Tianjing Li

The integrity of evidence synthesis is threatened by problematic randomised controlled trials, where there are serious concerns about the trustworthiness of the data or findings. Such concerns could be due to research misconduct, including fraud, or due to honest critical errors. If these trials are not detected, they may be inadvertently included in systematic reviews and guidelines, potentially distorting their results. To resolve this problem, the INSPECT-SR (Investigating Problematic Clinical Trials in Systematic Reviews) tool has been developed to assess the trustworthiness of randomised controlled trials. This tool will help researchers to identify and exclude problematic trials from systematic reviews. This article describes the development of INSPECT-SR and presents the tool and an associated guidance document.

Defining outcomes completely before conducting clinical trials helps to mitigate reporting biases; however, there is limited guidance to help investigators define outcomes completely. We aimed to develop a structured approach for defining trial outcomes completely and consistently.We reviewed literature, developed preliminary rules for defining outcomes, and refined them iteratively. We randomly selected randomized controlled trials (RCTs) on ClinicalTrials.gov that registered before their start dates and posted results by January 4, 2024. The 225 included RCTs evaluated 3,424 outcomes. Two raters independently applied preliminary rules to define each outcome. When raters encountered outcomes they could not define, we refined the rules. We continued this process until no further changes were needed. We discussed and finalized our approach in a consensus meeting.We define an “outcome” as a value for each participant that will be used in analysis to generate study results. A complete outcome definition includes six elements: outcome domain, specific measurement, specific metric, cutoff, variable type, and timepoint. We developed rules for naming specific measurements for both subjective and objective outcomes. We expanded on prior work by developing more comprehensive categories for specific metrics. We introduced "cutoff" as a distinct element with three subelements. To clarify the boundary between outcome definitions and statistical methods, we replaced a previously described element, "method of aggregation," with "variable type," which refers to whether the value for each individual is continuous or categorical.Trialists and sponsors could use this approach alongside other guidelines to define outcomes in trial registrations, protocols, and result reports.

Importance: Defining outcomes completely in trial registrations limits opportunities for reporting bias. That is, an incomplete outcome definition might be compatible with multiple results, so investigators could conduct many analyses and calculate different effects that would all appear to be consistent with the planned outcome. Objective: To evaluate outcome definitions for registered randomized controlled trials (RCTs) to determine whether they were complete or incomplete.Design: A cross-sectional study of RCTs registered on ClinicalTrials.gov and published in International Committee of Medical Journal Editors (ICMJE) member journals from 1 January to 30 June, 2025. Setting: ClinicalTrials.gov and RCTs published in ICMJE member journals.Participants: Outcomes registered before or soon after participant enrollment began for eligible RCTs.Exposure: This study characterized outcome registration completeness without a formal exposure or intervention.Main Outcomes and Measures: The proportion of outcomes that were defined completely (i.e., included a specific measurement, specific metric, variable type, time point, and cutoff, if applicable). Results were explored by outcome priority (i.e., Primary, Secondary, Other outcomes), trial regulatory status for RCTs governed by the US Food and Drug Administration Amendments Act (i.e., “Applicable Clinical Trials” versus other RCTs), by funder (i.e., Industry, National Institutes of Health, Other governmental organizations, Other), and journal.Results: Among 210 RCTs including 2543 registered outcomes, only 6 RCTs (2.9%) defined all outcomes completely and 326 outcomes (12.8%) were defined completely. Most outcomes defined the specific measurement (72.7%) and metric (85.7%), and the cutoff where applicable (65.0%) but not variable type (22.4%). In exploratory analyses, outcome definitions were more complete for Primary outcomes compared with Secondary and Other outcomes (20.5% versus 12.2% and 9.2%, respectively). Outcomes in Applicable Clinical Trials were better defined than outcomes in other RCTs (15.5% versus 11.0%), and outcomes in RCTs sponsored by Industry or funded by the National Institutes of Health were better defined than outcomes in Other RCTs (17.0% and 19.6% versus 7.4%, respectively). There was no evidence of differences across journals.Conclusions and Relevance: Most RCTs published in ICMJE member journals did not define their outcomes completely in registrations. Complete outcome definitions in trial registrations could prevent reporting biases.

Précis: A minority (13 of 91) of patient-reported outcome measures for adult glaucoma meet rigorous quality standards for development. Others, despite limited evidence describing development, have been validated in glaucoma populations and should be used judiciously. Purpose: To identify, characterize and assess the quality of existing patient-reported outcome measures (PROMs) relevant to adult glaucoma patients, focusing on development and validation. Methods: We searched multiple databases for studies reporting development, validation, adaptation, or extension of vision-related PROMs in adults with glaucoma. Eligible reports were grouped by instrument for assessment using the Consensus-based Standards for the selection of health Measurement Instruments (COSMIN) Risk of Bias checklist’s “Standards for evaluating the quality of PROM development.” Results: Among 91 instruments represented by 140 reports, 57 were developed for use in glaucoma patients; the remainder were validated in glaucoma patients. The instruments vary in length and complexity and evaluate domains within several constructs: vision-related QoL, health-related QoL, glaucoma knowledge, disease symptoms, vision-related activity limitation, treatment preferences, treatment side effects, patient satisfaction, and psychological status. According to COSMIN criteria, “adequate” development was documented for just 17 of 91 instruments, with descriptions of constructs, population(s) and context(s) for use, and published studies in the target population. Among these 17, concept elicitation was described for 13, of which 7 were developed since 2021. Conclusions: Thirteen instruments stand out in rigorous quality analysis based on the COSMIN Risk of Bias checklist. Other instruments without evidence of sound development have nevertheless been rigorously validated. PROMs we commonly use should be understood in the context of what they were designed to do, what they do well, their efficiency and limitations. It behooves us to follow and document best practices for developing and validating novel instruments.

Defining outcomes completely before conducting clinical trials helps to mitigate reporting biases; however, there is limited guidance to help investigators define outcomes completely. We aimed to develop a structured approach for defining trial outcomes completely and consistently.We reviewed literature, developed preliminary rules for defining outcomes, and refined them iteratively. We randomly selected randomized controlled trials (RCTs) on ClinicalTrials.gov that registered before their start dates and posted results by January 4, 2024. The 225 included RCTs evaluated 3,424 outcomes. Two raters independently applied preliminary rules to define each outcome. When raters encountered outcomes they could not define, we refined the rules. We continued this process until no further changes were needed. We discussed and finalized our approach in a consensus meeting.We define an “outcome” as a value for each participant that will be used in analysis to generate study results. A complete outcome definition includes six elements: outcome domain, specific measurement, specific metric, cutoff, variable type, and timepoint. We developed rules for naming specific measurements for both subjective and objective outcomes. We expanded on prior work by developing more comprehensive categories for specific metrics. We introduced "cutoff" as a distinct element with three subelements. To clarify the boundary between outcome definitions and statistical methods, we replaced a previously described element, "method of aggregation," with "variable type," which refers to whether the value for each individual is continuous or categorical.Trialists and sponsors could use this approach alongside other guidelines to define outcomes in trial registrations, protocols, and result reports.

Purpose: To identify, appraise, and synthesize the performance of artificial intelligence–based meibography reading as compared with human graders in diagnosing meibomian gland dysfunction. Methods: We followed Cochrane methodology and reporting guidelines for diagnostic test accuracy reviews. To assess potential risk of bias and applicability, we used a modified Quality Assessment of Diagnostic Accuracy Studies-2 checklist. We applied bivariate logistic models to estimate summary sensitivity and specificity when appropriate and used the GRADE framework to rate the certainty of the evidence. Results: We identified 14 eligible studies involving 5511 predominantly middle-aged participants (average age: 27–55 years) who were primarily female (≥54.5%). A total of 18,926 meibography images were obtained through noncontact infrared (11 studies) or in vivo confocal microscopy (three studies). Two studies reported external validation of deep learning models, 12 reported internally validated models, and one reported both. All but one study had high risk of bias in at least one domain; 12 studies raised high or intermediate concern about applicability. Based on three external evaluations, the summary sensitivity and specificity for diagnosing meibomian gland dysfunction from normal glands were 97.5% (95% confidence interval: 77.5%–99.8%) and 85.5% (95% confidence interval: 47.3%–97.5%). Sources of heterogeneity in internally validated models included study population, case mix, and others. The overall evidence was very low to low certainty because of imprecision, high risk of bias, and concerns about applicability. Conclusions: Artificial intelligence–based meibography grading appears less accurate than human graders. Future studies should adopt rigorous designs, including a more diverse participant pool (or image set), and external validation.

Rapid changes in the legalized cannabis market have led to the predominance of high-concentration delta-9-tetrahydrocannabinol (THC) cannabis products. To systematically review associations of high-concentration THC cannabis products with mental health outcomes. Ovid MEDLINE through May 2025; EMBASE, Allied and Complementary Medicine Database, Cochrane Library, Database of Abstracts of Reviews of Effects, CINAHL, and Toxicology Literature Online through August 2024. Two reviewers independently selected studies with high-concentration THC defined as greater than 5 mg or greater than 10% THC per serving or labeled as "high-potency concentrate," "shatter," or "dab." Outcomes included anxiety, depression, psychosis or schizophrenia, and cannabis use disorder (CUD). Results were categorized by association direction and by study characteristics. Therapeutic studies were defined by use of cannabis to treat medical conditions or symptoms. Ninety-nine studies (221 097 participants) were included: randomized trials (42%), observational studies (47%), and other interventional study designs (11%); more than 95% had moderate or high risk of bias. In studies not testing for therapeutic effects, high-concentration THC products showed consistent unfavorable associations with psychosis or schizophrenia (70%) and CUD (75%). No therapeutic studies reported favorable results for psychosis or schizophrenia. For anxiety and depression, 53% and 41% of nontherapeutic studies, respectively, reported unfavorable associations, especially among healthy populations. Among therapeutic studies, nearly half found benefits for anxiety (47%) and depression (48%), although some also found unfavorable associations (24% and 30%, respectively). Moderate and high risk of bias of individual studies and limited evaluation of contemporary products. High-concentration THC products are associated with unfavorable mental health outcomes, particularly for psychosis or schizophrenia and CUD. There was some low-quality evidence, inconsistent by population, for therapeutic benefits for anxiety and depression. Colorado General Assembly, House Bill 21-1317.

This scoping review represents the initial step in updating the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) extension for network meta-analysis (NMA) to improve the usability and relevance of systematic reviews with NMA for diverse audiences. The update will address gaps in reporting, such as documenting methods for assessing NMA homogeneity and transitivity, defining intervention nodes, and considering advances in statistical modeling. It will align with PRISMA 2020 and incorporate input from diverse knowledge users, including patients and the public. This scoping review included 61 studies, comprising 23 guidance documents and 38 overviews of reviews evaluating the completeness of reporting and methodological quality of published NMAs. We identified 37 additional NMA items for inclusion in the next step—the Delphi survey. Building on a 2014 scoping review by our team, we incorporated recent studies to inform the PRISMA-NMA update, adhering to established standards for guideline development. SUMMARY POINTS Network meta-analyses (NMAs) are increasingly prominent in evidence synthesis, but their methodological transparency and reporting completeness remain inconsistent, which affects the reliability of results. This scoping review highlights the need to update PRISMA-NMA to address evolving methodologies (e.g., component NMA) and reporting challenges to improve accessibility and utility of NMA findings. The review underscores progress and gaps in NMA reporting since the 2015 PRISMA-NMA extension, focusing on synthesis methods, homogeneity, consistency and transitivity assessments, and network geometry. Key recommendations, such as defining pre-specified nodes, specifying statistical methods, and addressing competing interests, were often overlooked in NMA reporting, but addressing these gaps could enhance transparency, reproducibility, and trust in NMA results.

Cataract surgery is one of the most performed surgical procedures worldwide. As a potential complication following cataract surgery, dry eye has the potential to impact visual outcomes, lower patient satisfaction, and be detrimental to quality of life. To evaluate the effect of cataract surgery on dry eye outcomes postoperatively. We searched Ovid MEDLINE and Embase from 01/01/2010 to 16/08/2021 and included observational studies of participants ≥ 18 years old undergoing any cataract surgical procedure. We compared postoperative dry eye outcomes with baseline including Ocular Surface Disease Index (OSDI), tear break up time (TBUT), Schirmer’s I test (ST1), and corneal fluorescein staining (CFS) at short-term (< 1 week) and medium-term (≥ 1 week to 3 months) follow-up. Our search yielded 11,133 records. After title and abstract, and then full text screening, we included 20 studies with 1,694 eyes. There was some evidence indicating a decrease in the TBUT during the short-term (within 1 week) and medium-term (1 week up to 3 months) periods following cataract surgery. There was a considerable degree of heterogeneity between studies across other outcomes. At medium-term follow-up most studies that reported ST1 and CFS showed deterioration of these outcomes but there was conflicting evidence of the effect of cataract surgery on OSDI. The review is limited by variability in follow-up timeframes which were unable capture potential clinical course like peak occurrence and duration. Dry eye may persist up to three months postoperatively following cataract surgery. Further studies are required to determine if dry eye outcomes return to baseline at longer term follow-up.