What question did this study set out to answer?

To determine whether historical influenza patient data can serve as an effective proxy to rapidly develop and validate personalized 30-day risk prediction models for adverse COVID-19 outcomes.

January 30, 2022BMC Medical Research MethodologyOpen Access

Seek COVER: using a disease proxy to rapidly develop and validate a personalized risk calculator for COVID-19 outcomes in an international network

View Full Paper

Ask AI

Bookmark

Authors

RWRoss D. WilliamsErasmus MC AMAniek F. MarkusErasmus MC CYCynthia YangErasmus MC

Discussion

Member takes

Implication

Multinational cohort study shows influenza-derived models predict severe outcomes in COVID-19 patients, highlighting the utility of disease proxies for rapid risk stratification during pandemics.

Key Points

To determine whether historical influenza patient data can serve as an effective proxy to rapidly develop and validate personalized 30-day risk prediction models for adverse COVID-19 outcomes.
Analyzed historical data from patients with influenza or flu-like illness across a federated network of 14 electronic medical record and claims databases in 6 countries.
Developed three 9-predictor risk scores (COVER-H for hospitalization, COVER-I for intensive services or death, and COVER-F for fatality) using two-step LASSO regularized logistic regression.
Externally validated the models in 44,507 patients with confirmed or suspected COVID-19 across 5 databases from South Korea, Spain, and the United States.
Identified age, sex, and 7 comorbidities (cancer, chronic obstructive pulmonary disease, diabetes, heart disease, hypertension, hyperlipidemia, kidney disease) as key discriminators for all endpoints.
Demonstrated good discrimination in COVID-19 validation cohorts, yielding AUC ranges of 0.69–0.81 for COVER-H, 0.73–0.91 for COVER-I, and 0.72–0.90 for COVER-F.
Identified variable calibration in COVID-19 cohorts that was poorer than in influenza cohorts, indicating a need for local recalibration prior to clinical deployment.

Cite This Study

Williams et al. (2022) studied this question.

synapsesocial.com/papers/6a9445193fda5c29338c1cfd https://doi.org/10.1186/s12874-022-01505-z

View Full Paper

Ask AI

Bookmark

Implication

Multinational cohort study shows influenza-derived models predict severe outcomes in COVID-19 patients, highlighting the utility of disease proxies for rapid risk stratification during pandemics.

Seek COVER: using a disease proxy to rapidly develop and validate a personalized risk calculator for COVID-19 outcomes in an international network

Discussion

Member takes

Implication

Key Points

Cite This Study

Seek COVER: using a disease proxy to rapidly develop and validate a personalized risk calculator for COVID-19 outcomes in an international network

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider