[Practice variation in Medical Decision in Critically Ill Patients During COVID-19 Pandemic in Spain]

Research on clinical decision-making in Intensive Care Units (ICUs) during the COVID-19 pandemic is limited. The objectives of this study are to quantify variation in ICU physicians' decisions in Spain and evaluate its association with contextual factors.

This was a mixed-methods, within-subjects, qualitative-quantitative experimental study using vignettes with manipulation of contextual factors (time of decision, age and pre-existing morbidity, incongruity of diagnostic results, pandemic wave) using an online questionnaire. A total of 146 physicians who worked in the ICU during the pandemic participated. Paired responses in each model were compared using the McNemar test (mid-P), and effect size was estimated using odds-ratios.

Wide practice variation in clinical decision-making was identified, associated with the analysed contextual factors. In model 1, 73% of physicians opted not to intervene at 3:00 AM, but only 14.5% at 10:00 AM. In model 2, most physicians opted to admit the youngest patient with multimorbidity to the ICU during the first wave, but not the older patient with a better pre-existing condition (OR=8.13), an effect that decreased in the last wave (OR=2.42). In model 3, managing patients with COVID-19 despite a negative PCR test was much more frequent in the first wave than in the last (78.8% vs. 56.7%).

High variation in ICU decisions during the pandemic was observed, sensitive to contextual factors such as time of day, age, and the influence of time. This suggests the use of heuristics by physicians and the need to incorporate the contextual component into their training.