Prediction of individual COVID-19 diagnosis using baseline demographics and lab data

Abstract The global surge in COVID-19 cases underscores the need for fast, scalable, and reliable testing. Current diagnostic tests are limited by turnaround time, availability, or occasional false findings. Here, we developed a machine learning-based framework predicting individual positive diagnosis relying only on readily-available baseline data, including patient demographics, comorbidities, common lab values. Leveraging cohort of 31,739 adults within an academic health system, trained tested multiple types learning models, achieving area under curve 0.75. Feature importance analyses highlighted serum calcium levels, temperature, age, lymphocyte count, smoking, hemoglobin aspartate aminotransferase oxygen saturation as key predictors. Additionally, single decision tree model that provided operable method stratifying sub-populations. Overall, this study provides proof-of-concept prediction models can be using data. resulting complement existing to enhance screening pandemic containment workflows.