Postpartum hospital use among survivors of intimate partner violence.

Abstract

More than 1 in 3 individuals who identify as female, experience either intimate partner violence (IPV) or sexual assault during their lifetime, and sexual violence committed by an intimate partner is at its highest during their reproductive years.1 As many as 20% of pregnant individuals may experience IPV, and IPV during pregnancy has been associated with an increased risk for adverse maternal and neonatal outcomes, making pregnant individuals an especially vulnerable population.1 In fact, >50% of pregnancy-associated suicides and >45% of pregnancy-associated homicides are associated with IPV and these often occur during the postpartum period.2 Although >50% of maternal deaths occur postpartum,3 little research has examined whether IPV is associated with markers of postpartum maternal morbidity, including hospital readmission and emergency department (ED) visits.4 In addition, few studies have examined the feasibility of ascertaining IPV at the delivery hospitalization using billing codes. Although the International Classification of Diseases, Tenth Revision (ICD-10) codes include factors related to social determinants of health, ICD-10 codes are largely underutilized for the purpose of understanding risk of disease and adverse outcomes.5 The primary objective of this study was to investigate the association of IPV screening at delivery with the incidence of postpartum hospital use. Another objective was to examine the possibility of using ICD-10 codes at the delivery hospitalization to identify IPV in pregnant individuals. This was a retrospective cohort of birth data linked with inpatient and outpatient hospital claims data, including deliveries of individuals residing in the New York City metropolitan area between 2016 and 2018. Thirty-day hospital use was ascertained by either a readmission or an ED visit within 30 days of discharge. We identified the incidence of IPV from the delivery hospital discharge records using 36 IPV-related ICD-10 codes that we identified in the literature, including those defined for adult psychological and sexual abuse. We estimated the associations between IPV identified during the delivery hospitalization and postpartum hospital use using a multivariable logistic regression and separately adjusting for demographic and structural determinants of health, psychosocial factors, comorbidities, and obstetrical complications. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). This study was approved by our institutional review board. IPV was indicated on the discharge records of 348 individuals (0.11%). As shown in the Table, the overall incidence of ED visits among individuals with an IPV-related diagnosis was 12.9%. The incidence of a postpartum ED visit was significantly higher among individuals with an IPV diagnosis than among those without (odds ratio [OR], 2.8; 95% confidence interval [CI], 2.1-3.9), and this was true after sequentially adjusting for demographic and structural determinants of health (OR, 2.0; 95% CI, 1.4-2.7), comorbidities and pregnancy complications (OR, 1.9; 95% CI, 1.4-2.6), psychosocial factors (OR, 1.5; 95% CI, 1.1-2.0), and obstetrical complications (OR, 1.5; 95% CI, 1.1-2.0). The incidence of either a postpartum ED visit or readmission was also higher among those patients with an IPV-related diagnosis (OR, 2.7; 95% CI, 2.0-3.6). However, there was no significant difference in postpartum readmissions alone among patients with or without an IPV-related diagnosis. This study established that postpartum ED visits are significantly higher among individuals with an IPV-related diagnosis during the delivery hospitalization in a large citywide database, even after adjusting for established risk factors for postpartum ED use. Because ED visits have been identified as a possible marker of maternal morbidity and mortality,4 this finding may suggest that individuals affected by IPV could benefit from screening throughout pregnancy, including during the delivery hospitalization, to prevent adverse postpartum outcomes. However, as established in this study, IPV identified solely by ICD-10 codes during the delivery hospitalization is rare and likely underreported. It is possible that underdetection of IPV is because of insufficient clinician screening, a lack of documentation in the medical records using ICD-10 codes, and the medical status of the pregnant individual at the time of delivery. This finding demonstrates a need to screen and record findings thoroughly during the pregnancy period, including at delivery hospitalization, for any IPV-related diagnoses. A limitation of our data is that we were not able to ascertain hospital use outside of New York City and did not include other time points during an individual's pregnancy. Future research should identify at which time points IPV screening occurs during care of a pregnant individual and whether this may affect postpartum ED visit rates. As a clinical outcome, maternal mortality is preventable and screening for risk factors such as IPV throughout the perinatal period, including at delivery admission and during the postpartum period, is imperative for comprehensive obstetrics care.