Abstract:Objective: To create and validate a simple and transferable machine learning model from electronic health record data to accurately predict clinical deterioration in patients with covid-19 across institutions, through use of a novel paradigm for model development and code sharing. Design: Retrospective cohort study. Setting: One US hospital during 2015-21 was used for model training and internal validation. External validation was conducted on patients admitted to hospital with covid-19 at 12 other US medical centers during 2020-21. Participants: 33 119 adults (≥18 years) admitted to hospital with respiratory distress or covid-19. Main outcome measures: An ensemble of linear models was trained on the development cohort to predict a composite outcome of clinical deterioration within the first five days of hospital admission, defined as in-hospital mortality or any of three treatments indicating severe illness: mechanical ventilation, heated high flow nasal cannula, or intravenous vasopressors. The model was based on nine clinical and personal characteristic variables selected from 2686 variables available in the electronic health record. Internal and external validation performance was measured using the area under the receiver operating characteristic curve (AUROC) and the expected calibration error-the difference between predicted risk and actual risk. Potential bed day savings were estimated by calculating how many bed days hospitals could save per patient if low risk patients identified by the model were discharged early. Results: 9291 covid-19 related hospital admissions at 13 medical centers were used for model validation, of which 1510 (16.3%) were related to the primary outcome. When the model was applied to the internal validation cohort, it achieved an AUROC of 0.80 (95% confidence interval 0.77 to 0.84) and an expected calibration error of 0.01 (95% confidence interval 0.00 to 0.02). Performance was consistent when validated in the 12 external medical centers (AUROC range 0.77-0.84), across subgroups of sex, age, race, and ethnicity (AUROC range 0.78-0.84), and across quarters (AUROC range 0.73-0.83). Using the model to triage low risk patients could potentially save up to 7.8 bed days per patient resulting from early discharge. Conclusion: A model to predict clinical deterioration was developed rapidly in response to the covid-19 pandemic at a single hospital, was applied externally without the sharing of data, and performed well across multiple medical centers, patient subgroups, and time periods, showing its potential as a tool for use in optimizing healthcare resources.

Prioritising deteriorating patients using time-to-event analysis: prediction model development and internal–external validation

Extracting Dynamic Information of Temporal Clinical Data to Predict the Outcome in Critically Ill Patients.

Clinician perspectives and recommendations regarding design of clinical prediction models for deteriorating patients in acute care

Improved Sensitivity For Detection Of Clinical Deterioration When Diagnostic Pathology And Patient Trends Are Included In Machine Learning Models

Evaluation of clinical prediction models (part 1): from development to external validation

Leveraging Clinical Time-Series Data for Prediction: A Cautionary Tale

Early identification of patients admitted to hospital for covid-19 at risk of clinical deterioration: model development and multisite external validation study

Patient outcomes following medical emergency team review on general wards: Development of predictive models

Prospective validation of clinical deterioration predictive models prior to intensive care unit transfer among patients admitted to acute care cardiology wards

Prediction of Clinical Deterioration in Hospitalized Adult Patients with Hematologic Malignancies Using a Neural Network Model

A tutorial on evaluating time-varying discrimination accuracy for survival models used in dynamic decision-making

Transformer-based time-to-event prediction for chronic kidney disease deterioration

Minimum sample size calculations for external validation of a clinical prediction model with a time‐to‐event outcome

Development and Validation of a Machine Learning Algorithm Using Clinical Pages to Predict Imminent Clinical Deterioration

Predicting Clinical Deterioration in Hospitals

Detecting Deteriorating Patients in the Hospital: Development and Validation of a Novel Scoring System

Targeted Development and Validation of Clinical Prediction Models in Secondary Care Settings: Opportunities and Challenges for Electronic Health Record Data

Prediction of ICU Patients’ Deterioration Using Machine Learning Techniques

Prediction of Short-Term Mortality in Acute Heart Failure Patients Using Minimal Electronic Health Record Data

A prediction model for prehospital clinical deterioration: The use of early warning scores

Towards Personalised Patient Risk Prediction Using Temporal Hospital Data Trajectories