Katy Sutcliffe

The Cochrane risk‐of‐bias tool (RoB 2) is used to assess the risk of bias in randomized trials in systematic reviews. However, studies have shown that the tool is often incorrectly used and published reviews rarely include detail regarding how it was applied or offer reflections on using it. This paper is a transparent applied case example with lessons learned and limitations of using RoB 2 in a (non‐Cochrane) systematic review evaluating the effectiveness of remote therapies for alcohol and drug use disorders. We share our reflections on aspects to plan in advance (pre‐selecting the outcome(s) and their results and piloting the tool); factors to consider when applying the tool (multiple outcomes within a study and navigating resources); overcoming common problems within specific domains (developing a review‐specific guidance document). Through our worked example, this article illustrates how review teams can engage with the RoB 2 tool within a specific review context to enhance its correct application.

This is a protocol for a Cochrane Review (qualitative). The objectives are as follows: This qualitative evidence synthesis (QES) aims to address the question: What is known from qualitative evidence about behavioural weight management for people living with mobility-limiting conditions, intellectual disabilities, or severe mental illness? The five review objectives are as follows. To explore individuals' experiences with weight and behavioural weight-management programmes. This includes understanding how individuals living with mobility-limiting conditions, intellectual disabilities, severe mental illness and their carers experience managing their weight and how they perceive and engage with behavioural weight-management programmes, including how accessible, supportive, and relevant they find the programme elements. To identify barriers to and facilitators of managing a healthy weight and accessing behavioural weight-management programmes. Focusing on the unique obstacles encountered by each population group, we will examine issues such as physical accessibility, communication challenges, financial constraints, and gaps in tailored support that may prevent equitable participation. We will also include people's perspectives on structural and system-level barriers, as well as the influence of intersectional factors (e.g. gender and ethnicity), on the challenges of maintaining weight and accessing and benefiting from support. To understand the impact of stigma and discrimination on programme engagement and outcomes. This objective seeks to uncover how stigma related to weight, disability, or mental health influences individuals' willingness to participate, their inclusion and sense of belonging with others within the programmes, and overall programme satisfaction. It also examines the role of stigma as a potential barrier to success. To capture the perspectives of programme providers and healthcare practitioners. Gathering insights from healthcare providers and programme facilitators, this objective assesses which programme elements are considered successful and which present ongoing challenges. This includes evaluating aspects like accessibility, engagement strategies, and support mechanisms that may require improvement. To produce an inclusive piece of research through incorporating the insight of people with lived experience and professional expertise in interpreting and considering the implications of the evidence. By synthesising this qualitative evidence, the review aims to provide a comprehensive understanding of how to design and implement behavioural weight-management programmes that are more inclusive, effective, and responsive to the needs of these populations.

Falls in hospitals are common events that cause considerable morbidity and mortality for older people. This is an update of the hospital components of a review of interventions in hospitals and care facilities first published in 2010 and updated in 2012 and 2018. To evaluate the benefits and harms of interventions designed to reduce the incidence of falls in older people in hospitals, including inpatient rehabilitation facilities. We searched CENTRAL, MEDLINE, Embase, CINAHL, and trial registers to 28 October 2025. Randomised controlled trials of any interventions for preventing falls in older people (aged 65 years and over) in hospitals, excluding trials where falls were recorded as adverse events of the intervention and those recruiting participants post-stroke or living with Parkinson's disease. Critical outcomes were rate of falls (number of falls per unit time) and risk of falling (number of fallers). Important outcomes were risk of fracture, adverse events, and economic outcomes. We evaluated risk of bias in the included trials against nine items (seven in the Cochrane RoB 1 tool, plus method of ascertaining falls and baseline imbalance). Two review authors independently performed study selection, risk of bias assessment, and data extraction. We calculated rate ratios (RaR) for rate of falls and risk ratios (RRs) for risk of falling and risk of fracture, with 95% confidence intervals (CIs). We adjusted for clustering if not undertaken by the trial authors. We pooled results where appropriate using the generic inverse variance method in RevMan. We conducted subgroup analyses according to intervention type, cognitive status, setting, and informed by an intervention component analysis. Where pooling was precluded by the nature of the data, we presented trial data in tables for illustrative purposes. We recategorised one trial considered as multifactorial in 2018 to multicomponent education for consistency within the group of trials considered as multifactorial. We used GRADE to assess the certainty of evidence for each outcome for the main comparisons. GRADE ratings of risk of bias were based on sensitivity analyses excluding trials at high risk. We included 32 new trials (35,235 participants) in this update, for a total of 55 trials (104,474 participants; mean age 79 years; 45% women). The majority of trials were at high risk of bias in one or more domains, mostly due to lack of blinding, which is not typically feasible for hospital fall prevention interventions. With few exceptions, the certainty of evidence for the critical outcomes of falls for individual interventions was rated as low or very low. We have reported outcomes for the main comparisons here. The outcomes of risk of fracture and economic outcomes were generally poorly reported and, where reported, the evidence was of very low certainty; we have not reported these data here. We only reported the outcome of adverse events when the certainty of evidence was stronger than very low. Exercise. The effect of exercise on the rate of in-hospital falls is uncertain (RaR 0.62, 95% CI 0.28 to 1.36; I² = 0%; 3 trials, 317 participants; very low-certainty evidence). Exercise may have little or no effect on the risk of falling in hospital settings overall (RR 0.73, 95% CI 0.36 to 1.46; I² = 27%; 6 trials, 668 participants; plus 1 trial, N = 12,863, adjusted odds ratio 0.8, 95% CI 0.5 to 1.1; P = 0.52; low-certainty evidence). Medication optimisation. We are uncertain of the effect of medication optimisation as a single intervention on the rate of falls (RaR 1.71, 95% CI 0.95 to 3.07; I² = 0%; 3 trials, 2093 participants; very low-certainty evidence) and risk of falling (RR 0.94, 95% CI 0.46 to 1.90; I² = 39%; 4 trials, 2255 participants; very low-certainty evidence). Medication optimisation may have no effect on adverse events overall (data not pooled, 3 trials; low-certainty evidence). Service model change (aspect of social environment category). Service model change (a change in a model of care or organisational system targeted at falls reduction) probably reduces the rate of falls in acute settings (RaR 0.45, 95% CI 0.27 to 0.74; I² = 0%; 6 trials, 10,825 participants; moderate-certainty evidence). However, the effect of service model change on risk of falling is uncertain (RR 0.93, 95% CI 0.25 to 3.46; I² = 70%; 3 trials, 5534 participants; very low-certainty evidence). Education (aspect of social environment category). Tailored education (patient, staff or multicomponent) probably reduces the rate of falls (RaR 0.73, 95% CI 0.56 to 0.94; I² = 0%; 3 trials, 4868 participants; moderate-certainty evidence) and risk of falling (RR 0.61, 95% CI 0.40 to 0.92; I² = 45%; 5 trials, 5035 participants; plus 1 trial, N = 3121 admissions, adjusted odds ratio 0.55, 95% CI 0.38 to 0.81; P = 0.003; moderate-certainty evidence). Multifactorial interventions. Multifactorial interventions, where two or more categories of interventions are given based on a fall risk assessment, probably reduce the rate of falls (RaR 0.87, 95% CI 0.69 to 1.08; I² = 40%; 5 trials, 42,256 participants; moderate-certainty evidence) and risk of falling (RR 0.82, 95% CI 0.63 to 1.07; I² = 0%; 5 trials, 41,141 participants; moderate-certainty evidence), but the CIs include the possibility of no effect or a slight increase in falls. Any intervention. Categorisation of intervention type was often not distinct. Overall, pooled data from all trials of any intervention to prevent falls had an uncertain impact on the rate of falls (RaR 0.85, 95% CI 0.71 to 1.02; I² = 51%; 27 trials, 72,016 participants; very low-certainty evidence), but suggested that there may be a reduced risk of falling (RR 0.83, 95% CI 0.73 to 0.94; I² = 0%; 35 trials, 56,452 participants; low-certainty evidence). A subgroup analysis found that trials of interventions that include integration with the local setting, tailoring approaches to the...

Background: Differential impacts of active travel interventions on disadvantaged population groups have been under-studied. A review of evidence from the UK was commissioned to inform the development of government policy and support the implementation of equitable and effective interventions.Methods: Systematic review, identifying studies via searching 31 electronic databases. Eligible studies evaluated active travel interventions in the UK and reported quantitative data on active travel behaviours and/or linked health outcomes for population subgroups defined by sex/gender, ethnicity, age, or socio-economic status. Two reviewers screened reports, extracted data and assessed risk of bias. Outcome data was summarised in relation to directions of intervention effects for population subgroups. Results: We included 18 study reports collectively evaluating five interventions. Evidence suggests that interventions are more likely to increase than decrease active travel among females and may also be more likely to benefit females relative to males. Interventions are also more likely to increase than decrease active travel among older adults, but there is no clear evidence that they will benefit this group relative to younger adults. Available evidence gives no clear indication of likely positive or negative impacts upon groups defined by ethnicity or socioeconomic status.Conclusion: There is some evidence that active travel interventions improve outcomes and may reduce inequalities in relation to sex, and improve outcomes among older adults. The available evidence for differential impacts of active travel interventions is otherwise largely inconclusive, but importantly provides no clear indication that implementing active travel interventions will widen inequalities.Funding: NIHR Policy Research Programme

Background: Diabetic retinopathy is a severe diabetes complication that can cause blindness. The United Kingdom’s pioneering diabetic eye screening programme has decreased blindness by early detection and treatment. Enhancing diabetic eye screening uptake requires a deeper understanding of the programme implementation. Objectives: This study aimed to develop a logic model depicting diabetic eye screening programme implementation and to systematically map evidence on the implementation of diabetic eye screening in the United Kingdom and countries with similar health systems: Australia, Canada, Ireland and New Zealand. Methods: A logic model was coproduced with UK National Screening Committee members and public coproducers with living experience of diabetic eye screening, informed by existing models and group knowledge. We searched 14 discipline-focused bibliographic databases, 3 academic search engines (Google Scholar, Bielefeld Academic Search Engine and OpenAlex) and targeted websites that covered the time frame up to December 2023. Eligible studies, from 2003 onwards, involved diabetic eye programme implementation in the target countries, covering a range of outcomes. Data extracted were publication year, study location (country), aim of study, study evaluation design, reported data (effectiveness outcomes, implementation outcomes, views/experiences data, observational data or data on resources required), study population, screening stage, intervention strategies and health inequality considerations. Findings are displayed as an interactive evidence map and searchable database. Results: The coproduced logic model depicted factors that could be mapped: screening stage, intervention strategy and evidence type as well as ‘black box’ factors that would require an in-depth synthesis to address: points for improvement and mechanisms of action. One hundred and thirty-three records were included the interactive map. The largest subset of studies provided information relevant to the entire screening pathway or multiple parts of this system ( n = 85), followed by interventions relating to delivery of the eye screening appointment ( n  = 36), while the fewest studies focused specifically on processes for identifying people eligible for screening. Few studies used experimental designs to evaluate the intervention effectiveness, and there were relatively few studies assessing how well interventions were implemented. Of the studies that reported the evaluation of some form of intervention, the most common type was environmental restructuring of the social and/or physical context ( n  = 40). The most common data types were observational (e.g. audit studies; n =  69) and views or experiences ( n  = 51). Most studies provided data that can contribute to tackling health inequalities ( n = 91). Limitations: We identified 328 additional records that met the general inclusion criteria but were not included in the map for pragmatic reasons (e.g. the record only presented a conference abstract or brief report with limited detail about the study). Thus, the map reflects a subset of the evidence base. Also, the review’s focus on five countries may omit valuable insights from elsewhere. Conclusions: A substantial body of evidence on diabetic eye programme implementation exists across five countries. However, evidence gaps remain, as certain process stages align with specific study types and data, highlighting areas for further research. The logic model and map may be useful for exploring ways to improve implementation of the programme. Future work: Future evidence syntheses could analyse subsets of studies on health inequalities, implementation experiences and outcomes, quality assurance processes or the underlying mechanisms of interventions. Primary research could address the various gaps in the evidence base. Funding: This article presents independent research funded by the National Institute for Health and Care Research (NIHR) Evidence Synthesis programme as award number NIHR159996.

Qualitative research seeks to provide deep understanding of people’s views, experiences, and behaviours and how these are shaped by social context. Qualitative evidence synthesis (QES) describes a range of methodologies for combining the findings of multiple qualitative studies to understand issues across a range of perspectives. QES is valuable when examining intervention complexity, as the nuanced insights into people’s lives help us to understand how well interventions are aligned with their needs and values and to explore the influence of context. Users’ or providers’ views on complex interventions QES can reveal overlooked features, important implementation factors, or unseen mechanisms that influence intervention success. In this chapter, we provide an overview of how to conduct a QES and options for integrating QES findings with findings from a complex intervention effects review in a mixed methods research synthesis (MMRS).

Background: Qualitative Comparative Analysis (QCA) uses Boolean set theory to identify complex causal ‘recipes’. QCA has been used in systematic reviews to identify configurations of intervention and contextual features associated with the effectiveness of complex multi-component interventions. Despite increasing use of QCA in systematic reviews, no tool currently exists to appraise reviews using QCA. To address this gap, we developed ‘CARU-QCA’ (Critical Appraisal of Reviews Using QCA). Methods: We used three sources of evidence to develop CARU-QCA: key papers on the logic and procedures of QCA; existing systematic review appraisal tools; and published QCA reviews. In addition to the existing six stages of QCA outlined elsewhere, we also distinguished a baseline stage (Stage 0) where we appraise the set-up and appropriateness of the use of QCA. Several rounds of user testing by the lead authors and colleagues in the field aimed to refine the tool and enhance its usability. Results: QCA blends a deep, holistic understanding of interventions and their underlying theory, with computer-based mathematical procedures for examining and testing patterns in the data. As such, elements of the CARU-QCA tool support the appraisal of the approach used to develop case-based knowledge and use of theory for exploring cross-case patterns as well as the technical analytic procedures. In line with systematic review guidance, CARU-QCA also appraises transparency of reporting. The tool is divided into seven sections to support users in appraising the set-up of QCA and the six analytical steps. The tool guides users to assess their confidence in the execution of each of the seven stages, before forming an overarching confidence assessment. A detailed guide accompanies the tool providing further background information and explanation of the rationale for each question.Conclusions: CARU-QCA provides a transparent and standardised approach for critiquing and evaluating reviews using QCA; it will be of value to those seeking to assess and implement the findings of QCA reviews. CARU-QCA will also be useful to systematic reviewers seeking guidance on how to undertake and report a QCA.

Background: Increased long-term antidepressant medication (ADM) use can lead to serious psychological and physical public health implications, as well as increased financial costs and social injustice. Given that health professionals (HPs) can influence decisions on ADM discontinuation, it is imperative that to promote and enable discontinuation when the medication is no longer indicated, their perspectives around this issue are explored in depth. This systematic review of qualitative evidence aimed to identify HPs’ perspectives on the barriers and facilitators to ADM discontinuation / deprescribing. Methods: Eligible studies were primary studies, on HPs’ views on ADM discontinuation / deprescribing, that used qualitative data collection and analysis, and were published in English. We searched PubMed, CINAHL, PsycINFO, the British Journal of General Practice, the Networked Digital Library of Theses and Dissertations, and Google Scholar, from July 2018 until May 2022, and harvested studies from a review published in 2019. The PubMed search was repeated post-synthesis, in January 2024, to identify and incorporate studies published since May 2022. Data was synthesised using thematic synthesis and further analysed using Bronfenbrenner’s Social Ecological Model (SEM) and a logic model. One reviewer extracted data and assessed study quality, which was checked by a second reviewer. Coding and development of themes and sub-themes was undertaken by one reviewer and further developed through discussion with a second reviewer. Results: Fourteen studies were included capturing the views of over 280 HPs. The synthesis yielded eight analytical themes: Perceptions of ADM; Perceptions of Depression; Sense of Professional Duty; Confidence in Supporting ADM Discontinuation; Assessment of Patients’ Circumstances and Characteristics; Assessment of Patients’ Desires, Motivations and Capabilities; Systemic Healthcare Delivery Issues; and, Societal Norms and Pressures. Conclusions: Use of the SEM and development of a logic model uncovered possible pathways and underlying reasons through which factors influencing ADM discontinuation interrelate both within and across societal levels. Societal norms and pressures, and systemic healthcare delivery issues, appear to influence directly or indirectly all aspects around ADM discontinuation / deprescribing. Future research and interventions could target barriers highlighted by these two themes demonstrating the largest impact.

Introduction: There is increasing work towards drawing on theory, implementing co‐production and accounting for complexity within the production of systematic reviews for public health. In this paper, we report on the process of co‐producing a theory; in this case, a graphical articulation of theory in the form of a logic model, which describes how contextual factors influence children's health. Methods: We undertook a series of three online co‐production workshops, involving 18–20 participants in each, and worked with an advisory group of experts with professional and lived expertise. An online virtual whiteboard was used to support the identification of factors that contributed to poorer childhood health, explanations for these factors, and connections between different factors. Results: Driven by government strategy, we initially focussed our work on childhood obesity. However, co‐production was transformational in switching the focus of the logic model away from a narrow focus on Body Mass Index as a measure of obesity, to a more holistic theory of factors that shape children's health, recognised as the intersection between healthy eating, physical activity and mental well‐being. Theorising with a diverse range of co‐producers helped us to recognise the stigmatising impacts that an exclusive focus on clinical measures of children's health can have, and the way that a narrow clinical focus inhibits theorising the complexity and drivers of poorer health. Conclusion: Co‐production led to a switch in theorising away from narratives of children's health that focus closely on personal responsibility, towards narratives that explore structural and contextual drivers of health. Patient or Public Contribution: The logic model was entirely driven by the contributions of researchers, those with lived experience (e.g., as parents and/or who have experienced poor health), and those with professional experience (e.g., as teachers) who worked together to co‐produce the model. An advisory group composed of people with a similar range of expertise helped to shape the conduct of co‐production and dissemination (including in the preparation of this manuscript).