Early predictability of the need for tracheotomy after admission to ICU: an observational study

The goal of this study was to explore the ability of professional judgment to predict the need for tracheotomy early among intensive care unit (ICU) patients.

Prospective study using daily questionnaires among ICU physicians in a mixed medical-surgical ICU. The prediction of tracheotomy was by a visual analogue scale (VAS, from 1 to 10, with 1 representing 'absolutely no need for tracheotomy' and 10 representing 'pertinent need for tracheotomy') during ICU stay until tracheal extubation or tracheotomy. For the purpose of this study, a VAS score ≥ 8 was considered a positive prediction for tracheotomy.

A total of 476 questionnaires were retrieved for 75 patients (6.4 ± 5.2 questionnaires per patient), of which 11 patients finally proceeded with a tracheostomy. At first assessment (mean of 2.4 ± 0.8 days after ICU admittance), ICU physicians predicted the need for tracheotomy 3.0 (2.0-6.0) higher VAS points for patients who were finally tracheotomized (P<0.01). Patients with a positive prediction had a 5.4 (1.2-24.1) higher chance of receiving tracheotomy (P=0.03). Considering the median VAS score over a maximum of 10 days before tracheotomy, ICU physicians scored tracheotomized patients significantly higher from day 8 onwards. When comparing ICU physicians, fellows and residents separately, only staff physicians scored a significant difference in the VAS score (P<0.05).

ICU physicians are able to differentiate between patients in need for tracheotomy from those who do not, within 2 days from admittance. The closer the time to the actual intervention, the better the physicians are able to predict this decision.