Respiratory failure & ventilation
Derivation and external validation of predictive models for invasive mechanical ventilation in intensive care unit patients with COVID-19
Authors' conclusion
In the development population of ICU patients with COVID-19, SOFA score, SpO2, and respiratory effort signals predicted the need for IMV better than the ROX index. In the external validation population, although the AUCs did not differ significantly, the accuracy was higher when using SOFA score, SpO2, and respiratory effort signals compared to the ROX index.
Study context
- Population
- Intensive care unit (ICU) patients aged ≥18 years with COVID-19, three hospitals in Rio de Janeiro, Brazil, March 2020–August 2021; excluded invasive mechanical ventilation (IMV) within first 24 h of ICU admission656 screened
- Participants
- 346 patients
- Exposure
- Prognostic models vs rate-OXygenation (ROX) index
- Outcome
- Predictive performance for IMV (external validation population)
Methodological detailsprognostic model derivation, external validation, multicentre
- Prognostic model derivation
- External validation
- Multicentre
Main finding
As reported in the abstract
| Group | AUC | Accuracy |
|---|---|---|
| ROX index | 0.683 | 63% |
| Akaike Information Criteria (AIC) model | 0.703 | 69% |
| Lowest AIC model with significant p values | 0.725 | 79% |
AUC
Accuracy
Observational data: associations, not effects.
Further findings
Predictors of IMV (lowest AIC with significant p values)
| OR (95% CI) | ||
|---|---|---|
| SOFA score | 1.46 (1.07–2.05) | |
| SpO2 | 0.81 (0.72–0.90) | |
| Respiratory effort signals | 9.13 (3.29–28.67) | |
| 0.050.21520Odds ratio (log scale) |
| Outcome | IMV | non-IMV | P value |
|---|---|---|---|
| ROX index at admission, IMV vs non-IMV | 7.3 (interval 5.2–9.8) | 9.6 (interval 6.8–12.9) | P < 0.001 |
IMV required (development cohort)
- Required IMV
- 155 (44.8%)
- Did not require IMV
- 191 (55.2%)
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →