No Benefit Seen in Routine Screening for Domestic Abuse
By John Gever, Senior Editor, MedPage Today
Published: August 04, 2009
Reviewed by Zalman S. Agus, MD; Emeritus Professor
University of Pennsylvania School of Medicine and
Dorothy Caputo, MA, RN, BC-ADM, CDE, Nurse Planner
Screening all women in medical clinics for intimate partner violence in a randomized trial did not reduce rates of abuse or improve their quality of life, researchers said.
Among those for whom the screening revealed a history of abuse, recurrence of partner violence during 18 months of follow-up was reduced by a nonsignificant 18% (odds ratio 0.82, 95% CI 0.32 to 2.12) compared with control participants, according to Harriet L. MacMillan, MD, of McMaster University in Hamilton, Ontario, and colleagues.
The study randomized more than 6,700 eligible and consenting women at 26 Ontario medical clinics to either screening before their medical evaluation, so that interventions could be recommended to women exposed to partner violence, or to completing the screening questionnaires after the clinic visit.
Writing in the Aug. 5 Journal of the American Medical Association, the researchers found that quality of life scores improved somewhat more at the 18-month follow-up in the screened group -- from 52.1 to 58.5, compared with an increase from 50.6 to 52.7 in the control group (P<0.05). Depression scores were also modestly reduced.
But when MacMillan and colleagues imputted outcomes for the more than 40% of participants in both groups who dropped out of the study during follow-up, these differences shrank to insignificance.
The screened women also showed no reduction in after-effects of domestic violence, such as post-traumatic stress disorder and drug and alcohol problems, compared with the unscreened group.
"These results do not provide sufficient evidence to support universal IPV screening in healthcare settings in the absence of an effective intervention to prevent or reduce intimate partner violence," the researchers concluded.
The trial did not employ a specific intervention for those with positive screening results. Clinicians were informed of available local services for victimized women, but they provided referrals at their own discretion.
In fact, the researchers said, "fewer than half [44%] of the screened-positive women reported having a discussion about violence with their clinician during their visit."
MacMillan and colleagues said it may be argued that even small improvements in outcomes justify screening, particularly as the trial did not identify any specific harms to participants from the screening.
For example, other researchers suggested last year that performing such screens on mothers bringing their children to pediatric clinics would be beneficial. (See Pediatricians Should Screen Moms for Domestic Violence)
But the researchers pointed out that 87% of participants were not exposed to partner violence, suggesting a significant amount of wasted effort in the screening.
They also noted that the screening had a false-positive rate of 11%, requiring additional clinical effort and opportunity costs to identify those with positive results who actually needed no intervention.
In an accompanying editorial, Kathryn E. Moracco, PhD, MPH, of the University of North Carolina in Chapel Hill, N.C., and Thomas B. Cole, MD, MPH, a contributing editor at JAMA, suggested that the effectiveness of medical screening for partner violence can't be properly evaluated until there are clinic-based interventions with proven efficacy.
"Specific interventions to prevent the recurrence of abuse for women at risk of violence should be implemented and rigorously tested, preferably in randomized trials, without further delay," they wrote.
"The results of the [current study] should dispel any illusions that universal screening with passive referrals to community services is an adequate response to violence in intimate relationships," they contended.
In the study, women presenting at 11 emergency departments, 12 family practice centers, and three ob/gyn clinics during an 18-month period beginning in July 2005 were nominally eligible to participate.
Those younger than 18 or older than 64 were excluded, as were women who did not make their own appointments or who did not have a male partner in the past year. Participation was also limited to those who spoke English, were able to be alone, and were not seriously ill.
These exclusions meant that, of more than 120,000 women presenting at these clinics, about 87,000 were ineligible. Another 13,000 either refused to provide information to determine eligibility or were missed by the investigators. About 1,500 more who were eligible declined to participate.
That left about 6,700 to be randomized. The investigators designated certain days as "screening days" during which all eligible and consenting female patients were asked to complete the screening questionnaire before they saw clinicians. An equal number of days provided the nonscreened control, with participants completing the questionnaire after the clinic visit.
The screening instrument was the Woman Abuse Screening Tool, with the Composite Abuse Scale used to determine exposure to violence during follow-up. Other outcomes such as PTSD, depression, and substance abuse were measured with standard instruments such as the Short Form-12 and Short Form-36. Quality of life was evaluated with the World Health Organization Quality of Life (WHOQOL)–Bref questionnaire.
Of 2,733 screened patients, 347 had positive results. A similar proportion of the 2,948 patients in the control group who completed the questionnaires also had positive results.
Those in the unscreened group could also be asked about domestic violence if the clinician chose.
Participants were interviewed every six months afterward through the 18-month follow-up period. Attrition was steady, reaching 43% in the screened group and 41% in the control patients at the end of follow-up.
Growth trajectories were calculated with a logistic model for abuse recurrence and with a linear model for quality of life. A secondary analysis used imputed observations based on partial follow-up data to compensate for the high attrition.
The researchers found that, among the screened-positive women still in the study at 18 months, 8% had called a crisis hotline, while 9% had contacted a counseling or advocacy center for abused women, and 4% had sought refuge at a shelter during the follow-up period. Some 26% had talked with a psychologist or social worker.
These figures were very similar to those in the control group whose screening results were positive.
MacMillan and colleagues said it was notable that absolute improvements in most outcomes were seen in both groups relative to baseline.
They speculated that participation in the study -- being asked about intimate partner violence and its effects -- may have been beneficial in itself. Every participant was given an information card with details on services for abused women in their communities.
The authors also noted that the study began during a period when crime statistics indicated violence was at a high point, and "was likely to decrease over time regardless of any intervention."
Besides the high loss to follow-up, limitations to the study included the reliance on participant self-report for baseline and follow-up data, as well as the carefully controlled trial conditions that would probably not occur in ordinary medical practice.
They also noted that significant costs were involved to support the highly trained research assistants who ensured that women filled out the written screen and that positive results were delivered immediately to the clinicians. In addition, each clinician received specific training.
Also, the study was conducted in Canada where women have universal access to most healthcare services.
Primary source: Journal of the American Medical Association
MacMillan H, et al "Screening for intimate partner violence in health care settings: a randomized trial" JAMA 2009; 302: 493-501.
Additional source: Journal of the American Medical Association
Moracco K, et al., "Preventing intimate partner violence: screening is not enough" JAMA 2009; 302: 568-69.