Systematic reviews are essential for evidence-based health care, but conducting them is time- and resource-consuming. To date, efforts have been made to accelerate and (semi-)automate various steps of systematic reviews through the use of artificial intelligence (AI) and the emergence of large language models (LLMs) promises further opportunities. One crucial but complex task within systematic review conduct is assessing the risk of bias (RoB) of included studies. Therefore, the aim of this study was to test the LLM Claude 2 for RoB assessment of 100 randomized controlled trials, published in English language from 2013 onwards, using the revised Cochrane risk of bias tool (‘RoB 2’; involving judgements for five specific domains and an overall judgement). We assessed the agreement of RoB judgements by Claude with human judgements published in Cochrane reviews. The observed agreement between Claude and Cochrane authors ranged from 41% for the overall judgement to 71% for domain 4 (‘outcome measurement’). Cohen’s κ was lowest for domain 5 (‘selective reporting’; 0.10 (95% confidence interval (CI): −0.10–0.31)) and highest for domain 3 (‘missing data’; 0.31 (95% CI: 0.10–0.52)), indicating slight to fair agreement. Fair agreement was found for the overall judgement (Cohen’s κ: 0.22 (95% CI: 0.06–0.38)). Sensitivity analyses using alternative prompting techniques or the more recent version Claude 3 did not result in substantial changes. Currently, Claude’s RoB 2 judgements cannot replace human RoB assessment. However, the potential of LLMs to support RoB assessment should be further explored.

Mapping reviews are valuable tools for synthesizing and visualizing research evidence, providing a comprehensive overview of studies within a specific field. Their visual approach enhances accessibility, enabling researchers, policymakers, and practitioners to efficiently identify key findings, trends, and knowledge gaps. These reviews are particularly significant in guiding future research, informing funding decisions, and shaping evidence-based policymaking. In environmental science—similar to health and social sciences—mapping reviews play a crucial role in identifying effective conservation strategies, tracking interventions, and supporting targeted programs.

Unlike systematic reviews, which assess intervention effectiveness, mapping reviews focus on broad research questions, aiming to chart the existing evidence on a given topic. They use structured methodologies to identify patterns, gaps, and trends, often employing visual tools to enhance data accessibility. A well-defined scope, guided by inclusion and exclusion criteria, ensures a transparent study selection process. Comprehensive search strategies, often spanning multiple databases, maximize evidence capture. Effective screening, combining automated and manual processes, ensures relevance, while data extraction emphasizes high-level categories such as study design and population demographics. Advanced software tools, including EPPI-Reviewer and MindMeister, support data extraction and visualization, with evidence gap maps highlighting robust areas and research voids.

Despite their advantages, mapping reviews present challenges. The categorization and coding of studies can introduce subjective biases, and the process demands substantial resources. Automation and artificial intelligence offer promising solutions, improving efficiency while addressing integration and multilingual limitations. As methodological advancements continue, interdisciplinary collaboration will be essential to fully realize the potential of mapping reviews across scientific disciplines.

Limited evidence on relative effectiveness is common in Health Technology Assessment (HTA), often due to sparse evidence on the population of interest or study-design constraints. When evidence directly relating to the policy decision is limited, the evidence base could be extended to incorporate indirectly related evidence. For instance, a sparse evidence base in children could borrow strength from evidence in adults to improve estimation and reduce uncertainty. In HTA, indirect evidence has typically been either disregarded (‘splitting’; no information-sharing) or included without considering any differences (‘lumping’; full information-sharing). However, sophisticated methods that impose moderate degrees of information-sharing have been proposed. We describe and implement multiple information-sharing methods in a case-study evaluating the effectiveness, cost-effectiveness and value of further research of intravenous immunoglobulin for severe sepsis and septic shock. We also provide metrics to determine the degree of information-sharing. Results indicate that method choice can have significant impact. Across information-sharing models, odds ratio estimates ranged between 0.55 and 0.90 and incremental cost-effectiveness ratios between £16,000–52,000 per quality-adjusted life year gained. The need for a future trial also differed by information-sharing model. Heterogeneity in the indirect evidence should also be carefully considered, as it may significantly impact estimates. We conclude that when indirect evidence is relevant to an assessment of effectiveness, the full range of information-sharing methods should be considered. The final selection should be based on a deliberative process that considers not only the plausibility of the methods’ assumptions but also the imposed degree of information-sharing.

The current meta-analytic structural equation modeling (MASEM) techniques cannot properly deal with cases where there are multiple effect sizes available for the same relationship from the same study. Existing applications either treat these effect sizes as independent, randomly select one effect size amongst many, or create an average effect size. None of these approaches deal with the inherent dependency in effect sizes, and either leads to biased estimates or loss of information and power. An alternative technique is to use univariate three-level modeling in the two-stage approach to model these dependencies. These different strategies for dealing with dependent effect sizes in the context of MASEM have not been previously compared in a simulation study. This study aims to compare the performance of these strategies across different conditions; varying the number of studies, the number of dependent effect sizes within studies, the correlation between the dependent effect sizes, the magnitude of the path coefficient, and the between-studies variance. We examine the relative bias in parameter estimates and standard errors, coverage proportions of confidence intervals, as well as mean standard error and power as measures of efficiency. The results suggest that there is not one method that performs well across all these criteria, pointing to the need for better methods.

When conducting a systematic review, screening the vast body of literature to identify the small set of relevant studies is a labour-intensive and error-prone process. Although there is an increasing number of fully automated tools for screening, their performance is suboptimal and varies substantially across review topic areas. Many of these tools are only trained on small datasets, and most are not tested on a wide range of review topic areas. This study presents two systematic review datasets compiled from more than 8600 systematic reviews and more than 540000 abstracts covering 51 research topic areas in health and medical research. These datasets are the largest of their kinds to date. We demonstrate their utility in training and evaluating language models for title and abstract screening. Our dataset includes detailed metadata of each review, including title, background, objectives and selection criteria. We demonstrated that a small language model trained on this dataset with additional metadata has excellent performance with an average recall above 95% and specificity over 70% across a wide range of review topic areas. Future research can build on our dataset to further improve the performance of fully automated tools for systematic review title and abstract screening.

Unanchored population-adjusted indirect comparisons (PAICs) such as matching-adjusted indirect comparison (MAIC) and simulated treatment comparison (STC) attracted a significant attention in the health technology assessment field in recent years. These methods allow for indirect comparisons by balancing different patient characteristics in single-arm studies in the case where individual patient-level data are only available for one study. However, the validity of findings from unanchored MAIC/STC analyses is frequently questioned by decision makers, due to the assumption that all potential prognostic factors and effect modifiers are accounted for. Addressing this critical concern, we introduce a sensitivity analysis algorithm for unanchored PAICs by extending quantitative bias analysis techniques traditionally used in epidemiology. Our proposed sensitivity analysis involves simulating important covariates that were not reported by the comparator study when conducting unanchored STC and enables the formal evaluating of the impact of unmeasured confounding in a quantitative manner without additional assumptions. We demonstrate the practical application of this method through a real-world case study of metastatic colorectal cancer, highlighting its utility in enhancing the robustness and credibility of unanchored PAIC results. Our findings emphasise the necessity of formal quantitative sensitivity analysis in interpreting unanchored PAIC results, as it quantifies the robustness of conclusions regarding potential unmeasured confounders and supports more robust, reliable, and informative decision-making in healthcare.

Network meta-analysis (NMA) enables simultaneous assessment of multiple treatments by combining both direct and indirect evidence. While NMAs are increasingly important in healthcare decision-making, challenges remain due to limited direct comparisons between treatments. This data sparsity complicates the accurate estimation of correlations among treatments in arm-based NMA (AB-NMA). To address these challenges, we introduce a novel sensitivity analysis tool tailored for AB-NMA. This study pioneers a tipping point analysis within a Bayesian framework, specifically targeting correlation parameters to assess their influence on the robustness of conclusions about relative treatment effects. The analysis explores changes in the conclusion based on whether the 95% credible interval includes the null value (referred to as the interval conclusion) and the magnitude of point estimates. Applying this approach to multiple NMA datasets, including 112 treatment pairs, we identified tipping points in 13 pairs (11.6%) for interval conclusion change and in 29 pairs (25.9%) for magnitude change with a threshold at 15%. These findings underscore potential commonality in tipping points and emphasize the importance of our proposed analysis, especially in networks with sparse direct comparisons or wide credible intervals for correlation estimates. A case study provides a visual illustration and interpretation of the tipping point analysis. We recommend integrating this tipping point analysis as a standard practice in AB-NMA.

The importance of network meta-analysis (NMA) methods for time-to-event (TTE) that do not rely on the proportional hazard (PH) assumption is increasingly recognized in oncology, where clinical trials evaluating new interventions versus standard comparators often violate this assumption. However, existing NMA methods that allow for time-varying treatment effects do not directly leverage individual events and censor times that can be reconstructed from Kaplan–Meier curves, which may be more accurate than discrete hazards. They are also challenging to implement given reparameterizations that rely on discrete hazards. Additionally, two-step methods require assumptions regarding within-study normality and variance. We propose a one-step fully Bayesian parametric individual patient data (IPD)-NMA model that fits TTE data with the exact likelihood and allows for time-varying treatment effects. We define fixed or random effects with the following distributions: Weibull, Gompertz, log-normal, log-logistic, gamma, or generalized gamma distributions. We apply the one-step model to a network of randomized controlled trials (RCTs) evaluating multiple interventions for advanced melanoma and compare results with those obtained with the two-step approach. Additionally, a simulation study was performed to compare the proposed one-step method to the two-step method. The one-step method allows for straightforward model selection among the “standard” distributions, now including gamma and generalized gamma, with treatment effects on either the scale alone or with multivariate treatment effects. Generalized gamma offers flexibility to model U-shaped hazards within a network of RCTs, with accessible interpretation of parameters that simplifies to exponential, Weibull, log-normal, or gamma in special cases.

Random-effects meta-analyses with only a few studies often face challenges in accurately estimating between-study heterogeneity, leading to biased effect estimates and confidence intervals with poor coverage. This issue is especially the case when dealing with rare diseases. To address this problem for normally distributed outcomes, two new approaches have been proposed to provide confidence limits of the global mean: one based on fiducial inference, and the other involving two modifications of the signed log-likelihood ratio test statistic in order to have improved performance with small numbers of studies. The performance of the proposed methods was evaluated numerically and compared with the Hartung–Knapp–Sidik–Jonkman approach and its modification to handle small numbers of studies. The simulation results indicated that the proposed methods achieved coverage probabilities closer to the nominal level and produced shorter confidence intervals compared to those based on existing methods. Two real examples are used to illustrate the proposed methods.

In meta-analyses of survival rates, precision information (i.e., standard errors (SEs) or confidence intervals) are often missing in clinical studies. In current practice, such studies are often excluded from the synthesis analyses. However, the naïve deletion of these incomplete data can produce serious biases and loss of precision in pooled estimators. To address these issues, we developed a simple but effective method to impute precision information using commonly available statistics from individual studies, such as sample size, number of events, and risk set size at a time point of interest. By applying this new method, we can effectively circumvent the deletion of incomplete data, resultant biases, and losses of precision. Based on extensive simulation studies, the developed method markedly improves the accuracy and precision of the pooled estimators compared to those of naïve analyses that delete studies with missing precision. Furthermore, the performance of the proposed method was not significantly inferior to the ideal case, where there was no missing precision information. However, for studies for which the risk set size at the time of interest was not available, the proposed method runs the risk of overestimating the SE. Although the proposed method is a single-imputation method, the simulations show that there is no underestimation bias of the SE, even though the proposed method does not consider the uncertainty of missing values. To demonstrate the robustness of our proposed methods, they were applied in a systematic review of radiotherapy data. An R package was developed to implement the proposed procedure.

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.

In small meta-analyses (e.g., up to 20 studies), the best-performing frequentist methods can yield very wide confidence intervals for the meta-analytic mean, as well as biased and imprecise estimates of the heterogeneity. We investigate the frequentist performance of alternative Bayesian methods that use the invariant Jeffreys prior. This prior has the usual Bayesian motivation, but also has a purely frequentist motivation: the resulting posterior modes correspond to the established Firth bias correction of the maximum likelihood estimator. We consider two forms of the Jeffreys prior for random-effects meta-analysis: the previously established “Jeffreys1” prior treats the heterogeneity as a nuisance parameter, whereas the “Jeffreys2” prior treats both the mean and the heterogeneity as estimands of interest. In a large simulation study, we assess the performance of both Jeffreys priors, considering different types of Bayesian estimates and intervals. We assess point and interval estimation for both the mean and the heterogeneity parameters, comparing to the best-performing frequentist methods. For small meta-analyses of binary outcomes, the Jeffreys2 prior may offer advantages over standard frequentist methods for point and interval estimation of the mean parameter. In these cases, Jeffreys2 can substantially improve efficiency while more often showing nominal frequentist coverage. However, for small meta-analyses of continuous outcomes, standard frequentist methods seem to remain the best choices. The best-performing method for estimating the heterogeneity varied according to the heterogeneity itself. Röver & Friede’s R package bayesmeta implements both Jeffreys priors. We also generalize the Jeffreys2 prior to the case of meta-regression.

In prognosis studies with time-to-event outcomes, the survivals of groups with high/low biomarker expression are often estimated by the Kaplan–Meier method, and the difference between groups is measured by the hazard ratios (HRs). Since the high/low expressions are usually determined by study-specific cutoff values, synthesizing only HRs for summarizing the prognostic capacity of a biomarker brings heterogeneity in the meta-analysis. The time-dependent summary receiver operating characteristics (SROC) curve was proposed as a cutoff-free summary of the prognostic capacity, extended from the SROC curve in meta-analysis of diagnostic studies. However, estimates of the time-dependent SROC curve may be threatened by reporting bias in that studies with significant outcomes, such as HRs, are more likely to be published and selected in meta-analyses. Under this conjecture, this paper proposes a sensitivity analysis method for quantifying and adjusting reporting bias on the time-dependent SROC curve. We model the publication process determined by the significance of the HRs and introduce a sensitivity analysis method based on the conditional likelihood constrained by some expected proportions of published studies. Simulation studies showed that the proposed method could reduce reporting bias given the correctly-specified marginal selection probability. The proposed method is illustrated on the real-world meta-analysis of Ki67 for breast cancer.

Model-based recursive partitioning (MOB) and its extension, metaMOB, are tools for identifying subgroups with differential treatment effects. When pooling data from various trials the metaMOB approach uses random effects to model the heterogeneity of treatment effects. In situations where interventions offer only small overall benefits and require extensive, costly trials with a large participant enrollment, leveraging individual-participant data (IPD) from multiple trials can help identify individuals who are most likely to benefit from the intervention. We explore the application of MOB and metaMOB in the context of non-specific low back pain treatment, using synthetic data based on a subset of the individual participant data meta-analysis by Patel et al.1 Our study underscores the need to explore heterogeneity in intercepts and treatment effects to identify subgroups with differential treatment effects in IPD meta-analyses.

Effect modification occurs when a covariate alters the relative effectiveness of treatment compared to control. It is widely understood that, when effect modification is present, treatment recommendations may vary by population and by subgroups within the population. Population-adjustment methods are increasingly used to adjust for differences in effect modifiers between study populations and to produce population-adjusted estimates in a relevant target population for decision-making. It is also widely understood that marginal and conditional estimands for non-collapsible effect measures, such as odds ratios or hazard ratios, do not in general coincide even without effect modification. However, the consequences of both non-collapsibility and effect modification together are little-discussed in the literature.

In this article, we set out the definitions of conditional and marginal estimands, illustrate their properties when effect modification is present, and discuss the implications for decision-making. In particular, we show that effect modification can result in conflicting treatment rankings between conditional and marginal estimates. This is because conditional and marginal estimands correspond to different decision questions that are no longer aligned when effect modification is present. For time-to-event outcomes, the presence of covariates implies that marginal hazard ratios are time-varying, and effect modification can cause marginal hazard curves to cross. We conclude with practical recommendations for decision-making in the presence of effect modification, based on pragmatic comparisons of both conditional and marginal estimates in the decision target population. Currently, multilevel network meta-regression is the only population-adjustment method capable of producing both conditional and marginal estimates, in any decision target population.

We developed a geographic search filter for retrieving studies about Germany from PubMed. In this study, we aimed to translate and validate it for use in Embase and MEDLINE(R) ALL via Ovid. Adjustments included aligning PubMed field tags with Ovid’s syntax, adding a keyword heading field for both databases, and incorporating a correspondence address field for Embase. To validate the filters, we used systematic reviews (SRs) that included studies about Germany without imposing geographic restrictions on their search strategies. Subsequently, we conducted (i) case studies (CSs), applying the filters to the search strategies of the 17 eligible SRs; and (ii) aggregation studies, combining the SRs’ search strategies with the ‘OR’ operator and applying the filters. In the CSs, the filters demonstrated a median sensitivity of 100% in both databases, with interquartile ranges (IQRs) of 100%–100% in Embase and 93.75%–100% in MEDLINE(R) ALL. Median precision improved from 0.11% (IQR: 0.05%–0.30%) to 1.65% (IQR: 0.78%–3.06%) and from 0.19% (IQR: 0.11%–0.60%) to 5.13% (IQR: 1.77%–6.85%), while the number needed to read (NNR) decreased from 893.40 (IQR: 354.81–2,219.58) to 60.44 (IQR: 33.94–128.97) and from 513.29 (IQR: 167.35–930.99) to 19.50 (IQR: 14.66–59.35) for Embase and MEDLINE(R) ALL, respectively. In the aggregation studies, the overall sensitivities were 98.19% and 97.14%, with NNRs of 83.29 and 33.34 in Embase and MEDLINE(R) ALL, respectively. The new Embase and MEDLINE(R) ALL filters for Ovid reliably retrieve studies about Germany, enhancing search precision. The approach described in our study can support search filter developers in translating filters for various topics and contexts.

Our objective was to evaluate the recall and number needed to read (NNR) for the Cochrane RCT Classifier compared to and in combination with established search filters developed for Ovid MEDLINE and Embase.com. A gold standard set of 1,103 randomized controlled trials (RCTs) was created to calculate recall for the Cochrane RCT Classifier in Covidence, the Cochrane sensitivity-maximizing RCT filter in Ovid MEDLINE and the Cochrane Embase RCT filter for Embase.com. In addition, the classifier and the filters were validated in three case studies using reports from the Swedish Agency for Health Technology Assessment and Assessment of Social Services to assess impact on search results and NNR. The Cochrane RCT Classifier had the highest recall with 99.64% followed by the Cochrane sensitivity-maximizing RCT filter in Ovid MEDLINE with 98.73% and the Cochrane Embase RCT filter with 98.46%. However, the Cochrane RCT Classifier had a higher NNR than the RCT filters in all case studies. Combining the RCT filters with the Cochrane RCT Classifier reduced NNR compared to using the RCT filters alone while achieving a recall of 98.46% for the Ovid MEDLINE/RCT Classifier combination and 98.28% for the Embase/RCT Classifier combination. In conclusion, we found that the Cochrane RCT Classifier in Covidence has a higher recall than established search filters but also a higher NNR. Thus, using the Cochrane RCT Classifier instead of current state-of-the-art RCT filters would lead to an increased workload in the screening process. A viable option with a lower NNR than RCT filters, at the cost of a slight decrease in recall, is to combine the Cochrane RCT Classifier with RCT filters in database searches.

Machine learning (ML) models have been developed to identify randomised controlled trials (RCTs) to accelerate systematic reviews (SRs). However, their use has been limited due to concerns about their performance and practical benefits. We developed a high-recall ensemble learning model using Cochrane RCT data to enhance the identification of RCTs for rapid title and abstract screening in SRs and evaluated the model externally with our annotated RCT datasets. Additionally, we assessed the practical impact in terms of labour time savings and recall improvement under two scenarios: ML-assisted double screening (where ML and one reviewer screened all citations in parallel) and ML-assisted stepwise screening (where ML flagged all potential RCTs, and at least two reviewers subsequently filtered the flagged citations). Our model achieved twice the precision compared to the existing SVM model while maintaining a recall of 0.99 in both internal and external tests. In a practical evaluation with ML-assisted double screening, our model led to significant labour time savings (average 45.4%) and improved recall (average 0.998 compared to 0.919 for a single reviewer). In ML-assisted stepwise screening, the model performed similarly to standard manual screening but with average labour time savings of 74.4%. In conclusion, compared with existing methods, the proposed model can reduce workload while maintaining comparable recall when identifying RCTs during the title and abstract screening stages, thereby accelerating SRs. We propose practical recommendations to effectively apply ML-assisted manual screening when conducting SRs, depending on reviewer availability (ML-assisted double screening) or time constraints (ML-assisted stepwise screening).