Among COVID-19 tests conducted in hospital settings, 13% of tests are estimated have false negative results, which can contribute to nosocomial transmission. Ghazi et al developed a model to predict false negative test results, and the cohort included 31,459 patients at Yale-New Haven Health System who had a COVID-19 test within 96 hours of hospital admission. The multivariable logistic regression model was trained on a subsample of patients who were tested only once. Higher age, black race, lower initial oxygen saturation, higher initial temperature, and lower white blood cell count were the factors most strongly associated with a positive test result among the training cohort. Subsequently, this model was validated using a subsample of the cohort which initially tested negative but was retested within 96 hours of admission. Out of the 3,511 patients in this subsample who were retested, 61 (1.7%) were false negatives (testing positive this time). In an attempt to predict the false negatives in the validation cohort, the model had an area under the receiver-operator characteristic of 0.76 (95% CI 0.70-0.83). The model predicted that 536 patients in the validation cohort were false negatives, which included 35 out of the 61 (57%) actual false negatives—a step in the right direction in hospital retesting.