The ctrdata package has been created to boost the use of data available in public registers of clinical trials. It enables user-friendly, reproducible workflows to identify trials of interest, download protocol- and results-related data, and conduct sophisticated analyses, across multiple registers and trials. ctrdata works in the widely used R environment, and its databases can be used with other tools. The package is open source with a permissive licence, to facilitate collaboration.

This report provides an overview of ctrdata, including its implementation, cases of interest to researchers in public health, medicines, and regulatory science, as well as potential limitations and further developments. At this time, ctrdata works with the European Union (EU) Clinical Trials Information System (CTIS), the EU Clinical Trials Register (EUCTR), the US Clinicaltrials.Gov (CTGOV), and the ISRCTN—the UK’s Clinical Study Registry. The registers are complementary in scope and scientific value, yet differences in data models, variable definitions, search parametrisations, and retrieval options hamper efficient scientific workflows, calling for a scientific-technical, programmatic solution and driving the development of ctrdata.

By employing ctrdata to comprehensively use and easily leverage trial register data, researchers can effectively address a variety of questions, gain useful insights into evolving policies and practices of drug development, and inform further clinical research. Patients and their organisations, developers, policymakers, and other interested parties can build on ctrdata to create solutions for their use cases.

High expressed emotion (EE) in families is known to increase risk of relapse for people with schizophrenia. This Cochrane Review by Chien et al suggests that family-based interventions such as psychoeducation may result in a reduction in caregiver burden, a shift from high to low EE, and perhaps also reduce patient relapse rates. However, there was a high degree of heterogeneity in the combined study sample and a significant risk of bias across studies. The authors’ decision to only include studies reporting both a family member and a patient outcome means relevant evidence in this area may not have been incorporated.

Accurate assessment of adverse event (AE) incidence is critical in clinical research for drug safety. While meta-analysis serves as an essential tool to comprehensively synthesize the evidence across multiple studies, incomplete AE reporting in clinical trials remains a persistent challenge. In particular, AEs occurring below study-specific reporting thresholds are often omitted from publications, leading to left-censored data. Failure to account for these censored AE counts can result in biased AE incidence estimates. We present an R Shiny application that implements a Bayesian meta-analysis model specifically designed to incorporate censored AE data into the estimation process. This interactive tool provides a user-friendly interface for researchers to conduct AE meta-analyses and estimate the AE incidence probability using an unbiased approach. It also enables direct comparisons between models that either incorporate or ignore censoring, highlighting the biases introduced by conventional approaches. This tutorial demonstrates the Shiny application’s functionality through an illustrative example on meta-analysis of PD-1/PD-L1 inhibitor safety and highlights the importance of this tool in improving AE risk assessment. Ultimately, the new Shiny app facilitates more accurate and transparent drug safety evaluations. The Shiny-MAGEC app is available at: https://zihanzhou98.shinyapps.io/Shiny-MAGEC/.

Infant self-regulation is shaped by early physiological systems and caregiver-infant co-regulatory interactions. Maternal perinatal (pre- and/or postnatal) depression may affect these processes and infants’ development of this critical construct. However, literature addressing the association between maternal perinatal depression and infant self-regulation has been mixed. We conducted a pre-registered meta-analysis of the association between maternal perinatal depression and several self-regulation constructs (e.g., effortful control, executive function) measured during the first 2 years of life. We included 68 reports comprising 193 effect sizes and 16,722 mother-infant dyads. On average, studies included an equal number of male and female infants, and, for most (68%) studies, most participants were White. Average infant age ranged from 0 – 16 months. Three-level random effects meta-analytic models indicated a small, significant overall association, with higher levels of depression associated with lower self-regulation (r = −.10, 95% CI = −.14, −.06, p < .001). There was substantial heterogeneity in this pooled effect. Subsequent analyses indicated moderation by methodological and conceptual variables. Evidence that maternal perinatal depression is associated with lower infant self-regulation underscores the importance of supporting dyads experiencing perinatal depression. Clarifying this association highlights a critical next step of examining potential causal processes linking maternal and infant well-being.

Polycystic ovary syndrome (PCOS) is a disorder characterized by insulin resistance, low-grade inflammation, and increased adipose tissue. The very low-carbohydrate ketogenic diet (VLCKD) has been suggested to reduce obesity risks in PCOS. This study aimed to update the evidence on effects of the very low-carbohydrate ketogenic diet in women with PCOS. Searches were conducted in electronic databases for randomized clinical trials (RCTs) addressing the research question. The values for the meta-analysis were presented as weighted mean difference (WMD). Twelve studies were included in the qualitative analysis and eleven in the quantitative analysis. Significant reductions were observed in anthropometric outcomes: weight [WMD: -9.57 kg; p < 0.0001], waist circumference [WMD: -7.75 cm; p < 0.0009], fat body mass [WMD: -7.44 kg; p = 0.0008], body mass index [WMD: -3.45 kg/m2; p < 0.0001] and waist-to-hip ratio [WMD: -0.02; p < 0.0034]. Hormonal improvements included free testosterone [WMD: -0.31 ng/dL; p < 0.0001], total testosterone [WMD: -7.21 ng/dL; p<0.0001], sex hormone binding globulin [WMD: 15.22 nmol/L; p =0.0035], luteinizing hormone (LH) [WMD: -3.97 U/L; p =0.0008], and LH/FSH ratio [WMD: -1.04; p =0.0053], but not for follicle-stimulating hormone (FSH) levels [WMD: 1.23 mUI/mL; p =0.12]. Significant changes in metabolic markers were seen in blood glucose [WMD: -9.65 mg/dL; p =0.0031], insulin [WMD: -2.41 mg/dL; p =0.0387], homeostatic model assessment for insulin resistance [WMD: -2.46; p =0.0123], and triglycerides [WMD: -29.95 mg/dL; p =0.0188]. VLCKD shows significant benefits in managing body composition, reducing hyperandrogenism, balancing sex hormones, and improving glucose metabolism in PCOS.

Meta-research and evidence synthesis require considerable resources. Large language models (LLMs) have emerged as promising tools to assist in these processes, yet their performance varies across models, limiting their reliability. Taking advantage of the large availability of small size (<10 billion parameters) open-source LLMs, we implemented an agreement-based framework in which a decision is taken only if at least a given number of LLMs produce the same response. The decision is otherwise withheld. This approach was tested on 1020 abstracts of randomized controlled trials in rheumatology, using 2 classic literature review tasks: (1) classifying each intervention as drug or nondrug based on text interpretation and (2) extracting the total number of randomized patients, a task that sometimes required calculations. Re-examining abstracts where at least 4 LLMs disagreed with the human gold standard (dual review with adjudication) allowed constructing an improved gold standard. Compared to a human gold standard and single large LLMs (>70 billion parameters), our framework demonstrated robust performance: several model combinations achieved accuracies above 95% exceeding the human gold standard on at least 85% of abstracts (e.g., 3 of 5 models, 4 of 6 models, or 5 of 7 models). Performance variability across individual models was not an issue, as low-performing models contributed fewer accepted decisions. This agreement-based framework offers a scalable solution that can replace human reviewers for most abstracts, reserving human expertise for more complex cases. Such frameworks could significantly reduce the manual burden in systematic reviews while maintaining high accuracy and reproducibility.

Nutraceuticals are increasingly of interest in nutritional psychiatry, where creatine has been investigated in several randomised trials for its effects on depressive symptoms. However, these findings have not yet been systematically synthesised. We conducted a systematic review to assess the effects of creatine supplementation on symptoms of depression. Four databases were searched up to February 2025 for trials comparing creatine with placebo in individuals with or without depression. Study selection, data extraction and risk of bias assessment (RoB 2) were conducted independently, and certainty of evidence was evaluated using Grading of Recommendations, Assessment, Development and Evaluation (GRADE). Random-effects meta-analyses with Hartung–Knapp–Sidik–Jonkman adjustment including eleven trials (1093 participants) found a standardised mean difference of −0·34 (95 % CI −0·68, −0·00; GRADE: very low quality of evidence), equivalent to 2·2 points on the seventeen-item Hamilton Depression Rating Scale, below the minimal important difference of 3·0 points. CI include non-clinically important effects, and heterogeneity was substantial (I2 = 71·3 %). While effects appeared larger in clinically depressed populations, subgroup analyses and trim-and-fill adjustments indicated substantial bias favouring creatine. Results for secondary endpoints were significant for remission (three trials, OR 3·60; 95 % CI 1·76, 7·56), but not for treatment response (two trials, OR 0·72; 95 % CI 0·28, 1·88). Our findings suggest that creatine may offer a small-to-moderate benefit for individuals with depression, but average effects were not clinically important and the true effect may be trivial or null. The evidence on which these results are based is very uncertain. Larger, more rigorous randomised trials are required to draw definitive conclusions.

Systematic reviews and meta-analyses are often considered the highest level in evidence hierarchies, and therefore are often drawn upon when considering changes in policy. Despite journals implementing measures aiming to enhance the quality of systematic reviews they publish, the authorship raise concerns about the quality of existing and ongoing systematic reviews, particularly relating to transparency and bias minimisation. Building on the current guidelines, standards and tools, we suggest a ‘meta checklist’ which aims to maximise methodologically sound, unbiased and reproducible reviews of the best scientific quality while considering feasibility throughout the process.

Suicide and self-harm in people with depression are major public health concerns; electroconvulsive therapy (ECT) is a treatment recommended in UK clinical guidelines for severe mood disorders. We aimed to investigate published literature on the effect of ECT on the incidence of suicide, self-harm, and the recorded presence of suicidal thoughts (suicide-related outcomes). We hypothesized that ECT would be associated with a reduced incidence of suicide-related outcomes and all-cause mortality. We reviewed systematically all eligible studies as specified in our protocol (PROSPERO 293393). We included studies that compared ECT against a comparator treatment, and which included suicide-related outcomes or mortality. We searched Medline, EMBASE, and PsycINFO on January 24, 2022, updated to February 12, 2025. We identified 12,313 records and, after deduplication, screened 8,281 records on title and abstract and 212 on full-text, identifying 17 eligible studies. Studies showed significant heterogeneity in methodology, outcomes, time points chosen, and study populations. Three included studies investigated change in the suicidality domain on psychological rating scales: two showed a reduction in the ECT group; the other was underpowered for this outcome. Meta-analysis of suicide outcomes showed significant statistical heterogeneity and did not detect differences in a consistent direction. Meta-analysis of other mortality outcomes showed reductions in the risk of all-cause mortality (log relative risk [logRR]: −0.29; 95% CI: −0.53, −0.05) and non-suicide mortality (logRR: −0.21; 95% CI: −0.35, −0.07). Further high-quality studies are needed, which should seek to minimize biases (particularly confounding by indication) and report a wider range of suicide-related outcomes.

The present study aimed to conduct a meta-analysis using the random-effects model to merge published genetic parameter estimates for metabolic disorders [mastitis (MA), metritis (ME), ketosis (KE), milk fever (MF), retained placenta (RP), displaced abomasum (DA), and lameness (LA)] in dairy cows. A total of 239 heritability and 355 genetic correlation estimates were extracted from 61 peer-reviewed articles published between 1991 and 2024. Heritability estimates for metabolic disorders were generally low, ranging from 0.034 for ME to 0.067 for DA and MF. These estimates were associated with small standard errors and narrow 95% confidence intervals, with all estimates being statistically significant (P < 0.05). The estimates for genetic correlations between metabolic disorders were positive, ranging from 0.038 (MA-DA) to 0.522 (KE-DA). Furthermore, only the genetic correlations for DA-LA (0.147), RP-ME (0.472), MA-LA (0.239), MA-MF (0.365), KE-DA (0.522), KE-RP (0.124), and KE-ME (0.157) were significant (P < 0.05). Positive and low genetic correlations were observed between MA-milk yield (MY), MF-MY, and LA-MY (0.291, 0.135, and 0.263, respectively; P < 0.05). Similarly, positive correlations were found between DA-fat yield (FY), MA-FY, MA-protein yield (PY), LA-FY, and LA-PY (0.121, 0.170, 0.303, 0.176, and 0.236, respectively; P < 0.05). The genetic correlations between KE-fat to protein ratio (FPR) and DA-FPR were positive and ranged from low to moderate (0.520 and 0.196, respectively; P < 0.05). Significant genetic correlations were also noted between MA-days open (DO) (0.416), MA-days from calving to first service (DCFS) (0.207), and LA-calving interval (CI) (0.210; P < 0.05). This meta-analysis provides valuable insights into the genetic contribution to metabolic disorders in dairy cows and highlights the generally low heritability and the variable genetic correlations between these disorders and performance traits. These findings may contribute to more effective breeding strategies aimed at reducing the occurrence of metabolic disorders in dairy herds.

To understand a treatment’s potential impact at the individual level, it is crucial to explore whether the effect differs across patient subgroups and covariate values. Meta-analysis provides an important tool for detecting treatment–covariate interactions, as it can improve power compared to a single study. However, aggregation bias can occur when estimating individual-level treatment–covariate interactions in meta-analysis, due to trial-level confounding. This refers to when the association between the covariate and treatment effect across trials (at the aggregate level) differs from that observed within trials (at the individual level). It is, thus, recommended that heterogeneity in the treatment effect at the individual level should be disentangled from that at the trial level, ideally using an individual participant data (IPD) meta-analysis. Here, we explain this issue and provide new intuition about how trial-level confounding is impacted by differences in within-trial distributions of covariates and how this corresponds to asymmetry in subgroup-specific funnel plots in the case of categorical covariates. We then propose a sensitivity analysis to assess the robustness of interaction estimates to potential trial-level confounding. We illustrate these concepts using simulated and real data from an IPD meta-analysis of trials conducted on the TICO/ACTIV-3 platform, which assessed passive immunotherapy treatments for inpatients with COVID-19.

Sepsis affects 50 million people globally, contributing to 20 % of all deaths and significantly increasing healthcare costs due to intensive care needs. Although the role of n-3 fatty acids in reducing sepsis mortality remains debated, recent studies suggest their potential in modulating immune responses and improving outcomes. This umbrella review aims to clarify the benefits of n-3 supplementation on mortality rate, length of intensive care unit (ICU) stays and days on mechanical ventilation in patients with sepsis. Following Cochrane and Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodologies, a systematic search was conducted across multiple databases up to February 2025. After independent screening, data extraction and critical appraisal, meta-analyses were reassessed using the DerSimonian and Laird model. Evidence was graded using the GRADE approach, categorising outcomes based on strength and quality. A comprehensive search identified 934 records, of which thirty-four randomised controlled trials (RCT) from twenty-one systematic reviews and meta-analyses focused on n-3 supplementation in sepsis patients. n-3 significantly reduced mortality (risk ratio: 0·79, 95 % CI 0·69, 0·90), length of ICU stays (mean difference (MD): −3·6 d, 95 % CI −4·39, −2·81) and ventilation days (MD: −2·86 d, 95 % CI −4·46, −1·26). Parenteral nutrition showed slightly better outcomes than enteral nutrition, and EPA and DHA provided superior results compared with mixed oils. These findings suggest n-3 supplementation could improve mortality, ICU stays and ventilator dependency in patients with sepsis. However, the certainty of the evidence ranges from low to very low, emphasising the need for further high-quality RCT to validate these benefits. Also, clinicians should prescribe n-3 supplements cautiously in this regard.

The ICH E9(R1) addendum provides guidelines on accounting for intercurrent events in clinical trials using the estimands framework. However, there has been limited attention to the estimands framework for meta-analysis. Using treatment switching, a well-known intercurrent event that occurs frequently in oncology, we conducted a simulation study to explore the bias introduced by pooling together estimates targeting different estimands in a meta-analysis of randomized clinical trials (RCTs) that allowed treatment switching. We simulated overall survival data of a collection of RCTs that allowed patients in the control group to switch to the intervention treatment after disease progression under fixed effects and random effects models. For each RCT, we calculated effect estimates for a treatment policy estimand that ignored treatment switching, and a hypothetical estimand that accounted for treatment switching either by fitting rank-preserving structural failure time models or by censoring switchers. Then, we performed random effects and fixed effects meta-analyses to pool together RCT effect estimates while varying the proportions of trials providing treatment policy and hypothetical effect estimates. We compared the results of meta-analyses that pooled different types of effect estimates with those that pooled only treatment policy or hypothetical estimates. We found that pooling estimates targeting different estimands results in pooled estimators that do not target any estimand of interest, and that pooling estimates of varying estimands can generate misleading results, even under a random effects model. Adopting the estimands framework for meta-analysis may improve alignment between meta-analytic results and the clinical research question of interest.

This study synthesized 65 (quasi-)experimental studies published between 2010 and 2024 that examined the use of mobile applications to develop language learners’ vocabulary learning. Bayesian meta-analysis was adopted to assess (1) overall effect size; (2) subgroup analyses (i.e. education level, vocabulary knowledge, aspects of vocabulary learning, learning environment, sample size, mobile application type, gender, and cultural background); and (3) publication bias. A large effect size of 1.28 was found for the overall effectiveness of using mobile applications for vocabulary learning when we restricted the studies to long-term treatment duration of 10 weeks or above. Each moderator was analyzed and discussed, and implications for language teaching and research were provided.