Julian P T Higgins’s research while affiliated with University of Bristol and other places

Background: Public health attempts to prevent obesity in children and young people should aim to minimize health inequalities. Two Cochrane reviews examining interventions aiming to prevent childhood obesity found that interventions promoting (only) physical activity have a small beneficial effect on BMI for people aged 5–18 years, as do interventions promoting physical activity alongside healthy eating for 5–11 year olds. We examined whether the effectiveness of the interventions included in these reviews differed according to eight factors associated with inequity: place, race/ethnicity, occupation, gender/sex, religion, education, socio-economic status, and social capital (the PROGRESS framework). Methods: We collected data on change in BMI (standardized or unstandardized), subgrouped by baseline measures of PROGRESS factors, for intervention and control groups, from trial authors. We calculated the intervention effect per subgroup (mean difference), then contrasted these to estimate interactions between intervention and the baseline factors. We combined interaction estimates for each factor across trials using meta-analyses. Findings: We collected subgrouped data from 81 trials that took place between 2001 and 2020, involving 84,713 participants. We found no substantial differences in effectiveness of interventions for PROGRESS subgroups in most scenarios. However, in the younger age group (5–11 years), the effect of interventions on standardized BMI appeared to be higher in boys (average difference in mean differences 0.03; 95% CI 0.01 to 0.06; 45 studies, n = 44,740), which was consistent in direction with the BMI effect (average difference in mean differences 0.06 kg/m2; 95% CI −0.02 to 0.13; 31 studies, n = 27,083). Interpretation: Our findings suggest that those responsible for public health can promote these beneficial interventions without major concerns about increasing inequalities but should be mindful that these interventions may work better in boys aged 5–11 years than girls. More data are needed, so we encourage future trialists to perform subgroup analyses on PROGRESS factors. Funding: National Institute for Health and Care Research (NIHR).

Background: Reported results of clinical trials assessing higher-dose anticoagulation in patients hospitalized for COVID-19 have been inconsistent. Purpose: To estimate the association of higher- versus lower-dose anticoagulation with clinical outcomes. Data sources: Randomized trials were identified from the World Health Organization's International Clinical Trials Registry Platform and ClinicalTrials.gov with no restriction by trial status or language. Study selection: Eligible randomized trials assigned patients hospitalized for COVID-19 to higher- versus lower-dose anticoagulation. Data extraction: 20 eligible trials provided data in a prospectively agreed format. Two further studies were included based on published data. The primary outcome was all-cause mortality 28 days after randomization. Secondary outcomes were progression to invasive mechanical ventilation or death, thromboembolic events, and major bleeding. Data synthesis: Therapeutic- compared with prophylactic-dose anticoagulation with heparins reduced 28-day mortality (OR, 0.77 [95% CI, 0.64 to 0.93]; I 2 = 29%; 11 trials, 6297 patients, of whom 5456 required low or no oxygen at randomization). The ORs for 28-day mortality were 1.21 (CI, 0.93 to 1.58; I 2 = 0%) for therapeutic-dose compared with intermediate-dose anticoagulation (6 trials, 1803 patients, 843 receiving noninvasive ventilation at randomization) and 0.95 (CI, 0.76 to 1.19; I 2 = 0%; 10 trials, 3897 patients, 2935 receiving no or low oxygen at randomization) for intermediate- versus prophylactic-dose anticoagulation. Treatment effects appeared broadly consistent across predefined patient subgroups, although some analyses were limited in power. Higher- compared with lower-dose anticoagulation was associated with fewer thromboembolic events, but a greater risk for major bleeding. Conclusion: Therapeutic-dose compared with prophylactic-dose anticoagulation reduced 28-day mortality. Mortality was similar for intermediate-dose compared with prophylactic-dose anticoagulation and higher for therapeutic-dose compared with intermediate-dose anticoagulation, although this comparison was not estimated precisely. Primary funding source: No direct funding. (PROSPERO: CRD42020213461).

Background Trace amine-associated receptor 1 (TAAR1) agonists offer a new approach, but there is uncertainty regarding their effects, exact mechanism of action and potential role in treating psychosis. Aims To evaluate the available evidence on TAAR1 agonists in psychosis, using triangulation of the output of living systematic reviews (LSRs) of animal and human studies, and provide recommendations for future research prioritisation. Method This study is part of GALENOS (Global Alliance for Living Evidence on aNxiety, depressiOn and pSychosis). In the triangulation process, a multidisciplinary group of experts, including those with lived experience, met and appraised the first co-produced living systematic reviews from GALENOS, on TAAR1 agonists. Results The animal data suggested a potential antipsychotic effect, as TAAR1 agonists reduced locomotor activity induced by pro-psychotic drug treatment. Human studies showed few differences for ulotaront and ralmitaront compared with placebo in improving overall symptoms in adults with acute schizophrenia (four studies, n = 1291 participants, standardised mean difference (SMD) 0.15, 95% CI −0.05 to 0.34). Large placebo responses were seen in ulotaront phase three trials. Ralmitaront was less efficacious than risperidone (one study, n = 156 participants, SMD = −0.53, 95% CI −0.86 to −0.20). The side-effect profile of TAAR1 agonists was favourable compared with existing antipsychotics. Priorities for future studies included (a) using different animal models of psychosis with greater translational validity; (b) animal and human studies with wider outcomes including cognitive and affective symptoms and (c) mechanistic studies and investigations of other potential applications, such as adjunctive treatments and long-term outcomes. Recommendations for future iterations of the LSRs included (a) meta-analysis of individual human participant data, (b) including studies that used different methodologies and (c) assessing other disorders and symptoms. Conclusions This co-produced, international triangulation examined the available evidence and developed recommendations for future research and clinical applications for TAAR1 agonists in psychosis. Broader challenges included difficulties in assessing the risk of bias, reproducibility, translation and interpretability of animal models to clinical outcomes, and a lack of individual and clinical characteristics in the human data. The research will inform a separate, independent prioritisation process, led by lived experience experts, to prioritise directions for future research.

Immortal time arises when individuals in the analysis are either selected based on post-assignment eligibility criteria or assigned to treatment strategies based on post-eligibility information. Explicit target trial emulation prevents the introduction of immortal time in survival analyses of observational data because it synchronizes eligibility and treatment assignment at the start of follow-up. Describing the structure of the biases that generate immortal time is facilitated by specifying the target trial so that the procedures to determine eligibility and assignment can be appropriately evaluated. Selection based on eligibility criteria applied after treatment assignment at the start of follow-up results in immortal time when the analysis starts the follow-up at assignment. Misclassification of assignment to treatment strategies based on treatment received after the start of follow-up results in immortal time when the treatment strategies are not distinguishable at the start of follow-up. The above selection and misclassification can be represented using causal diagrams. We summarize analytic approaches that prevent immortal time when longitudinal data are available from the time of treatment assignment. The term “immortal time bias” suggests that the source of the bias is the immortal time, but it is selection or misclassification that generates the immortal time, leading to bias.

Network meta‐analysis (NMA) combines evidence from multiple trials to compare the effectiveness of a set of interventions. In many areas of research, interventions are often complex, made up of multiple components or features. This makes it difficult to define a common set of interventions on which to perform the analysis. One approach to this problem is component network meta‐analysis (CNMA) which uses a meta‐regression framework to define each intervention as a subset of components whose individual effects combine additively. In this article, we are motivated by a systematic review of complex interventions to prevent obesity in children. Due to considerable heterogeneity across the trials, these interventions cannot be expressed as a subset of components but instead are coded against a framework of characteristic features. To analyse these data, we develop a bespoke CNMA‐inspired model that allows us to identify the most important features of interventions. We define a meta‐regression model with covariates on three levels: intervention, study, and follow‐up time, as well as flexible interaction terms. By specifying different regression structures for trials with and without a control arm, we relax the assumption from previous CNMA models that a control arm is the absence of intervention components. Furthermore, we derive a correlation structure that accounts for trials with multiple intervention arms and multiple follow‐up times. Although, our model was developed for the specifics of the obesity data set, it has wider applicability to any set of complex interventions that can be coded according to a set of shared features.

Quantitative evidence synthesis methods aim to combine data from multiple medical trials to infer relative effects of different interventions. A challenge arises when trials report continuous outcomes on different measurement scales. To include all evidence in one coherent analysis, we require methods to “map” the outcomes onto a single scale. This is particularly challenging when trials report aggregate rather than individual data. We are motivated by a meta‐analysis of interventions to prevent obesity in children. Trials report aggregate measurements of body mass index (BMI) either expressed as raw values or standardized for age and sex. We develop three methods for mapping between aggregate BMI data using known or estimated relationships between measurements on different scales at the individual level. The first is an analytical method based on the mathematical definitions of z‐scores and percentiles. The other two approaches involve sampling individual participant data on which to perform the conversions. One method is a straightforward sampling routine, while the other involves optimization with respect to the reported outcomes. In contrast to the analytical approach, these methods also have wider applicability for mapping between any pair of measurement scales with known or estimable individual‐level relationships. We verify and contrast our methods using simulation studies and trials from our data set which report outcomes on multiple scales. We find that all methods recreate mean values with reasonable accuracy, but for standard deviations, optimization outperforms the other methods. However, the optimization method is more likely to underestimate standard deviations and is vulnerable to non‐convergence.

Importance Randomized clinical trials (RCTs) are widely regarded as the methodological benchmark for assessing clinical efficacy and safety of health interventions. There is growing interest in using nonrandomized studies to assess efficacy and safety of new drugs. Objective To determine how treatment effects for the same drug compare when evaluated in nonrandomized vs randomized studies. Data Sources Meta-analyses published between 2009 and 2018 were identified in MEDLINE via PubMed and the Cochrane Database of Systematic Reviews. Data analysis was conducted from October 2019 to July 2024. Study Selection Meta-analyses of pharmacological interventions were eligible for inclusion if both randomized and nonrandomized studies contributed to a single meta-analytic estimate. Data Extraction and Synthesis For this meta-analysis using a meta-epidemiological framework, separate summary effect size estimates were calculated for nonrandomized and randomized studies within each meta-analysis using a random-effects model and then these estimates were compared. The reporting of this study followed the Guidelines for Reporting Meta-Epidemiological Methodology Research and relevant portions of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guideline. Main Outcome and Measures The primary outcome was discrepancies in treatment effects obtained from nonrandomized and randomized studies, as measured by the proportion of meta-analyses where the 2 study types disagreed about the direction or magnitude of effect, disagreed beyond chance about the effect size estimate, and the summary ratio of odds ratios (ROR) obtained from nonrandomized vs randomized studies combined across all meta-analyses. Results A total of 346 meta-analyses with 2746 studies were included. Statistical conclusions about drug benefits and harms were different for 130 of 346 meta-analyses (37.6%) when focusing solely on either nonrandomized or randomized studies. Disagreements were beyond chance for 54 meta-analyses (15.6%). Across all meta-analyses, there was no strong evidence of consistent differences in treatment effects obtained from nonrandomized vs randomized studies (summary ROR, 0.95; 95% credible interval [CrI], 0.89-1.02). Compared with experimental nonrandomized studies, randomized studies produced on average a 19% smaller treatment effect (ROR, 0.81; 95% CrI, 0.68-0.97). There was increased heterogeneity in effect size estimates obtained from nonrandomized compared with randomized studies. Conclusions and Relevance In this meta-analysis of treatment effects of pharmacological interventions obtained from randomized and nonrandomized studies, there was no overall difference in effect size estimates between study types on average, but nonrandomized studies both overestimated and underestimated treatment effects observed in randomized studies and introduced additional uncertainty. These findings suggest that relying on nonrandomized studies as substitutes for RCTs may introduce additional uncertainty about the therapeutic effects of new drugs.

Background and aims Chronic non‐cancer pain (CNCP) is one of the most common causes of disability globally. Opioid prescribing to treat CNCP remains widespread, despite limited evidence of long‐term clinical benefit and evidence of harm such as problematic pharmaceutical opioid use (POU) and overdose. The study aimed to measure the prevalence of POU in CNCP patients treated with opioid analgesics. Method A comprehensive systematic literature review and meta‐analysis was undertaken using MEDLINE, Embase and PsycINFO databases from inception to 27 January 2021. We included studies from all settings with participants aged ≥ 12 with non‐cancer pain of ≥ 3 months duration, treated with opioid analgesics. We excluded case–control studies, as they cannot be used to generate prevalence estimates. POU was defined using four categories: dependence and opioid use disorder (D&OUD), signs and symptoms of D&OUD (S&S), aberrant behaviour (AB) and at risk of D&OUD. We used a random‐effects multi‐level meta‐analytical model. We evaluated inconsistency using the I ² statistic and explored heterogeneity using subgroup analyses and meta‐regressions. Results A total of 148 studies were included with > 4.3 million participants; 1% of studies were classified as high risk of bias. The pooled prevalence was 9.3% [95% confidence interval (CI) = 5.7–14.8%; I ² = 99.9%] for D&OUD, 29.6% (95% CI = 22.1–38.3%, I ² = 99.3%) for S&S and 22% (95% CI = 17.4–27.3%, I ² = 99.8%) for AB. The prevalence of those at risk of D&OUD was 12.4% (95% CI = 4.3–30.7%, I ² = 99.6%). Prevalence was affected by study setting, study design and diagnostic tool. Due to the high heterogeneity, the findings should be interpreted with caution. Conclusions Problematic pharmaceutical opioid use appears to be common in chronic pain patients treated with opioid analgesics, with nearly one in 10 experiencing dependence and opioid use disorder, one in three showing signs and symptoms of dependence and opioid use disorder and one in five showing aberrant behaviour.