Estimating the Prevalence of Intimate Partner Violence at an Urban Hospital Before and During the COVID-19 Pandemic Using a Novel Natural Language Processing Algorithm

The impact of COVID-19 on intimate partner violence (IPV) in the United States is still relatively unknown, although some early data demonstrate that cases of IPV increased during COVID-19. The objective of this study was to measure the prevalence of IPV before and during the COVID-19 pandemic in a southeastern urban hospital. We performed a retrospective analysis of IPV encounters at a single high-volume Level I trauma hospital. IPV encounters were identified through a novel natural language processing algorithm using IPV-related words and phrases within unstructured clinical notes. IPV encounters from February to August 2019 (pre-COVID-19 period) were compared with encounters from February to August 2020 (COVID-19 period). The IPV visit rate during the COVID-19 period was higher than that during the pre-COVID-19 period (0.82% of all visits in 2020 vs. 0.72% of all visits in 2019). The number of IPV encounters for patients with no prior IPV visits was higher in 2020, whereas the number of revisits, patients with prior IPV encounters, was lower in 2020. There was an increased incidence of IPV during the COVID-19 pandemic with an increase in the number of patients presenting with first time IPV encounters. Future hospital and community pandemic preparedness protocols must include expansion of screening, resource allocation, and protective policies for those in unsafe situations.

Duke Scholars

Author Rishi Kamaleswaran Trauma, Acute, and Critical Care Surgery