A detailed description of the trial methods and participants was previously published.⁴ All participants provided written informed consent as approved by institutional review boards at the US Centers for Disease Control and Prevention, Cook County Hospital and Health Services, and Rush University.

Trained research assistants recruited adult women from May 2009 until April 2010 from 8 public and 2 private primary health care clinics in Cook County, Illinois. Use of health services from enrollment to 3 years later was the main outcome prespecified in a protocol amendment (Supplement). Participants’ electronic medical records were searched for outpatient care visits, emergency department visits, and hospitalizations.

The mixed-models linear regression command for SPSS version 18 (SPSS Inc) was used to estimate intervention effects on the mean number of visits or hospitalizations while adjusting for age, race/ethnicity, education, type of insurance, and clustering of data by clinic in the overall sample and among the subgroup of women reporting partner violence in the year before enrollment. Results are presented as estimated marginal means. One-tailed significance tests (P<.05) were used.

Of 2708 women randomized, 8 were unenrolled, leaving 2700 women with electronic medical records; 15% reported partner violence in the year before enrollment. Baseline characteristics have been reported, with no significant differences between groups.⁴ The mean (SD) age was 38.7 (14.9) years; 54.9% were black and 36.8% Latina. There were minimal differences between unadjusted and adjusted means so only adjusted estimates are shown in the Table.

For the full sample, adjusted estimates showed no statistically significant differences between study groups in the mean number of hospitalizations (0.2; 95% CI, 0.2–0.3), emergency department visits (0.7; 95% CI, 0.4–0.9), or outpatient care visits (12.2; 95% CI, 10.0–14.4) in the 3 years following enrollment. No differences in these outcomes were found among the subgroup of women who reported experiencing partner violence in the year before enrollment.

Screening women for partner violence and providing a resource list did not influence the number of hospitalizations, emergency department, or outpatient care visits compared with women only receiving a resource list or receiving no intervention over 3 years. Our data do not support providing a partner violence resource list with or without computerized screening of women in urban health care settings to improve health outcomes.

Our trial has the advantages of a large sample, random assignment, a true control group, blinded assessment of outcomes, and 3-year follow-up. Generalizability of the findings are limited by the urban setting; exclusion of participants without telephones, those accompanied by partners or children older than 3 years at the time of their visit, non-English or non-Spanish speaking; and the limited number of college-educated and white, Asian, or Native American participants in the sample. Health visits for participants using health services outside the county system were not captured.

The consistency of the results at 1 year and 3 years contributes to greater confidence in the findings. These null findings are consistent with other trials in primary care settings.⁵ Research should focus on more intensive interventions among women already identified as abused.⁶