Abstract One-to-one peer support is widely used in mental health services, but the components associated with better outcomes remain unclear. We systematically reviewed randomised controlled trials and conducted additive component network meta-analyses to identify which components of one-to-one peer support worker interventions were associated with outcomes for adults using mental health services. CINAHL Ultimate, Embase, MEDLINE, PsycINFO, CENTRAL, ClinicalTrials.gov and ISRCTN were searched, supplemented by citation tracking, previous reviews and expert consultation. Interventions were coded for seven components: Training and development, Maintaining peer support worker wellbeing, Relationship-building, Social support, Emotional support, Practical support and Cultural adaptation. The review followed PRISMA-NMA reporting guidance and was registered with PROSPERO (CRD42022355291). Thirty-six trials randomised 6,645 participants across nine countries. Only quality of life and recovery yielded estimable component effects at one or more follow-up points. For quality of life, Practical support had a positive incremental estimate at 3 months (standardised mean difference 0.52, 95% confidence interval 0.17 to 0.87); no component showed clear evidence of benefit at 6 months; and at 12 months Social support had a positive estimate (1.57, 0.12 to 3.01), whereas Maintaining peer support worker wellbeing had a negative estimate (-1.66,-3.05 to-0.28). These estimates were not consistent across follow-up points. For recovery, Relationship-building had positive estimates at 6 months (0.90, 0.03 to 1.78) and 12 months (0.50, 0.29 to 0.72). Networks were sparse and often disconnected, and additivity could not be tested in disconnected networks. Current trials do not permit definitive prioritisation of peer-support components. Relationship-building was the most consistent candidate component, but all findings remain provisional. Future trials should prospectively specify, manipulate and measure component delivery.

BACKGROUND:Pain arising after surgery and persisting beyond 3 months is termed 'chronic postsurgical pain' (CPSP). Prevention of CPSP is a research priority. Our objectives were to assess whether regional anaesthesia can reduce the incidence of CPSP and long-term opioid use and to determine which factors influence its effectiveness. METHODS:We conducted a Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA)-compliant systematic review and network meta-analysis of randomised controlled trials assessing the effectiveness of regional anaesthesia for reduction of CPSP. We searched various databases up to October 2025. We performed pairwise meta-analysis and network meta-analysis to determine whether type, timing and method (catheter or single injection) of regional anaesthesia influenced the incidence of CPSP. We investigated the effect of surgery type, baseline risk, and sex via meta-regression. We used Grading of Recommendations Assessment, Development and Evaluation (GRADE)/ Confidence In Network Meta-Analysis (CINeMA) to assess certainty of evidence. RESULTS:We included 158 randomised controlled trials involving 18 794 subjects. Overall, regional anaesthesia reduced the incidence of CPSP compared with no block (risk ratio [RR]: 0.73, 95% confidence interval [CI]: 0.67-0.80), with effects observed up to 12 months after surgery. Reductions in CPSP were also observed within specific surgery types, including mastectomy (RR: 0.69, 95% CI: 0.59-0.79), thoracotomy (RR: 0.72, 95% CI: 0.55-0.96), video-assisted thoracoscopic surgery (RR: 0.73, 95% CI: 0.56-0.96), and knee arthroplasty (RR: 0.71, 95% CI: 0.52-0.97). Opioid use was not statistically significantly different between groups (RR: 0.88, 95% CI: 0.61-1.28). CPSP evidence was low certainty; opioid use was very low certainty. Network meta-analysis suggested that neuraxial techniques reduced the incidence of CPSP more effectively than peripheral techniques after thoracotomy (neuraxial RR: 0.64 95% CI: 0.49-0.83; peripheral RR: 0.84, 95% CI 0.73-0.96) and video-assisted thoracoscopic surgery (neuraxial RR: 0.60, 95% CI: 0.45-0.80; peripheral RR: 0.77, 95% CI: 0.63-0.94), compared with no block. Differences in effectiveness based on administration time point and method of administration varied by surgery type. Network meta-analysis evidence was low to very low certainty. Meta-regression showed no significant effect of surgery type, sex or baseline risk. CONCLUSIONS:Our findings indicate that regional anaesthesia reduces the incidence of CPSP for up to 12 months after surgery. The effectiveness of regional anaesthesia was influenced by the type of regional technique used, whereas the type of surgery, sex, or baseline risk did not have a significant impact.

OBJECTIVES:Living systematic reviews (LSRs) are an emerging type of review that continuously updates as new evidence becomes available. A previous methodological survey conducted in 2021 identified and studied all health-based LSRs. Since then, the landscape has changed, including the on-going accumulation of COVID-19 research and availability of automation tools. Furthermore, various methods and guidance exist for conducting LSRs and review authors are often encouraged to explore opportunities to maximize dissemination. We conducted an LSR survey update to describe LSRs in a "post-COVID" era. Our objectives were to summarize the uptake of LSRs, describe their characteristics, including methodological and communicative characteristics, and identify patterns in LSR attributes. STUDY DESIGN AND SETTING:We systematically searched for new LSRs and any updates-including updates from LSRs identified previously-published between May 2021 and March 2023 in any health field. Eligible articles were identified and data extracted and combined with data from the original survey. Outcomes broadly included LSR characteristics and uptake, and methodological and communicative characteristics. Analyses were descriptive and included visualizations to explore distributions, combinations, and any time effects of characteristics. RESULTS:A total of 549 records across 168 individual LSRs were identified (of which 92 LSRs were newly detected). Although the presence of COVID-19 LSRs dominated in later years, there was an increased uptake in non-COVID-19 LSRs; the former were found to search the evidence and update/publish results more frequently. Where reported, the approach to conducting updates varied considerably, including a wide range of prespecified frequencies and/or triggers. Of the 337 updates, 25.5% reported on ongoing studies, and among LSRs with published results, 58.5% used the Grading of Recommendations, Assessment, Development and Evaluation system. The proportion of LSRs with a centralized platform for sharing results was higher among (i) those that included updates, (ii) Cochrane reviews, (iii) non-COVID-19 LSRs, and (iv) funded LSRs. Few LSRs included interactive features. CONCLUSION:The number of LSRs is growing at an accelerating rate, but this survey illustrates that there are still methodological limitations and challenges that carefully need addressing. Key areas for improvement include more explicit prespecified updating strategies and better use of web-based platforms for disseminating results. PLAIN LANGUAGE SUMMARY:Every year, a huge amount of health-related research is published and it is difficult for busy doctors and health care workers to keep up to date with all of the new evidence. To help with this, the research can be summarized by carrying out a review. This is known as a "systematic review" if it is carefully carried out by two or more researchers. We use systematic reviews to get an accurate and fair summary of all the research available. However, these reviews can take a long time to create and can quickly become out-of-date. There is a new and exciting type of review called a "LSR" which is continually updated with new research evidence as it becomes available. This type of review is hugely important for "high-priority" health questions that need to be put into practice straight away, such as new treatments for COVID-19. As LSRs have become more popular since the COVID-19 pandemic, it is important to understand how LSRs are actually being performed and how well they share their results. In 2021, a group of researchers collected all the health-related LSRs that exist. In March 2023, we aimed to collect any new LSRs since then, and summarize what they look like now. Including LSRs found by the previous group, we found 168 LSRs, where 92 were found since May 2021. There now exist more COVID-19-related LSRs than non-COVID-19-related ones, but both groups are increasing in popularity. How LSRs decided when to "update" their results varied a lot. Non-COVID-19-related LSRs and those that had funding were more likely to use online tools to share their results, but very few used tools to let readers interact with the results. By looking at how LSRs are performed, we have shown that there are still challenges that need more work. Key areas for improvement include creating better ways on deciding when and how these reviews should be updated, and building tools to help researchers summarize the findings in a way that is reliable and easily understandable to everyone involved, including patients.

Network meta-analysis has emerged as a method for analysing clinical trials, with a large increase in the number of publications over the past decade. Network meta-analysis offers advantages over traditional pairwise meta-analysis, including increased power, the ability to compare treatments not compared in the original trials, and the ability to rank treatments. However, network meta-analyses are inherently more complex than pairwise meta-analyses, requiring additional statistical expertise and assumptions. Many factors can affect the certainty of evidence from pairwise meta-analysis and can often lead to unreliable results. Network meta-analysis is prone to all these issues, although it has the additional assumption of transitivity. Here we review network meta-analyses, problems with their conduct and reporting, and methodological strategies that can be used by those conducting reviews to help improve the reliability of their findings. We provide evidence that violation of the assumption of transitivity is relatively common and inadequately considered in published network meta-analyses. We explain key concepts with clinically relevant examples for those unfamiliar with network meta-analysis to facilitate their appraisal and application of their results to clinical practice.

BACKGROUND AND OBJECTIVES:The MetaInsight web application (https://apps.crsu.org.uk/MetaInsight) allows users to carry out network meta-analyses (NMAs) via a point-and-click interface, without the need for statistical programming. Network meta-regression (NMR) is an extension of NMA that adds covariates to the model and is used to investigate heterogeneity and inconsistency within the network. Specifically, NMR allows users to explore the interaction between treatment and study-level covariates. The aim of this paper is to describe the implementation and application of NMR in MetaInsight for a single covariate, which may be a study-level variable or baseline risk with the uncertainty in the latter correctly accounted for. METHODS:NMR has been added to MetaInsight using the R packages gemtc and bnma. The type of regression coefficients fitted can be set to shared, exchangeable, or unrelated, relating to whether the same or different relationships are assumed between the covariate and each treatment. A graph has been added to show the distribution of covariate values, and a novel visualization has been developed to show which studies contribute to which comparisons for multiple comparisons simultaneously. RESULTS:The new functionality is described and illustrated with an example, and screenshots of the app are included. CONCLUSION:This extensive update of the app greatly facilitates such complex analyses making them freely accessible to researchers from a wide range of backgrounds. This in turn should improve the reporting and reliability of published NMA which ultimately should positively impact clinical decision-making.

Background:Pulmonary rehabilitation (PR) is a key treatment for chronic obstructive pulmonary disease (COPD) recommended by all guidelines. However, programmes vary widely and the optimal combination of components to maximise benefits and efficiency remains unknown. We aimed to use the novel technique of component network meta-analysis (cNMA) to investigate the relative contribution of 1) exercise modality and intensity, 2) non-exercise components, 3) type of supervision, and 4) programme duration of PR for people with COPD. Methods:MEDLINE, EMBASE, CINAHL, and Cochrane databases searched in October 2023 with no date or language restrictions. We included randomised controlled trials (RCTs) which included an intervention involving exercise for people with COPD. We present outcomes of exercise capacity, breathlessness and health related quality of life (HRQoL). Screening and eligibility were assessed by two independent reviewers. cNMA, a technique developed to investigate complex interventions such as PR, was conducted to examine the contribution of single components within diverse multicomponent interventions controlling for cohort demographics. PROSPERO: CRD42022322058. Findings:We included 337 RCTs with 18,911 participants and 227 intervention components. In-person supervision enhanced gains in exercise capacity (Standardised mean difference (SMD) 0.41, 95% CrI 0.20; 0.63), HRQoL (0.43 95% CrI 0.19; 0.68) and breathlessness (0.31 95% CrI 0.04; 0.58) over exercise training alone with moderate to high certainty. Remote supervision increased gains in exercise capacity (0.40 95% CrI 0.08; 0.73) with trends towards improvements in HRQoL and breathlessness, with low certainty. Aerobic training appeared to be most effective for all outcomes at high or very high intensity but with low certainty. Addition of structured education did not improve any outcome. Psychological interventions led improvements in exercise capacity (0.37 95% CrI 0.01; 0.73, low certainty) and HRQoL (0.54 95% CrI 0.18; 0.91, moderate certainty). There was trend towards improvements in breathlessness with addition of breathing exercises (0.26 95% CrI -0.04; 0.56, low certainty). Programme duration did not impact outcomes. For outcomes of exercise capacity, HRQoL and breathlessness there were 60%, 63% and 59% studies at high risk of bias respectively. Interpretation:This large-scale analysis of over 300 randomised PR trials found the strongest effects for in-person supervised and prescribed aerobic exercise training with less certainty for the benefit of other commonly used PR components and delivery methods. Funding:This research was funded through a National Institute for Health and Care Research (NIHR) Applied Research Collaboration East Midlands grant (2.12) and carried out at the NIHR Leicester Biomedical Research Centre (BRC).

Since 2015, the Complex Reviews Synthesis Unit (CRSU) has developed a suite of web-based applications (apps) that conduct complex evidence synthesis meta-analyses through point-and-click interfaces. This has been achieved in the R programming language by combining existing R packages that conduct meta-analysis with the shiny web-application package. The CRSU apps have evolved from two short-term student projects into a suite of eight apps that are used for more than 3,000 h per month. Here, we present our experience of developing production grade web-apps from the point-of-view of individuals trained primarily as statisticians rather than software developers in the hopes of encouraging and inspiring other groups to develop valuable open-source statistical software whilst also learning from our experiences. We discuss how we have addressed challenges to research software development such as responding to feedback from our real-world users to improve the CRSU apps, the implementation of software engineering principles into our app development process and gaining recognition for non-traditional research work within the academic environment. The CRSU continues to seek funding opportunities both to maintain and further develop our shiny apps. We aim to increase our user base by implementing new features within the apps and building links with other groups developing complementary evidence synthesis tools.

In the context of an imperfect gold standard, latent class modelling can be used to estimate accuracy of multiple medical tests. However, the conditional independence (CI) assumption is rarely thought to be clinically valid. Two models accommodating conditional dependence are the latent class multivariate probit (LC-MVP) and latent trait models. Despite LC-MVP's greater flexibility - modelling full correlation matrices versus the latent trait's restricted structure - the latent trait has been more widely used. No simulation studies have directly compared these two models. We conducted a comprehensive simulation study comparing both models across five data generating mechanisms: CI, low-heterogeneity (latent trait-generated), and high-heterogeneity (LC-MVP-generated) correlation structures. We evaluated multiple priors, including novel constrained correlation priors using Pinkney's method that preserves prior interpretability. Models were fit using our BayesMVP R package, which achieves GPU-like speed-ups on these inherently serial models. The LC-MVP model demonstrated superior overall performance. Whilst the latent trait model performed acceptably on its own generated data, it failed for high-heterogeneity structures, sometimes performing worse than the CI model. The CI model did badly for most dependent structures. We also found ceiling effects: high sensitivities reduced the importance of correlation recovery, explaining paradoxes where models achieved good performance despite poor correlation recovery. Our results strongly favour LC-MVP for practical applications. The latent trait model's severe consequences under realistic correlation structures make it a more risky choice. However, LC-MVP with custom correlation constraints and priors provides a safer, more flexible framework for test accuracy evaluation without a perfect gold standard.

Standard (network) meta-analysis methods for medical test accuracy evaluation analyse the data separately for each test threshold - wasting data - unless every study reports all thresholds. Previously proposed "multiple threshold" models either fail to provide threshold-specific summary estimates, or they assume that ordinal tests (e.g., questionnaires) are continuous. We propose two ordinal regression models - ordinal-bivariate and ordinal-HSROC - using an induced-Dirichlet framework for cutpoint parameters, enabling intuitive priors and both fixed-effects and random-effects cutpoints. We conducted a simulation study to evaluate the performance of our proposed models, with the simulated data being based on real anxiety screening data spanning 7, 22, and 64 ordinal categories, with 15 Our proposed ordinal-bivariate model with fixed-effect cutpoints tended to obtain the best RMSE and bias, including when data was generated from a recently proposed continuous-assumption model. For instance - even with 64 categories - continuous models performed 10 We implemented the models in the MetaOrdDTA R package (https://github.com/CerulloE1996/MetaOrdDTA), which provides features such as: Stan estimation, K-fold cross-validation for model selection, meta-regression, network meta-analysis extensions, and visualisation tools including sROC plots with credible/prediction regions. Overall, our simulation study suggests that our proposed models may obtain better accuracy estimates than previous approaches for ordinal tests, even when the number of ordinal categories is very high.

Graphical displays are often utilised for high-quality reporting of meta-analyses. Previous work has presented augmentations to funnel plots that assess the impact that an additional trial would have on an existing meta-analysis. However, decision-makers, such as the National Institute for Health and Care Excellence in the United Kingdom, assess health technologies based on their cost-effectiveness, as opposed to efficacy alone. Motivated by this fact, this article outlines a novel approach, developed for augmenting funnel plots, based on the ability of an additional trial to change a decision regarding the optimal intervention. The approach is presented for a generalised class of economic decision models, where the clinical effectiveness of the health technology of interest is informed by a meta-analysis, and is illustrated with an example application. The ‘decision contours’ produced from the proposed methods have various potential uses not only for decision-makers and research funders but also for other researchers, such as meta-analysts and primary researchers designing new studies, as well as those developing health technologies, such as pharmaceutical companies. The relationship between the new approach and existing methods for determining sample size calculations for future trials is also considered.

Objective To rank commonly used patient-reported outcome measures (PROMs) for assessing pain in osteoarthritis trials according to their assay sensitivity, defined as the ability of a PROM to distinguish an effective from a less effective intervention or placebo, proposing a hierarchy for PROM selection in trials and data-extraction in meta-analyses. Design Analysis of trials with placebo, sham, or non-intervention control that included ≥100 patients per arm with knee/hip osteoarthritis, reporting treatment effects on ≥2 pain PROMs. Treatment effects from all PROMs were standardized on a 0-100 scale. Negative mean differences indicated a larger effect of the experimental treatment compared to control. We ranked PROMs by assay sensitivity using a Bayesian multi-outcome synthesis random-effects model. Results 135 trials comprising 57,141 participants were included. The ranking of PROMs from highest to lowest assay sensitivity was as follows: pain overall, pain on stairs, pain at night, pain on walking, pain at rest, WOMAC pain, WOMAC global, Lequesne index. Pain overall, the highest-ranked PROM, had a pooled mean difference of -6.96 (95%CrI -7.94, -6.02), while WOMAC pain, the most reported PROM in our study, had a pooled mean difference of -4.90 (95%CrI -5.55, -4.26). The pooled ratio of mean differences between pain overall and WOMAC pain was 1.42 (95%CrI 1.30, 1.55) representing a 42% larger effect size with pain overall. Conclusions Pain overall has better assay sensitivity than other pain PROMs. Investigators should consider the hierarchy proposed in this study to guide PROM selection in osteoarthritis clinical trials and data extraction in osteoarthritis meta-analyses.

OBJECTIVE:To quantify the effectiveness and safety of intra-articular interventions for knee and hip osteoarthritis (OA) through a systematic review and Bayesian random-effects network meta-analysis. DESIGN:We searched CENTRAL and regulatory agency websites (inception-2023) for large, English-language, randomized controlled trials (RCTs) (≥100 patients/group) examining any intra-articular intervention. PRIMARY OUTCOME:pain intensity. SECONDARY OUTCOMES:physical function and safety outcomes. Pain and function outcomes were analyzed at 2, 6, 12, 24, and 52 weeks post-randomization, and presented as standardized mean differences (SMDs) (95% credible intervals, 95% CrI). The prespecified minimal clinically important between-group difference (MID) was -0.37 SMD. Safety outcomes were presented as odds ratios (OR) (95% CrI). FINDINGS:Among 57 RCTs (22,795 participants) examining 18 intra-articular interventions, usual care or placebo, treatment effects were larger in 35 high-risk-of-bias trials than in 22 low/unclear-risk-of-bias trials. In the main analysis (excluding high-risk-of-bias trials), triamcinolone had the highest probabilities of reaching the MID at weeks 2 and 6 (75.3% and 90%, respectively) with corresponding SMDs of -0.48 (95% CrI,-0.85 to -0.10) and -0.53 (95% CrI,-0.79 to -0.27) compared to placebo (1 trial). The complex homeopathic products Tr14/Ze14 showed therapeutic potential at week 6 compared to placebo (SMD:-0.42, 95% CrI,-0.71 to -0.11, 63.5% probability of reaching the MID, 1 trial). Hyaluronic acid had no effect on pain (SMD:-0.04, 95% CrI,-0.19 to 0.11, 11 trials) but a higher risk of dropouts due to adverse events (OR: 2.01, 95% CrI,1.08 to 3.77) and serious adverse events (OR: 1.86, 95% CrI, 1.16 to 3.03) than placebo. CONCLUSION:Triamcinolone had the highest probabilities to have a treatment effect beyond the MID at weeks 2-6. Large RCTs with lower risk of bias indicate that the effects of 16 intra-articular interventions in knee or hip OA were smaller than the MID, and that most were consistent with placebo effects. Lack of evidence of long-term effectiveness underscores the need for further research beyond 24 weeks.

We introduce OrigamiPlot, an open-source R package and Shiny web application designed to enhance the visualization of multivariate data. This package implements the origami plot, a novel visualization technique proposed by Duan et al. in 2023, which improves upon traditional radar charts by ensuring that the area of the connected region is invariant to the ordering of attributes, addressing a key limitation of radar charts. The software facilitates multivariate decision-making by supporting comparisons across multiple objects and attributes, offering customizable features such as auxiliary axes and weighted attributes for enhanced clarity. Through the R package and user-friendly Shiny interface, researchers can efficiently create and customize plots without requiring extensive programming knowledge. Demonstrated using network meta-analysis as a real-world example, OrigamiPlot proves to be a versatile tool for visualizing multivariate data across various fields. This package opens new opportunities for simplifying decision-making processes with complex data.