Adverse health behaviors and the detection of partner violence by clinicians.

BACKGROUND Intimate partner violence (IPV) is a major public health problem in the United States, and victims are commonly encountered in medical settings. Many barriers exist to clinician-initiated screening for IPV. However, smoking and problem drinking are conditions that clinicians commonly screen for and both have been strongly associated with IPV in prior studies. By estimating the predicted probability of 12-month and lifetime IPV for a given patient based on whether she presents with these conditions, our study gives clinicians information that can help them identify patients at risk for IPV. METHODS A cross-sectional written patient survey was administered to 2386 female patients at 8 different health care settings in the Greater Boston (Mass) metropolitan area. The probabilities of 12-month and lifetime IPV were estimated based on the women's self-report of smoking and drinking behaviors. RESULTS A woman who neither smoked nor engaged in problem drinking had a 10% probability of IPV in the preceding 12 months and a 39% chance of IPV in her lifetime. Smoking increased the probability to 14% and 49%, respectively. Problem drinking resulted in a doubling of the predicted probability of 12-month IPV to 21%, with a lifetime probability of 43%. When both conditions were present, the effects were additive, with a woman having a 27% probability of experiencing IPV in the preceding 12 months and 54% chance of IPV in her lifetime. CONCLUSIONS The presence of smoking or problem drinking should raise clinicians' suspicion for IPV. This paradigm should not replace direct questioning about IPV but may aid in the detection of abuse in patient populations.