A Comparison of Relative-Efficacy Estimate(S) Derived From Both Matching-Adjusted Indirect Comparisons and Standard Anchored Indirect Treatment Comparisons: A Review of Matching-Adjusted Indirect Comparisons.

OBJECTIVES: We present an empirical comparison of relative-efficacy estimate(s) from matching-adjusted indirect comparisons (MAICs) with estimates from corresponding standard anchored indirect treatment comparisons. METHODS: A total of 80 comparisons were identified from 17 publications through a systematic rapid review. A standardized metric that used reported relative treatment efficacy estimates and their associated uncertainty was used to compare the methods across different treatment indications and outcome measures. RESULTS: On aggregate, MAICs presented for connected networks tended to report a more favorable relative-efficacy estimate for the treatment for which individual-level patient data were available relative to the reported indirect treatment comparison estimate. CONCLUSIONS: Although we recognize the importance of MAIC and other population adjustment methods in certain situations, we recommend that results from these analyses are interpreted with caution. Researchers and analysts should carefully consider if MAICs are appropriate where presented and whether MAICs would have added value where omitted.

Comparative effectiveness of neutralising monoclonal antibodies in high risk COVID-19 patients: a Bayesian network meta-analysis.

The purpose of this work was to review and synthesise the evidence on the comparative effectiveness of neutralising monoclonal antibody (nMAB) therapies in individuals exposed to or infected with SARS-CoV-2 and at high risk of developing severe COVID-19. Outcomes of interest were mortality, healthcare utilisation, and safety. A rapid systematic review was undertaken to identify and synthesise relevant RCT evidence using a Bayesian Network Meta-Analysis. Relative treatment effects for individual nMABs (compared with placebo and one another) were estimated. Pooled effects for the nMAB class compared with placebo were estimated. Relative effects were combined with baseline natural history models to predict the expected risk reductions per 1000 patients treated. Eight articles investigating four nMABs (bamlanivimab, bamlanivimab/etesevimab, casirivimab/imdevimab, sotrovimab) were identified. All four therapies were associated with a statistically significant reduction in hospitalisation (70-80% reduction in relative risk; absolute reduction of 35-40 hospitalisations per 1000 patients). For mortality, ICU admission, and invasive ventilation, the risk was lower for all nMABs compared with placebo with moderate to high uncertainty due to small event numbers. Rates of serious AEs and infusion reactions were comparable between nMABs and placebo. Pairwise comparisons between nMABs were typically uncertain, with broadly comparable efficacy. In conclusion, nMABs are effective at reducing hospitalisation among infected individuals at high-risk of severe COVID-19, and are likely to reduce mortality, ICU admission, and invasive ventilation rates; the effect on these latter outcomes is more uncertain. Widespread vaccination and the emergence of nMAB-resistant variants make the generalisability of these results to current patient populations difficult.

A text-mining tool generated title-abstract screening workload savings: performance evaluation versus single-human screening.

BACKGROUND AND OBJECTIVES: Text-mining tool, Abstrackr, may potentially reduce the workload burden of title and abstract screening (Stage 1), using screening prioritization and truncation. This study aimed to evaluate the performance of Abstrackr's text-mining functions ('Abstrackr-assisted screening'; screening undertaken by a single-human screener and Abstrackr) vs. Single-human screening. METHODS: A systematic review of treatments for relapsed/refractory diffuse large B cell lymphoma (n = 7,723) was used. Citations, uploaded to Abstrackr, were screened by a human screener until a pre-specified maximum prediction score of 0.39540 was reached. Abstrackr's predictions were compared with the judgments of a second, human screener (who screened all citations in Covidence). The performance metrics were sensitivity, specificity, precision, false negative rate, proportion of relevant citations missed, workload savings, and time savings. RESULTS: Abstrackr reduced Stage 1 workload by 67% (5.4 days), when compared with Single-human screening. Sensitivity was high (91%). The false negative rate at Stage 1 was 9%; however, none of those citations were included following full-text screening. The high proportion of false positives (n = 2,001) resulted in low specificity (72%) and precision (15.5%). CONCLUSION: Abstrackr-assisted screening provided Stage 1 workload savings that did not come at the expense of omitting relevant citations. However, Abstrackr overestimated citation relevance, which may have negative workload implications at full-text screening.