Linking Electronic Health Records With Wearable Technology From the All of Us Research Program

Abstract

ObjectiveTo evaluate the feasibility of using electronic health records (EHRs) and wearable data to describe patterns of longitudinal change in day-level heart rate before, during, and after pregnancy and how these patterns differ by age and body mass index. DesignDescriptive secondary analysis feasibility study using data from the National Institutes of Health All of Us Research Program. SettingUnited States. ParticipantsWomen (N = 89) who had a birth or length of gestation code in the EHR and at least 60 days of Fitbit heart rate data during pregnancy. MethodsWe estimated pregnancy-related episodes using EHR codes. Time consisted of five 3-month periods: before pregnancy, first trimester, second trimester, third trimester, and after birth. We analyzed data using descriptive statistics and locally estimated scatterplot smoothing. ResultsAn average of 330 days (SD = 112) of Fitbit heart rate data (29,392 days) were available from participants. During pregnancy, distinct peaks in heart rate occurred during the first trimester (6% increase) and third trimester (15% increase). ConclusionFuture researchers can examine whether longitudinal timing and patterns of heart rate from wearable devices could be leveraged to detect health problems early in pregnancy.