Feltner C, Wallace I, Berkman N, Kistler C, Middleton JC, Barclay C, Higginbotham L, Green JT, Jonas DE. Screening for Intimate Partner Violence, Elder Abuse, and Abuse of Vulnerable Adults: An Evidence Review for the U.S. Preventive Services Task Force. Evidence Synthesis No. 169. AHRQ Publication No. HHSA-290-2012-00015-I. Rockville, MD: Agency for Healthcare Research and Quality; 2018
To systematically review the evidence on screening for intimate partner violence (IPV), elder abuse, and abuse of vulnerable adults for populations and settings relevant to primary care in the United States.
PubMed/MEDLINE, the Cochrane Library, Embase, and trial registries through October 4, 2017; reference lists of retrieved articles; outside experts; reviewers; and active surveillance of literature since August 2018.
Two investigators independently selected English-language studies using a priori criteria. Eligible studies included randomized, controlled trials (RCTs) of screening or treatment for abuse victimization, studies evaluating accuracy of screening tests to detect abuse, and cohort studies with a concurrent control group assessing the harms of screening or treatment for abuse.
One investigator extracted data and a second checked accuracy. Two reviewers independently rated quality for all included studies using predefined criteria.
Thirty studies (14,959 participants) were included. Three RCTs (3,759 participants) compared IPV screening (with brief intervention and information about referral options for screen positive women) with no screening; none found significant improvements in any outcome over 3 to 18 months (e.g., IPV, quality of life, or depression) and two RCTs (1,051 participants) reported no harms associated with screening. Fifteen studies assessed the accuracy of one or more abuse screening tools (1,051 participants); studies reported on different measures (e.g., current, past-year, or lifetime IPV). Nine studies assessed tools to detect any past-year or current IPV in women; for past-year IPV (5 studies; n=6,331), sensitivity of five tools ranged from 65 to 87 percent, and specificity ranged from 80 to 95 percent. The accuracy of five tools (4 studies; n=1,795) for detecting current abuse varied widely; sensitivity ranged from 46 to 94 percent, and specificity ranged from 38 to 95 percent. Eleven RCTs (6,740 participants) evaluated interventions for adult women with screen-detected IPV or who were considered at risk for IPV. Eight reported on the incidence of any category of IPV; two of these (575 participants) found a statistically significant benefit in favor of the intervention, one home visiting intervention (standardized mean difference [SMD] -0.34; 95% CI, -0.59 to -0.08) and one behavioral counseling intervention addressing multiple risk factors (SMD -0.40; 95% CI, -0.68 to -0.12). Of the six other RCTs reporting on measures of any IPV exposure, one home visiting intervention (N=643) found an association with reduced IPV, but differences were not statistically significant (SMD -0.04; 95% CI, -0.23 to 0.14), and five RCTs (7,283 participants) found similar rates of IPV in both groups with no statistically significant differences between groups. Two RCTs (210 participants) reported on subtypes of violence only and found mixed results. One RCT assessing a behavioral counseling intervention targeted at multiple risk factors (IPV, smoking, depression, tobacco exposure) reported on birth outcomes among the subgroup of women who screened positive for IPV at baseline (306 of 1,044 enrolled participants) and found no significant difference between groups in rates of low birth weight neonates (<2,500 g) or preterm birth (<37 weeks) or very low birth weight neonates (<1,500 g); however, significantly fewer women in the intervention group had very preterm neonates (≤33 weeks) (2 vs. 9 women; p=0.03). Five RCTs assessing interventions for women with IPV reported on depression outcomes and found inconsistent results (3 found benefit and 2 did not). Three RCTs (506 participants) measured quality of life, two found no difference between groups on SF-12 scores, and one found mixed results across SF-36 subdomains. No studies evaluated screening for elder abuse or abuse of vulnerable adults. We identified one study assessing a screening tool for elder abuse that had poor accuracy (sensitivity 46% and specificity 73% for detecting physical or verbal abuse). We found no RCTs of treatment specific to populations with elder abuse or abuse in vulnerable adults.
RCTs of IPV screening and treatment interventions were heterogeneous in terms of setting, intervention content, and intensity. We were not able to pool study results due to heterogeneity. Strength of evidence was low or insufficient for benefits of treatment (depending on the outcome); evidence was graded as insufficient for birth outcomes because of imprecision, unknown consistency, few events from one subgroup analysis, and uncertainty about whether results could be attributed to IPV counseling. No studies assessed screening or treatment for elder abuse and abuse of vulnerable adults. Most screening tools were assessed in only one study; several enrolled participants from emergency department settings and may have unclear applicability to primary care settings.
Although available screening tools may reasonably identify women experiencing past 12-month IPV, RCTs of screening in adult women do not show a reduction in IPV exposure or improvement in quality of life over 3 to 18 months. Interventions for women with screen-detected IPV show inconsistent results; limited evidence from some RCTs suggested that home visiting interventions and behavioral counseling interventions that address multiple risk factors may lead to reduced IPV among pregnant or postpartum women. No eligible studies assessed screening or treatment for elder abuse and abuse of vulnerable adults.