Objective This study was aimed to review 4 weeks of universal novel coronavirus disease 2019 (COVID-19) screening among delivery hospitalizations, at two hospitals in March and April 2020 in New York City, to compare outcomes between patients based on COVID-19 status and to determine whether demographic risk factors and symptoms predicted screening positive for COVID-19.
Study Design This retrospective cohort study evaluated all patients admitted for delivery from March 22 to April 18, 2020, at two New York City hospitals. Obstetrical and neonatal outcomes were collected. The relationship between COVID-19 and demographic, clinical, and maternal and neonatal outcome data was evaluated. Demographic data included the number of COVID-19 cases ascertained by ZIP code of residence. Adjusted logistic regression models were performed to determine predictability of demographic risk factors for COVID-19.
Results Of 454 women delivered, 79 (17%) had COVID-19. Of those, 27.9% (n = 22) had symptoms such as cough (13.9%), fever (10.1%), chest pain (5.1%), and myalgia (5.1%). While women with COVID-19 were more likely to live in the ZIP codes quartile with the most cases (47 vs. 41%) and less likely to live in the ZIP code quartile with the fewest cases (6 vs. 14%), these comparisons were not statistically significant (p = 0.18). Women with COVID-19 were less likely to have a vaginal delivery (55.2 vs. 51.9%, p = 0.04) and had a significantly longer postpartum length of stay with cesarean (2.00 vs. 2.67days, p < 0.01). COVID-19 was associated with higher risk for diagnoses of chorioamnionitis and pneumonia and fevers without a focal diagnosis. In adjusted analyses, including demographic factors, logistic regression demonstrated a c-statistic of 0.71 (95% confidence interval [CI]: 0.69, 0.80).
Conclusion COVID-19 symptoms were present in a minority of COVID-19-positive women admitted for delivery. Significant differences in obstetrical outcomes were found. While demographic risk factors demonstrated acceptable discrimination, risk prediction does not capture a significant portion of COVID-19-positive patients.
Key Points
Understanding Demographic Risk Factors for Adverse Outcomes in COVID-19 Patients: Explanation of a Deep Learning Model
