EMRs Could Aid Detection of Domestic Abuse Early

Advertisement

Domestic abuse is not always easy to spot except in hindsight or when it becomes severe or chronic. Electronic medical records (EMRs) could aid in detecting domestic abuse early.

Ben Reis, PhD and colleagues have reported the results of their study in September issue of the British Medical Journal. Using a prediction model based on diagnostic history the researchers were able to pick up cases of domestic abuse an average of 10 to 30 months before eventual diagnosis with 88% accuracy.

The American Medical Association and other professional organizations recommend routine screening for abuse. Too often due to the brevity of office visits, actual screening and detection rates remain low. It is often difficult within the time limit of a brief office visit to notice the pattern formed by longitudinal data. This is where the model approach of the EMR is useful.

Advertisement

The researchers analyzed a state-wide claims database over a six year period. The data included more than 550,000 adults with at least four years between their earliest and most recent visits to a hospital or emergency department. By using a narrow case definition of codes that explicitly refer to abuse, 1.04% of patients in the database were classified as victims of domestic abuse.

A broader case definition including the explicit abuse codes as well as those associated with intentional assault and injury suggested 3.44% were domestically abused.

Analysis of the prior-visit diagnoses that predicted high risk of a future abuse diagnosis revealed that injury and psychological health were the most predictive.

The impact of such a model on clinical practice would need further study, but EMRs could be programmed to trigger an alert for further screening of key patients who were domestic abuse victims. Healthcare worker could then do further screening, scheduling a longer office visit.

Source:
Reis BY, et al "Longitudinal histories as predictors of future diagnoses of domestic abuse: modelling study" BMJ 2009; 339: b3677.

Share this content.

If you liked this article and think it may help your friends, consider sharing or tweeting it to your followers.
Advertisement