than planned. The process evaluation (n = 116) demonstrated that, despite participants engaging in risky substance use, they did not often acknowledge this nor felt that they needed help. Children in care had complex, chaotic lives and children’s services departments were less research mature and extremely stretched; this, coupled with the multiple steps in the intervention pathway and study protocol, resulted in low adherence to the intervention and the trial. Conclusions: The SOLID trial demonstrated successful engagement with children in care to adapt the motivational enhancement therapy and social behaviour and network therapy interventions. However, the pilot randomised controlled trial found that a definitive trial is not feasible. The current screen, refer and treat pathway for children in care did not work. There is an urgent need to radically rethink how we deliver therapeutic services for children in care. A pragmatic evaluation design, coupled with additional research resource for children’s services, is needed to evaluate these novel models of care at scale. Trial registration: This study is registered as PROSPERO CRD42018098974 and Current Controlled Trials ISRCTN80786829. Funding: This project was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 8, No. 13. See the NIHR Journals Library website for further project information.

Background: The visually impaired have a higher risk of falling and are likely to avoid activity. Objectives: To adapt the existing Falls Management Exercise (FaME) programme, which is delivered in the community, for visually impaired older people (VIOP) and to investigate the feasibility of conducting a definitive randomised controlled trial of this adapted intervention. Design: Phase I – consultation with stakeholders to adapt the existing programme. Two focus groups were conducted, each with 10 VIOP across the study sites. Phase II – two-centre randomised pilot trial and economic evaluation of the adapted programme for VIOP versus usual care. Phases III and IV – qualitative interviews with VIOP and Postural Stability Instructors regarding their views and experiences of the research process, undertaking the intervention and its acceptability. Intervention: This was adapted from the group-based FaME programme. A 1-hour exercise programme ran weekly over 12 weeks at the study sites (Newcastle upon Tyne and Glasgow) and was delivered by third-sector organisations. Participants were advised to also exercise at home for 2 hours per week. Those randomised to the usual activities group received no intervention. Outcome measures: These were completed at baseline, week 12 and week 24. The primary potential outcome measure used was the Short Form Falls Efficacy Scale – International. Secondary outcome assessment measures were activity avoidance, current activity, balance/falls risk, physical activity, loneliness, anxiety and depression, work and social adjustment, quality of life and economic costs. Participants’ compliance was assessed by reviewing attendance records and self-reported compliance with the home exercises. Instructors’ compliance with the course content (fidelity) was assessed by a researcher attending a sample of exercise sessions. Adverse events were collected in a weekly telephone call for all participants in both the intervention and control arm. Findings: An adapted exercise programme was devised with stakeholders. In the pilot trial, 82 participants drawn from community-living VIOP were screened, 68 met the inclusion criteria and 64 were randomised, with 33 allocated to the intervention and 31 to the usual activities arm. A total of 94% of participants provided data at week 12 and 92% at week 24. Adherence to the study was high. The intervention was found to be both safe and acceptable to participants, with 76% attending nine or more classes. Median time for home exercise was 50 minutes per week. There was little or no evidence that fear of falling, exercise, attitudinal or quality-of-life outcomes differed between trial arms at follow-up. Thematic analysis of the interviews with VIOP participants identified facilitators of and barriers to exercise, including perceived relevance to health, well-being and lifestyle, social interaction, self-perception and practical assistance. Instructors identified issues regarding level of challenge and assistance from a second person. Limitations: The small sample size and low falls risk of the study sample are study limitations. Conclusion: Although adaptation, recruitment and delivery were successful, the findings (particularly from qualitative research with instructors and participants) indicated that VIOP with low to moderate falls risk could be integrated into mainstream programmes with some adaptations. A future definitive trial should consider graduated exercises appropriate to ability and falls risk within mainstream provision. Other outcome measures may additionally be considered. Trial registration: Current Controlled Trials ISRCTN16949845. Funding: This project was funded by the NIHR Public Health Research programme and will be published in full in Public Health Research; Vol. 7, No. 4. See the NIHR Journals Library website for further project information.