Dados do Trabalho

Título

Do clinicians detect asynchronies at the bedside? A diagnostic accuracy study in Latin-American ICUs

Objetivo

Our aim is to compare the rate of asynchronies detection by ICU professionals and by a trained expert, to assess the reliability of bedside asynchronies detection by visual inspection of ventilatory curves.

Métodos

This is an ongoing sub study of a multicenter cohort on mechanical ventilation. The ICU professionals of participating centers recorded a video of the ventilator screen during transition from controlled to spontaneous ventilation. The videos were independently analyzed by the ICU professionals and by a trained expert, our gold standard, for the presence of asynchronies. We measured sensitivity, specificity, positive predictive value, and negative predictive value of the ICU professionals’ analysis.

Resultados

A total of 892 videos have been included in the study. The videos sum 20,803 seconds of duration and 9,403 respiratory cycles. The ICU professionals detected asynchronies in 118 videos: 28 with double triggering, 34 with ineffective effort, 38 with delayed cycling, and 27 with other types. The expert identified asynchronies in 202 videos: 50 with double triggering, 77 with ineffective effort, 55 with delayed cycling, and 53 with other types. The detection of asynchronies by ICU professionals had a sensitivity of 37.6%, a specificity of 92.3%, a positive predictive value of 66.0%, and negative predictive value of 83.7%.

Conclusão

Our results reveal a low reliability of bedside asynchronies detection by ICU professionals. Very low sensitivity suggests many asynchronies are not detected at the bedside. To provide better ventilatory support, ICU professionals should be trained to detect asynchronies.