Who is Being Screened for Intimate Partner Violence in Primary Care Settings? Secondary Data Analysis of a Cluster Randomised Trial

Objectives To assess sociodemographic differences in postpartum women screened for intimate partner violence and who disclosed to their Maternal and Child Health nurses. Methods Secondary analyses of survey data from women participating in a cluster randomised trial. The trial tested a nurse-designed, enhanced violence screening model-versus routine screening among eight community nurse clinics in Melbourne, Australia. Self-completion anonymous surveys were sent to all clinic attendees who had given birth in the previous eight months. We measured intimate partner violence with the Composite Abuse Scale and other sociodemographic variables. Multivariable logistic regression was used to analyse characteristics of screened versus unscreened women and those who did or did not disclose. Results 91 clinics (163 nurses) participated in the trial. 2621/10,472 (25%) women responded to the survey. Notable characteristics, such as level of intimate partner violence (AdjOR 1.14, CI 0.94-1.40), parity (AdjOR 1.13, CI 0.94-1.35), education (AdjOR 1.20 CI 0.91-1.58) and being born in Australia (AdjOR 0.94, CI 0.86-1.03) made no significant difference to screening. However, nurses were significantly less likely to screen women with a lower income than those with a higher one (AdjOR 0.59, CI 0.40-0.87) with a dose response relationship. Women on the lowest levels of income were significantly more likely to disclose abuse (AdjOR 3.06, CI 1.02-9.17), indicating missed opportunities for nurses to provide timely care. Conclusions for practice Despite being required to screen all women, nurses are almost twice as likely to screen more affluent women, who would be less likely to be experiencing or disclose intimate partner violence.