S13 Methodological Considerations of Network Meta-Analyses: Improving Evaluation of Comparative Effectiveness and Safety of Advanced Therapies Treating Moderate to Severe Ulcerative Colitis
Bibliographic record
Abstract
Background: The growing number of treatments for moderate to severe ulcerative colitis (UC) necessitates the use of network meta-analyses (NMAs) to better understand the comparative efficacy and safety of treatments not directly compared in clinical trials. However, different methods used across NMAs can make it difficult for reviewers to interpret their results. This work aims to provide an overview of key methodological considerations for conducting and reviewing NMAs in moderate to severe UC and suggests a preferred approach for the identified evidence base by refining methods from previous work. Methods: We reviewed NMAs published or submitted for health technology appraisals (TAs) through the National Institute for Health and Care Excellence (NICE) and identified method variations with implications for NMA results and their interpretation. We compared methodologies of previous NMAs to determine preferred methods to compare outcomes among randomized clinical trials evaluating biological therapies (adalimumab, golimumab, infliximab, ustekinumab, vedolizumab) or oral small molecules (filgotinib, ozanimod, tofacitinib) for adults with moderate to severe UC identified through systematic review. Results: We identified and implemented NMA methods relevant to our identified evidence base in UC based on NICE guidelines (Dias et al., 2011) and methodological decisions made in past TAs (NICE TA547, NICE TA342, NICE TA633, NICE TA828): (1) random-effects models with informative priors (Turner et al., 2015) instead of fixed-effect models; (2) analyzed clinical response and clinical remission jointly instead of as separate, uncorrelated endpoints; (3) adjusted maintenance data from treat-through trial designs to approximate re-randomized trial designs instead of splitting networks by trial design; and (4) compared adverse events (AEs) and serious AEs (SAEs) among both induction and longer-term maintenance phases, drawing conclusions from relative treatment effects instead of surface under the cumulative ranking (SUCRA) values. In this NMA with preferred methods, ozanimod had similar efficacy to all other authorized therapies and was significantly more efficacious than adalimumab for the induction of clinical response and clinical remission in bio-experienced patients. Vedolizumab was significantly superior to adalimumab for induction of clinical response and clinical remission in bio-experienced patients in a previous NMA using a fixed-effect model (ICER 2020), compared with no significant difference in our NMA with a random-effects model. We found no significant differences between any active therapy or placebo in AE or SAE rates for induction or maintenance. Based on SUCRA values, ozanimod ranked 2/12 for SAEs (second “best”) during maintenance, which was a marked improvement from analyses assessing the induction period only. Despite no evidence of actual relative treatment effect differences, this might point to ozanimod’s longer-term safety benefit, which was not captured in a previous NMA that only evaluated short-term safety (Lasa et al., 2022). Conclusions: Comparative results from NMAs can differ based on methods used. We refined our NMA using preferred methods from NICE technical guidelines relevant to trials identified in UC to facilitate a wholistic evaluation of the evidence base. Understanding the impact of different methodological choices on NMA results facilitates better interpretation of the comparative efficacy and safety of treatments to guide clinical decision-making.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.543 | 0.840 |
| Meta-epidemiology (narrow) | 0.004 | 0.003 |
| Meta-epidemiology (broad) | 0.008 | 0.026 |
| Bibliometrics | 0.010 | 0.013 |
| Science and technology studies | 0.002 | 0.004 |
| Scholarly communication | 0.010 | 0.007 |
| Open science | 0.007 | 0.006 |
| Research integrity | 0.007 | 0.005 |
| Insufficient payload (model declined to judge) | 0.012 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".