Effects of allocation concealment and blinding in trials addressing treatments for COVID-19: A methods study
Bibliographic record
Abstract
Abstract Objective Assess the impact of allocation concealment and blinding on the results of trials addressing COVID-19 therapeutics. Data sources World Health Organization (WHO) COVID-19 database and the Living Overview of the Evidence (L-OVE) COVID-19 platform by the Epistemonikos Foundation (up to February 4 th 2022) Methods We included trials that compared drug treatments, antiviral antibodies and cellular therapies with placebo or standard care. For the five most commonly reported outcomes, if sufficient data were available, we performed random-effects meta-regression comparing the results of trials with and without allocation concealment and trials in which both healthcare providers and patients were blinded with trials in which healthcare providers and/or patients were aware of the intervention. A ratio of odds ratios (ROR) > 1 or a difference in mean difference (DMD) > 0 indicates that trials without allocation concealment or open-label trials produced larger effects than trials with allocation concealment or blinded trials. Results As of February 4 th 2022, we have identified 488 trials addressing COVID-19 drug treatments and antiviral antibodies and cellular therapies. Of these, 436 trials reported on one or more of our outcomes of interest and were included in our analyses. We found that trials without allocation concealment probably overestimate mortality (ROR 1.14 [95% CI 0.92 to 1.41]), need for mechanical ventilation (ROR 1.26 [95% CI 0.97 to 1.64]), admission to hospital (ROR 1.93 [95% CI 0.83 to 4.48]), duration of hospitalization (DMD 1.94 [95% CI 0.86 to 3.02]), and duration of mechanical ventilation (DMD 2.64 [95% CI −0.90 to 6.18]), but results were imprecise. We did not find compelling evidence that double-blind and open-label trials produce consistently different results for mortality (ROR 1.00 [95% CI 0.87 to 1.15]), need for mechanical ventilation (ROR 1.03 [95% CI 0.84 to 1.26]), and duration of hospitalization (DMD 0.47 days [95% CI −0.38 to 1.32]). We found that open-label trials may overestimate the beneficial effects of interventions for hospitalizations (ROR 1.87 [95% CI 0.95 to 3.67] and duration of mechanical ventilation (DMD 1.02 days [95% CI −1.30 to 3.35]), but results were imprecise. Conclusion We found compelling evidence that, compared to trials with allocation concealment, trials without allocation concealment may overestimate the beneficial effects of treatments. We did not find evidence that trials without blinding addressing COVID-19 interventions produce consistently different results from trials with blinding. Our results suggest that consideration of blinding status may not be sufficient to judge risk of bias due to imbalances in co-interventions. Evidence users may consider evidence of differences in co-interventions between trial arms when judging the trustworthiness of open-label trials. We suggest, however, evidence users to remain skeptical of trials without allocation concealment. What’s new? key findings Trials without blinding did not produce consistently different results from trials with blinding. Additional information Previous studies have had conflicting results with regards to the effects of blinding on trial results. Our study supports the assertion that results from blinded trials may not differ significantly from unblinded ones. Implications Our study suggest that risk of bias assessment of blinding needs to be more nuanced and that lack of blinding may not be a definite indication of risk of bias.
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 | 0.349 | 0.553 |
| Meta-epidemiology (narrow) | 0.008 | 0.003 |
| Meta-epidemiology (broad) | 0.012 | 0.040 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.005 | 0.007 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.010 | 0.006 |
| Insufficient payload (model declined to judge) | 0.014 | 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".