Abstract: Depression in older adults is common and is associated with poor quality of life, increased morbidity and early mortality, and increased health and social care use. Collaborative care, a low-intensity intervention for depression that is shown to be effective in working-age adults, has not yet been evaluated in older people with depression who are managed in UK primary care. The CollAborative care for Screen-Positive EldeRs (CASPER) plus trial fills the evidence gap identified by the most recent guidelines on depression management. Objectives To establish the clinical effectiveness and cost-effectiveness of collaborative care for older adults with major depressive disorder in primary care. Design A pragmatic, multicentred, two-arm, parallel, individually randomised controlled trial with embedded qualitative study. Participants were automatically randomised by computer, by the York Trials Unit Randomisation Service, on a 1 : 1 basis using simple unstratified randomisation after informed consent and baseline measures were collected. Blinding was not possible. Setting Sixty-nine general practices in the north of England. Participants A total of 485 participants aged ≥ 65 years with major depressive disorder. Interventions A low-intensity intervention of collaborative care, including behavioural activation, delivered by a case manager for an average of six sessions over 7–8 weeks, alongside usual general practitioner (GP) care. The control arm received only usual GP care. Main outcome measures The primary outcome measure was Patient Health Questionnaire-9 items score at 4 months post randomisation. Secondary outcome measures included depression severity and caseness at 12 and 18 months, the EuroQol-5 Dimensions, Short Form questionnaire-12 items, Patient Health Questionnaire-15 items, Generalised Anxiety Disorder-7 items, Connor–Davidson Resilience Scale-2 items, a medication questionnaire, objective data and adverse events. Participants were followed up at 12 and 18 months. Results In total, 485 participants were randomised (collaborative care, n  = 249; usual care, n  = 236), with 390 participants (80%: collaborative care, 75%; usual care, 86%) followed up at 4 months, 358 participants (74%: collaborative care, 70%; usual care, 78%) followed up at 12 months and 344 participants (71%: collaborative care, 67%; usual care, 75%) followed up at 18 months. A total of 415 participants were included in primary analysis (collaborative care, n  = 198; usual care, n  = 217), which revealed a statistically significant effect in favour of collaborative care at the primary end point at 4 months [8.98 vs. 10.90 score points, mean difference 1.92 score points, 95% confidence interval (CI) 0.85 to 2.99 score points; p   〈  0.001], equivalent to a standard effect size of 0.34. However, treatment differences were not maintained in the longer term (at 12 months: 0.19 score points, 95% CI –0.92 to 1.29 score points; p  = 0.741; at 18 months: 〈  0.01 score points, 95% CI –1.12 to 1.12 score points; p  = 0.997). The study recorded details of all serious adverse events (SAEs), which consisted of ‘unscheduled hospitalisation’, ‘other medically important condition’ and ‘death’. No SAEs were related to the intervention. Collaborative care showed a small but non-significant increase in quality-adjusted life-years (QALYs) over the 18-month period, with a higher cost. Overall, the mean cost per incremental QALY for collaborative care compared with usual care was £26,016; however, for participants attending six or more sessions, collaborative care appears to represent better value for money (£9876/QALY). Limitations Study limitations are identified at different stages: design (blinding unfeasible, potential contamination), process (relatively low overall consent rate, differential attrition/retention rates) and analysis (no baseline health-care resource cost or secondary/social care data). Conclusion Collaborative care was effective for older people with case-level depression across a range of outcomes in the short term though the reduction in depression severity was not maintained over the longer term of 12 or 18 months. Participants who received six or more sessions of collaborative care did benefit substantially more than those who received fewer treatment sessions but this difference was not statistically significant. Future work recommendations Recommendations for future research include investigating the longer-term effect of the intervention. Depression is a recurrent disorder and it would be useful to assess its impact on relapse and the prevention of future case-level depression. Trial registration Current Controlled Trials ISRCTN45842879. 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. 21, No. 67. See the NIHR Journals Library website for further project information.

Abstract: Progesterone is essential to maintain a healthy pregnancy. Guidance from the Royal College of Obstetricians and Gynaecologists and a Cochrane review called for a definitive trial to test whether or not progesterone therapy in the first trimester could reduce the risk of miscarriage in women with a history of unexplained recurrent miscarriage (RM). The PROMISE trial was conducted to answer this question. A concurrent cost-effectiveness analysis was conducted. Design and setting A randomised, double-blind, placebo-controlled, international multicentre study, with economic evaluation, conducted in hospital settings across the UK (36 sites) and in the Netherlands (nine sites). Participants and interventions Women with unexplained RM (three or more first-trimester losses), aged between 18 and 39 years at randomisation, conceiving naturally and giving informed consent, received either micronised progesterone (Utrogestan ® , Besins Healthcare) at a dose of 400 mg (two vaginal capsules of 200 mg) or placebo vaginal capsules twice daily, administered vaginally from soon after a positive urinary pregnancy test (and no later than 6 weeks of gestation) until 12 completed weeks of gestation (or earlier if the pregnancy ended before 12 weeks). Main outcome measures Live birth beyond 24 completed weeks of gestation (primary outcome), clinical pregnancy at 6–8 weeks, ongoing pregnancy at 12 weeks, miscarriage, gestation at delivery, neonatal survival at 28 days of life, congenital abnormalities and resource use. Methods Participants were randomised after confirmation of pregnancy. Randomisation was performed online via a secure internet facility. Data were collected on four occasions of outcome assessment after randomisation, up to 28 days after birth. Results A total of 1568 participants were screened for eligibility. Of the 836 women randomised between 2010 and 2013, 404 received progesterone and 432 received placebo. The baseline data (age, body mass index, maternal ethnicity, smoking status and parity) of the participants were comparable in the two arms of the trial. The follow-up rate to primary outcome was 826 out of 836 (98.8%). The live birth rate in the progesterone group was 65.8% (262/398) and in the placebo group it was 63.3% (271/428), giving a relative risk of 1.04 (95% confidence interval 0.94 to 1.15; p  = 0.45). There was no evidence of a significant difference between the groups for any of the secondary outcomes. Economic analysis suggested a favourable incremental cost-effectiveness ratio for decision-making but wide confidence intervals indicated a high level of uncertainty in the health benefits. Additional sensitivity analysis suggested the probability that progesterone would fall within the National Institute for Health and Care Excellence’s threshold of £20,000–30,000 per quality-adjusted life-year as between 0.7145 and 0.7341. Conclusions There is no evidence that first-trimester progesterone therapy improves outcomes in women with a history of unexplained RM. Limitations This study did not explore the effect of treatment with other progesterone preparations or treatment during the luteal phase of the menstrual cycle. Future work Future research could explore the efficacy of progesterone supplementation administered during the luteal phase of the menstrual cycle in women attempting natural conception despite a history of RM. Trial registration Current Controlled Trials ISRCTN92644181; EudraCT 2009-011208-42; Research Ethics Committee 09/H1208/44. 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. 20, No. 41. See the NIHR Journals Library website for further project information.

Abstract: Malnutrition is a common problem in critically ill patients in UK NHS critical care units. Early nutritional support is therefore recommended to address deficiencies in nutritional state and related disorders in metabolism. However, evidence is conflicting regarding the optimum route (parenteral or enteral) of delivery. Objectives To estimate the effect of early nutritional support via the parenteral route compared with the enteral route on mortality at 30 days and on incremental cost-effectiveness at 1 year. Secondary objectives were to compare the route of early nutritional support on duration of organ support; infectious and non-infectious complications; critical care unit and acute hospital length of stay; all-cause mortality at critical care unit and acute hospital discharge, at 90 days and 1 year; survival to 90 days and 1 year; nutritional and health-related quality of life, resource use and costs at 90 days and 1 year; and estimated lifetime incremental cost-effectiveness. Design A pragmatic, open, multicentre, parallel-group randomised controlled trial with an integrated economic evaluation. Setting Adult general critical care units in 33 NHS hospitals in England. Participants 2400 eligible patients. Interventions Five days of early nutritional support delivered via the parenteral ( n  = 1200) and enteral ( n  = 1200) route. Main outcome measures All-cause mortality at 30 days after randomisation and incremental net benefit (INB) (at £20,000 per quality-adjusted life-year) at 1 year. Results By 30 days, 393 of 1188 (33.1%) patients assigned to receive early nutritional support via the parenteral route and 409 of 1195 (34.2%) assigned to the enteral route had died [ p  = 0.57; absolute risk reduction 1.15%, 95% confidence interval (CI) −2.65 to 4.94; relative risk 0.97 (0.86 to 1.08)]. At 1 year, INB for the parenteral route compared with the enteral route was negative at −£1320 (95% CI −£3709 to £1069). The probability that early nutritional support via the parenteral route is more cost-effective – given the data – is 〈  20%. The proportion of patients in the parenteral group who experienced episodes of hypoglycaemia ( p  = 0.006) and of vomiting ( p   〈  0.001) was significantly lower than in the enteral group. There were no significant differences in the 15 other secondary outcomes and no significant interactions with pre-specified subgroups. Limitations Blinding of nutritional support was deemed to be impractical and, although the primary outcome was objective, some secondary outcomes, although defined and objectively assessed, may have been more vulnerable to observer bias. Conclusions There was no significant difference in all-cause mortality at 30 days for early nutritional support via the parenteral route compared with the enteral route among adults admitted to critical care units in England. On average, costs were higher for the parenteral route, which, combined with similar survival and quality of life, resulted in negative INBs at 1 year. Future work Nutritional support is a complex combination of timing, dose, duration, delivery and type, all of which may affect outcomes and costs. Conflicting evidence remains regarding optimum provision to critically ill patients. There is a need to utilise rigorous consensus methods to establish future priorities for basic and clinical research in this area. Trial registration Current Controlled Trials ISRCTN17386141. Funding This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment ; Vol. 20, No. 28. See the NIHR Journals Library website for further project information.

Abstract: Up to 10% of children and young people have a specific phobia that can significantly affect their mental health, development and daily functioning. Cognitive–behavioural therapy-based interventions remain the dominant treatment, but limitations to their provision warrant investigation into low-intensity alternatives. One-session treatment is one such alternative that shares cognitive–behavioural therapy principles but has a shorter treatment period. Objective This research investigated the non-inferiority of one-session treatment to cognitive–behavioural therapy for treating specific phobias in children and young people. The acceptability and cost-effectiveness of one-session treatment were examined. Design A pragmatic, multicentre, non-inferiority randomised controlled trial, with embedded economic and qualitative evaluations. Settings There were 26 sites, including 12 NHS trusts. Participants Participants were aged 7–16 years and had a specific phobia defined in accordance with established international clinical criteria. Interventions Participants were randomised 1 : 1 to receive one-session treatment or usual-care cognitive–behavioural therapy, and were stratified according to age and phobia severity. Outcome assessors remained blind to treatment allocation. Main outcome measures The primary outcome measure was the Behavioural Avoidance Task at 6 months’ follow-up. Secondary outcomes included the Anxiety Disorder Interview Schedule, Child Anxiety Impact Scale, Revised Children’s Anxiety and Depression Scale, a goal-based outcome measure, Child Health Utility 9D, EuroQol-5 Dimensions Youth version and resource usage. Treatment fidelity was assessed using the Cognitive Behaviour Therapy Scale for Children and Young People and the One-Session Treatment Rating Scale. Results A total of 274 participants were recruited, with 268 participants randomised to one-session treatment ( n  = 134) or cognitive–behavioural therapy ( n  = 134). A total of 197 participants contributed some data, with 149 participants in the intention-to-treat analysis and 113 in the per-protocol analysis. Mean Behavioural Avoidance Task scores at 6 months were similar across treatment groups when both intention-to-treat and per-protocol analyses were applied [cognitive–behavioural therapy: 7.1 (intention to treat), 7.4 (per protocol); one-session treatment: 7.4 (intention to treat), 7.6 (per protocol); on the standardised scale adjusted mean difference for cognitive–behavioural therapy compared with one-session treatment –0.123, 95% confidence interval –0.449 to 0.202 (intention to treat), mean difference –0.204, 95% confidence interval –0.579 to 0.171 (per protocol)] . These findings were wholly below the standardised non-inferiority limit of 0.4, which suggests that one-session treatment is non-inferior to cognitive–behavioural therapy. No between-group differences in secondary outcome measures were found. The health economics evaluation suggested that, compared with cognitive–behavioural therapy, one-session treatment marginally decreased the mean service use costs and maintained similar mean quality-adjusted life-year improvement. Nested qualitative evaluation found one-session treatment to be considered acceptable by those who received it, their parents/guardians and clinicians. No adverse events occurred as a result of phobia treatment. Limitations The COVID-19 pandemic meant that 48 children and young people could not complete the primary outcome measure. Service waiting times resulted in some participants not starting therapy before follow-up. Conclusions One-session treatment for specific phobia in UK-based child mental health treatment centres is as clinically effective as multisession cognitive–behavioural therapy and highly likely to be cost-saving. Future work could involve improving the implementation of one-session treatment through training and commissioning of improved care pathways. Trial registration This trial is registered as ISRCTN19883421. Funding This project was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment ; Vol. 26, No. 42. See the NIHR Journals Library website for further project information.

Abstract: Asthma exacerbations affect the quality of life of patients with asthma and have a major effect on the overall costs of asthma care. An asthma self-management plan that advises the temporary quadrupling of inhaled corticosteroid dose may prevent asthma exacerbations, but this needs to be confirmed before being adopted widely. Objectives To compare the clinical effectiveness and cost-effectiveness of an asthma self-management plan that advises patients to temporarily quadruple the dose of inhaled corticosteroid when asthma control starts to deteriorate with a standard self-management plan. Design A multicentre, parallel-group, pragmatic randomised trial, with follow-up for 12 months. Setting Primary and secondary care across 207 sites in the UK. Participants Asthma patients aged ≥ 16 years treated with an inhaled corticosteroid who had experienced at least one exacerbation in the previous 12 months. Interventions Participants were randomised (1 : 1) to a usual-care self-management plan or to a modified self-management plan that advised a temporary quadrupling of the inhaled corticosteroid at the point of asthma deterioration, both of which were actively implemented and supported by local research staff. Primary outcome The primary outcome of ‘time to first asthma exacerbation’ was defined as the need for systemic corticosteroids (for at least 3 consecutive days) and/or unscheduled health-care consultations for asthma (i.e. reaching zone 3 or 4 of the Asthma UK self-management plan). Results A total of 1922 participants were randomised: the primary analysis included 938 participants (97%) in the usual-care group and 933 participants (97%) in the modified self-management group. The number of participants having at least one exacerbation of asthma in the year after randomisation was 484 (51.6%) in the usual-care group and 420 (45.0%) in the modified self-management group [adjusted hazard ratio 0.81, 95% confidence interval (CI) 0.71 to 0.92; p  = 0.002]. There were fewer serious adverse events reported in the modified self-management group than in the usual-care group (11 vs. 32, respectively). Eight and six events of pneumonia, lower respiratory tract infections or influenza were reported in the usual-care group and the modified self-management group, respectively. Health-care-related costs were lower in the modified self-management group. The modified self-management group was £24 (bootstrapped 95% CI –£122 to £71) less costly than usual care, with a greater quality-adjusted life-year gain of 0.02 (bootstrapped 95% CI –0.005 to 0.04). Therefore, the modified self-management group was ‘dominant’, with a 94–95% probability of being cost-effective at the £20,000–30,000 threshold. Limitations As the Fourfold Asthma STudy (FAST) was an open-label pragmatic trial, the possibility of treatment bias that may have affected the participants in the modified self-management group cannot be ruled out. Poorer than expected completion of participant diary cards, particularly within the usual-care self-management group, could have led to a null bias, underestimating the true effect of the intervention. Conclusions An asthma self-management plan that advises patients to temporarily quadruple their dose of inhaled corticosteroid at the point of asthma symptoms worsening does reduce clinically important asthma exacerbations. In addition, the plan is cost-effective compared with the usual-care self-management plan. Future work To effectively implement asthma self-management plans that advise a temporary quadrupling of inhaled steroid at asthma deterioration into routine practice. Trial registration Current Controlled Trials ISRCTN15441965. Funding This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment ; Vol. 22, No. 70. See the NIHR Journals Library website for further project information.

Abstract: In the UK, 6% of the UK population have diabetes mellitus, 90% of whom have type 2 diabetes mellitus (T2DM). Diabetes mellitus accounts for 10% of NHS expenditure (£14B annually). Good self-management may improve health outcomes. NHS policy is to refer all people with T2DM to structured education, on diagnosis, to improve their self-management skills, with annual reinforcement thereafter. However, uptake remains low (5.6% in 2014–15). Almost all structured education is group based, which may not suit people who work, who have family or other caring commitments or who simply do not like group-based formats. Moreover, patient needs vary with time and a single education session at diagnosis is unlikely to meet these evolving needs. A web-based programme may increase uptake. Objectives Our aim was to develop, evaluate and implement a web-based self-management programme for people with T2DM at any stage of their illness journey, with the goal of improving access to, and uptake of, self-management support, thereby improving health outcomes in a cost-effective manner. Specific objectives were to (1) develop an evidence-based theoretically informed programme that was acceptable to patients and health-care professionals (HCPs) and that could be readily implemented within routine NHS care, (2) determine the clinical effectiveness and cost-effectiveness of the programme compared with usual care and (3) determine how best to integrate the programme into routine care. Design There were five linked work packages (WPs). WP A determined patient requirements and WP B determined HCP requirements for the self-management programme. WP C developed and user-tested the Healthy Living for People with type 2 Diabetes (HeLP-Diabetes) programme. WP D was an individually randomised controlled trial in primary care with a health economic analysis. WP E used a mixed-methods and case-study design to study the potential for implementing the HeLP-Diabetes programme within routine NHS practice. Setting English primary care. Participants People with T2DM (WPs A, D and E) or HCPs caring for people with T2DM (WPs B, C and E). Intervention The HeLP-Diabetes programme; an evidence-based theoretically informed web-based self-management programme for people with T2DM at all stages of their illness journey, developed using participatory design principles. Main outcome measures WPs A and B provided data on user ‘wants and needs’, including factors that would improve the uptake and accessibility of the HeLP-Diabetes programme. The outcome for WP C was the HeLP-Diabetes programme itself. The trial (WP D) had two outcomes measures: glycated haemoglobin (HbA 1c ) level and diabetes mellitus-related distress, as measured with the Problem Areas in Diabetes (PAID) scale. The implementation outcomes (WP E) were the adoption and uptake at clinical commissioning group, general practice and patient levels and the identification of key barriers and facilitators. Results Data from WPs A and B supported our holistic approach and addressed all areas of self-management (medical, emotional and role management). HCPs voiced concerns about linkage with the electronic medical records (EMRs) and supporting patients to use the programme. The HeLP-Diabetes programme was developed and user-tested in WP C. The trial (WP D) recruited to target ( n  = 374), achieved follow-up rates of over 80% and the intention-to-treat analysis showed that there was an additional improvement in HbA 1c levels at 12 months in the intervention group [mean difference –0.24%, 95% confidence interval (CI) –0.44% to –0.049%]. There was no difference in overall PAID score levels (mean difference –1.5 points, 95% CI –3.9 to 0.9 points). The within-trial health economic analysis found that incremental costs were lower in the intervention group than in the control group (mean difference –£111, 95% CI –£384 to £136) and the quality-adjusted life-years (QALYs) were higher (mean difference 0.02 QALYs, 95% CI 0.000 to 0.044 QALYs), meaning that the HeLP-Diabetes programme group dominated the control group. In WP E, we found that the HeLP-Diabetes programme could be successfully implemented in primary care. General practices that supported people in registering for the HeLP-Diabetes programme had better uptake and registered patients from a wider demographic than those relying on patient self-registration. Some HCPs were reluctant to do this, as they did not see it as part of their professional role. Limitations We were unable to link the HeLP-Diabetes programme with the EMRs or to determine the effects of the HeLP-Diabetes programme on users in the implementation study. Conclusions The HeLP-Diabetes programme is an effective self-management support programme that is implementable in primary care. Future work The HeLP-Diabetes research team will explore the following in future work: research to determine how to improve patient uptake of self-management support; develop and evaluate a structured digital educational pathway for newly diagnosed people; develop and evaluate a digital T2DM prevention programme; and the national implementation of the HeLP-Diabetes programme. Trial registration Research Ethics Committee reference number 10/H0722/86 for WPs A–C; Research Ethics Committee reference number 12/LO/1571 and UK Clinical Research Network/National Institute for Health Research (NIHR) Portfolio 13563 for WP D; and Research Ethics Committee 13/EM/0033 for WP E. In addition, for WP D, the study was registered with the International Standard Randomised Controlled Trial Register as reference number ISRCTN02123133. Funding details This project was funded by the NIHR Programme Grants for Applied Research programme and will be published in full in Programme Grants for Applied Research ; Vol. 6, No. 5. See the NIHR Journals Library website for further project information.

Abstract: Social skills interventions are commonly recommended to help children and young people with autism spectrum disorder develop social skills, but some struggle to engage in these interventions. LEGO ® (LEGO System A/S, Billund, Denmark) based therapy, a group social skills intervention, aims to be more interesting and engaging. Objective To evaluate the clinical effectiveness of LEGO ® based therapy on the social and emotional skills of children and young people with autism spectrum disorder in school settings compared with usual support. Secondary objectives included evaluations of cost-effectiveness, acceptability and treatment fidelity. Design A cluster randomised controlled trial randomly allocating participating schools to either LEGO ® based therapy and usual support or usual support only. Setting Mainstream schools in the north of England. Participants Children and young people (aged 7–15 years) with autism spectrum disorder, their parent/guardian, an associated teacher/teaching assistant and a facilitator teacher/teaching assistant (intervention schools only). Intervention Schools randomised to the intervention arm delivered 12 weekly sessions of LEGO ® based therapy, which promotes collaborative play and encourages social problem-solving in groups of three children and young people with a facilitator (trained teacher or teaching assistant). Participants received usual support from school and community services. Participants in the control arm received usual support only. Research assistants and statisticians were blind to treatment allocation. Main outcome measure The social skills subscale of the Social Skills Improvement System (SSIS), completed by the children and young people’s unblinded teacher pre randomisation and 20 weeks post randomisation. The SSIS social skills subscale measures social skills including social communication, co-operation, empathy, assertion, responsibility and self-control. Participants completed a number of other pre- and post-randomisation measures of emotional health, quality of life, loneliness, problem behaviours, academic competence, service resource utilisation and adverse events. Results A total of 250 children and young people from 98 schools were randomised: 127 to the intervention arm and 123 to the control arm. Intention-to-treat analysis of the main outcome measure showed a modest positive difference of 3.74 points (95% confidence interval –0.16 to 7.63 points, standardised effect size 0.18; p  = 0.06) in favour of the intervention arm. Statistical significance increased in per-protocol analysis, with a modest positive difference (standardised effect size 0.21; p  = 0.036). Cost-effectiveness of the intervention was found in reduced service use costs and a small increase in quality-adjusted life-years. Intervention fidelity and acceptability were positive. No intervention-related adverse events or effects were reported. Conclusions The primary and pre-planned sensitivity analysis of the primary outcome consistently showed a positive clinical difference, with modest standardised effect sizes of between 0.15 and 0.21. There were positive health economics and qualitative findings, corroborated by the difference between arms for the majority of secondary outcomes, which were not statistically significant but favoured the intervention arm. Post hoc additional analysis was exploratory and was not used in drawing this conclusion. Given the small positive change, LEGO ® based therapy for children and young people with autism spectrum disorder in schools should be considered. Limitations The primary outcome measure was completed by an unblinded teacher (rather than by the facilitator). Future work The study team recommends future research into LEGO ® based therapy, particularly in school environments. Trial registration This trial is registered as ISRCTN64852382. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Public Health Research programme (NIHR award ref: 15/49/32) and is published in full in Public Health Research ; Vol. 11, No. 12. See the NIHR Funding and Awards website for further award information.

Abstract: There is a high prevalence of smoking among people with severe mental ill health (SMI). Helping people with SMI to quit smoking could improve their health and longevity, and reduce health inequalities. However, those with SMI are less likely to access and engage with routine smoking cessation services than the general population. Objectives To compare the clinical effectiveness and cost-effectiveness of a bespoke smoking cessation (BSC) intervention with usual stop smoking services for people with SMI. Design A pragmatic, two-arm, individually randomised controlled trial. Setting Primary care and secondary care mental health services in England. Participants Smokers aged ≥ 18 years with SMI who would like to cut down on or quit smoking. Interventions A BSC intervention delivered by mental health specialists trained to deliver evidence-supported smoking cessation interventions compared with usual care. Main outcome measures The primary outcome was self-reported, CO-verified smoking cessation at 12 months. Smoking-related secondary outcomes were self-reported smoking cessation, the number of cigarettes smoked per day, the Fagerström Test for Nicotine Dependence and the Motivation to Quit questionnaire. Other secondary outcomes were Patient Health Questionnaire-9 items, Generalised Anxiety Disorder Assessment-7 items and 12-Item Short-Form Health Survey, to assess mental health and body mass index measured at 6 and 12 months post randomisation. Results The trial randomised 526 people (265 to the intervention group, 261 to the usual-care group) aged 19 to 72 years (mean 46 years). About 60% of participants were male. Participants smoked between 3 and 100 cigarettes per day (mean 25 cigarettes per day) at baseline. The intervention group had a higher rate of exhaled CO-verified smoking cessation at 6 and 12 months than the usual-care group [adjusted odds ratio (OR) 12 months: 1.6, 95% confidence interval (CI) 0.9 to 2.8; adjusted OR 6 months: 2.4, 95% CI 1.2 to 4.7]. This was not statistically significant at 12 months ( p  = 0.12) but was statistically significant at 6 months ( p  = 0.01). In total, 111 serious adverse events were reported (69 in the BSC group and 42 in the usual-care group); the majority were unplanned hospitalisations due to a deterioration in mental health ( n  = 98). The intervention is likely (57%) to be less costly but more effective than usual care; however, this result was not necessarily associated with participants’ smoking status. Limitations Follow-up was not blind to treatment allocation. However, the primary outcome included a biochemically verified end point, less susceptible to observer biases. Some participants experienced difficulties in accessing nicotine replacement therapy because of changes in service provision. Efforts were made to help participants access nicotine replacement therapy, but this may have affected participants’ quit attempt. Conclusions People with SMI who received the intervention were more likely to have stopped smoking at 6 months. Although more people who received the intervention had stopped smoking at 12 months, this was not statistically significant. Future work Further research is needed to establish how quitting can be sustained among people with SMI. Trial registration Current Controlled Trials ISRCTN72955454. 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. 23, No. 50. See the NIHR Journals Library website for further project information.