This article was originally published here

General psychiatric hospital. November 2, 2021; 74: 9-17. doi: 10.1016 / j.genhospppsych.2021.10.005. Online ahead of print.


OBJECTIVE: To validate a previously published machine learning model of the risk of delirium in hospitalized patients with coronavirus disease 2019 (COVID-19).

METHOD: Using data from six hospitals in two academic medical networks covering care occurring after initial model development, we calculated the predicted risk of delirium using a developed risk model previously applied to diagnosis. , drugs, laboratory and other clinical features available in the electronic health system. (FSD) upon admission to hospital. We assessed the accuracy of these predictions with respect to subsequent delirium diagnoses during this admission.

RESULTS: Of the 5102 patients in this cohort, 716 (14%) developed delirium. The model’s risk predictions produced a c-index of 0.75 (95% CI, 0.73-0.77) with 27.7% of cases occurring in the top decile of predicted risk scores. The model calibration has been decreased from the initial COVID-19 wave.

CONCLUSION: This EHR delirium risk prediction model, developed in the initial wave of COVID-19 patients, produced consistent discrimination over subsequent larger waves; however, as the composition of cohorts and rates of occurrence of delirium changed, model calibration decreased. These results underscore the importance of benchmarking and the challenge of developing risk models for clinical settings where the standard of care and clinical populations may change.

PMID: 34798580 | DOI: 10.1016 / j.genhospppsych.2021.10.005

Source link