Screening for Interpersonal Violence: Potential Harm of Modifying Tools

Abstract

Research ObjectiveHITS, the Hurt Insult Threat Scream tool, was developed to detect intimate partner violence (IPV) in primary care settings. The Center for Medicare and Medicaid Innovation (CMMI) modified the tool to screen more broadly for interpersonal violence—rather than restricting to IPV—and integrated it into a multidomain social risk screening tool in an ongoing national demonstration project. HITS includes four questions: one on physical abuse and three on verbal abuse. The physical abuse question asks: “How often does anyone, including family and friends, physically hurt you?” The three verbal abuse questions ask about being insulted, threatened, or screamed at, respectively. Questions are scored based on frequency of experiencing abuse (1 = “Never”; 5 = “Frequently”). Scores range from 4 to 20. A scoring system for the original tool was validated in adult female survivors of IPV, with a score ≥11 signaling a safety concern. In the current demonstration, CMMI recommends using the original HITS scoring cutoff, despite having broadened the scope of the questions. The objective of this study was to evaluate the use of the original scoring system for detecting rates of reported physical and verbal interpersonal violence.Study DesignCross‐sectional survey design. Descriptive analyses using chi‐square with two‐sided Fisher's exact.Population StudiedA convenience sample of adult English and/or Spanish speaking/reading patients or adult caregivers of pediatric patients at seven primary care clinics and four emergency departments across nine states.Principal FindingsOf 1014 participants, 66 (6.5%) screened positive for any frequency of physical abuse. Using the recommended score cutoff of ≥11, 12/66 (18.2%) screened positive for a safety concern. Of those with scores <11, 3/54 participants (5.6%) reported experiencing physical abuse “fairly often” or “frequently” and 14/54 (25.9%) “sometimes.” 394/1014 participants (38.9%) screened positive for any frequency of verbal abuse; 344/1014 (33.9%) screened positive for verbal abuse without physical abuse. Eighteen of 394 (4.6%) scored ≥11. Three of 394 (7.6%) had total scores <11 and reported “sometimes” or “often” being threatened with harm, while 286/394 (72.6%) reported “rarely” or “sometimes” being insulted or screamed at.ConclusionsUsing the original scoring criteria for the modified HITS, patients reporting physical violence often did not reach the recommended score cutoff to signal a safety concern to the care team. Patient abuse disclosure that lacks adequate follow‐up may result in or exacerbate patient mistrust in the health care system, in addition to being a safety risk for patients. Given that over 30% of patients in this study reported verbal abuse in the absence of physical abuse, and each question is scored with equal weighting, a more nuanced scoring system may be necessary to identify patients at highest risk.Implications for Policy or PracticeScreening for interpersonal violence should at a minimum identify patients experiencing abuse. Identifying patients at risk for abuse can enable interventions to prevent escalation. The scoring system recommended by the original HITS tool for IPV does not allow health systems to reliably identify patients experiencing/at high risk for interpersonal violence. More stakeholder input is needed on how to best score the modified HITS to improve patient safety.Primary Funding SourceThe Commonwealth Fund.