Polaris Research Identifies Factors for Predicting Likelihood of Revictimization Among Abused Women

In response to the high rates of revictimization among abused women, Polaris Health Directions has added predictive analytics to its outcomes system for domestic violence to help treatment providers gauge the probability of repeat abuse.

To develop this feature, 181 women were assessed for revictimization with the Polaris-DV (domestic violence) system, which has been used in a number of medical settings, domestic violence agencies and Family Justice Centers.

From the data collected, nine symptom and strength factors were identified as significant in determining whether a woman is likely to avoid physical, sexual or psychological revictimization, including the severity of post-traumatic stress disorder, indicators related to a woman’s readiness to change, attachment and relationship factors, her subjective well-being and depression. All were incorporated into the resulting predictive model.

During a follow-up visit 90 days after taking the initial Polaris-DV assessment, about 41 percent reported recurring abuse. Among those women, the predictive model correctly identified 77 percent as being at risk for continued abuse. Of the non-victimized women, 72 percent were correctly identified as being not at risk.

“Revictimization often involves victims engaging in exploitative relationships with family members, employers and others as a way of unconsciously reenacting the psychological terror, power and control these women have encountered in their primary relationships” said Dr. Linda Toche-Manley, vice president of Polaris. “This predictive model emphasizes behaviors and conditions that can be addressed in clinical care, providing caseworkers with case-level data they can act on to help women achieve resilience and improved functioning, and have a better chance of avoiding revictimization.”

To request more information about Polaris-DV and pilot opportunities, send an e-mail to moreinfo@polarishealth.com.


This project was supported by Award Number 2R44HD054079-02 from the National Institute of Child Health and Human Development. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Child Health and Human Development or the National Institutes of Health.