LAUSR

406 www.pccmjournal.org April 2020 • Volume 21 • Number 4 specificity of 99% (95% CI, 95–100%). For the sensitivity of 83%, the 95% CI was quite large (66–93%). The same seems to be observed in the study by de Castro et al (1), which reports excellent specificity (99.3%) with a narrow CI (95% CI, 96– 100%) and a sensitivity of 90.9% with a large CI (95% CI, 58.7–99.8%). Despite having more patients (116), the prevalence was 7% (only eight cases), which raises concern about the true sensitivity value of the test, which is somewhere in the reported CI. The sample was of convenience, and it seems difficult to find a more accurate CI: in a sample calculation, to demonstrate a sensitivity of 90%, with an accuracy of 0.05 and a 95% CI for the prevalence of 7%, 1975 patients would be needed. It is also worthy of attention the number of patients excluded from the study due to the impossibility to validate the examination in 30 minutes, with the neurologist or the psychiatrist, which was five times more than the final number of patients. Other point to clarify and referenced by de Castro et al (1) was how many of the critical ill patients had not consented in participate in the study since frequently these are the patients that have delirium, as observed by Dervan et al (5). Also referred in the study limitations is the age of the patients, as the tool applied just for children older than 5 years, and patients under 2 years old are more susceptible to develop delirium (5). We welcome this publication and look forward for more studies like this, so that in the future a meta-analysis can be made, clarifying the still obscure points of this condition, and helping in the development of useful tools. The authors have disclosed that they do not have any potential conflicts of interest.