Emergency consent procedure and intervention effect on mortality in critical care: a meta-epidemiological study of randomized controlled trials

Position du problème et objectif(s) de l’étude

Obtaining informed consent from patients or proxies in critical care randomized controlled trials (RCTs) can be challenging and may delay randomization, potentially affecting intervention efficacy. Research without prior consent (RWPC) procedures are increasingly used to facilitate timely inclusion but their impact on trial outcomes remains uncertain. We aimed to assess whether RWPC procedures are associated with differences in intervention effects on mortality in critical care RCTs.

Matériel et méthodes

We searched PubMed and the Cochrane Database of Systematic Reviews from inception to August 1, 2024. We included meta-analyses of RCTs evaluating therapeutic interventions in critically ill adults, reporting mortality as a primary or secondary outcome. We conducted a meta-epidemiological study using a two-step approach. First, we calculated the ratio of odds ratios (ROR) within each meta-analysis to compare the effect of interventions on mortality between RCTs using RWPC and those using standard consent. Second, we pooled these RORs across meta-analyses using a random-effects model. Secondary outcomes included the delay from eligibility to randomization and the recruitment rate.

Résultats & Discussion

We included 42 meta-analyses comprising 323 RCTs and 103,011 patients, of which 59 RCTs (18%) used a RWPC procedure. Trials using RWPC were more recent (median year: 2015 [2008–2019] vs 2012 [2007–2017]; p<0.01), larger (sample size: 203 [101–605] vs 72 [40–162]; p<0.01), more frequently multicenter (80% vs 43%; p<0.01), and had lower overall risk of bias. There was no significant difference in intervention effect on mortality between trials with and without RWPC (pooled ROR, 1.05 [95% CI 0.83–1.34]; I²=71.7%) (Figure 1). RWPC was associated with shorter time to randomization (3 [1–9] vs 11 [4–23] hours; p<0.01) (Figure 2) and higher recruitment rates (9.6 [4.7–18.7] vs 4.5 [1.9–8.6] patients/month; p=0.01). In this meta-epidemiological study including 42 meta-analyses and 323 RCTs enrolling 103,011 critically ill patients, nearly 20% of trials reported using a research without prior consent (RWPC) procedure. No significant difference in intervention effect on mortality was found between RWPC and non-RWPC trials, including in subgroup analyses by intervention type or clinical condition. RWPC trials were more recent, multicenter, had larger sample sizes, and a lower risk of bias. RWPC use was associated with faster randomization and higher recruitment rates. These trials also showed a slightly higher withdrawal rate (1.3%). Factors like the overall lack of effective interventions, smaller sample sizes, and higher bias in non-RWPC trials may explain the absence of observed differences in intervention effect. Despite substantial heterogeneity across meta-analyses, RWPC appears to improve trial feasibility by better integrating research into clinical workflow. This is especially relevant in acute care where delays in intervention can impact outcomes. While RWPC is generally accepted, post-trial communication remains essential, as many patients remain unaware of their participation. Limitations include incomplete reporting of consent procedures and timing, a focus solely on mortality, and exclusion of pediatric trials, potentially affecting generalizability.

Conclusion

In critical care RCTs, RWPC procedures were not associated with differences in intervention effect on mortality but were linked to shorter time to randomization and higher recruitment rates.

Auteurs

Geoffroy HARIRI (1) , Jacqueline LOUIE (1), Aqsa KHAN (1), Peggy TAHIR (1), Guillaume MARTIN (2), Agnès DECHARTRES (2), Matthieu LEGRAND (1) - (1)Ucsf, San Francisco, États-Unis, (2)Sorbonne Université, Paris, France

Orateur(s)

Geoffroy HARIRI (Paris)