skip navigation

OffLine Summer 1998

Volume 11 Number 1

According to former Surgeon General C. Everett Koop, the impact of domestic violence on today's society is comparable with that of communicable diseases in the prior two centuries. Often referred to as spousal or partner abuse, domestic violence (DV) accounts for 50 percent of all homicides in women and is estimated to occur in 16 percent of couples each year. Despite these statistics, health care providers are often reluctant to raise the specter of DV, comparing it to "opening Pandora's box." For example, one study found that emergency room physicians failed to ask about DV in 40 percent of their interactions with battered women. Only 2 to 7 percent of primary care patients report physician inquiry about verbal or physical abuse. This gap between the magnitude of the problem and the response of health care providers led policymakers to include the effective identification and management of DV as one of the goals of Healthy People 2000.

Ninety percent of injury-related visits at Group health take place in primary care, making this a logical setting to concentrate efforts to improve the identification and management of domestic violence. To this end, Robert S. Thompson MD, Director of the Department of Preventive Care, and Frederick P. Rivara MD, Director of the Harborview Injury Prevention and Research Center, have headed up a "first-of-its-kind" primary care study examining a systematic longitudinal domestic violence intervention directed at the entire health care team. This issue of OffLine describes the background, content and results of this intervention.

Results are also presented from a study related to childhood maltreatment, a phenomenon affecting nearly 3 million children per year in the United States, according to reports made to child protective services. UW scientist Ed Walker MD, in collaboration with Group Health Center for Health Studies, examined the relationship between childhood maltreatment and adult physical health problems. While the effect of childhood abuse on psychological outcomes has been well-studied, little data exist on its relationship to adult physical health, particularly in a defined HMO population. Walker hypothesized that a history of childhood maltreatment would be associated with higher numbers of distressing physical symptoms, health risk behaviors, and adverse sexual and reproductive outcomes, as well as increased functional disability and elevated health care costs.

Domestic violence and childhood maltreatment both involve the intentional infliction of injury or distress. These events can profoundly impact the victim's health status and quality of life. Thompson's and Walker's research illustrate the potential role of primary care providers in improving these outcomes.

DOMESTIC VIOLENCE

As mentioned above, the purpose of the DV randomized trial was to improve the identification and management of DV in primary care. Five Group Health clinics—two intervention and three control—participated in the the trial. The Intervention was unique in that it was directed at the entire primary care practice teams, including physicians, nurses, physician assistants, medical assistants and receptionists. The rationale was that since all staff interact with patients, all staff should be trained to handle specific roles within the team framework. The intervention was also novel in its intensity and length (one year), reflecting the researchers' belief that one-time trainings and written protocols are insufficient to achieve sustained changed in provider behavior.

Provider survey: The first step in the trial was to assess the baseline knowledge, attitudes, beliefs and self-efficacy regarding DV of the staff at the five participating clinics (N=287). Thompson and colleagues used the survey results to help shape the content of the intervention (described below). According to this baseline survey, a large portion of primary care providers estimated the prevalence of current DV in their practice at less than 1 percent. These estimates conflict with the weight of the scientific evidence pegging the prevalence of current DV in primary care at 5 to 25 percent. Providers' likely underestimation of the magnitude of the problem in their practice is of concern because it suggests they will be less likely to devote valuable clinic time to DV. The survey also revealed that 45 percent of clinicians reported that they seldom or never asked about DV when treating injured patients, even though this situation presents a critical opportunity to diagnose DV. Less than one-quarter (23 percent) of providers felt they had strategies to help victims, such as the ability to make appropriate referrals. Only 39 percent of providers expressed confidence in asking about DV, while 99 percent and 98 percent of clinicians were comfortable asking patients about smoking and alcohol, respectively.

Content of Intervention: The first component of the intervention, two half-day training sessions, featured didactics, role-playing, discussions with former victims, and case presentations designed to correct misconceptions and increase providers' self-efficacy regarding DV. The second part of the intervention involved improvements to the practice infrastructure intended to increase the likelihood of a patient-provider dialogue about DV. Two DV related questions were added to the routine physical exam questionnaire and pamphlets, brochures and provider cue cards about DV were placed in strategic clinic locations. The third component of the intervention was reinforcement. A bimonthly newsletter reiterated messages from the training and reminded providers of available resources. In addition, "designated" leaders were recruited from clinic staff to receive more intensive training about DV and to serve as on-site consultive resources.

Results: The effect of the intervention on providers' knowledge, attitudes and beliefs was assessed by comparing responses to the baseline survey with those from surveys administered at 9 and 21–23 months post-training. Response rates at all 3 time points were 80 percent or higher. Results indicate that intervention providers felt increased self-efficacy in dealing with DV. They reported changes in attitudes, knowledge and beliefs consistent with an increased predisposition to intervention. For example, post-training survey results indicated that providers were less fearful that asking about DV might offend patients. This was the first DV intervention in primary care to demonstrate sustained effectiveness over time.

Did these changes in attitudes, beliefs, and self-efficacy translate into improved and increased action by intervention providers? Researchers reviewed nearly 4,000 medical records at baseline and 12 months later (post-intervention) in order to assess changes in provider rates of: 1)asking about DV; 2)case finding; and 3)documenting of appropriate plans for management, referral and follow-up care. This medical record cohort consisted of a random sample of patients presenting in primary care with injuries, depression, chronic pelvic pain or for a routine examination. (Depression and chronic pelvic pain were among the "red flag" conditions for DV emphasized in training.) Preliminary analysis indicates that, compared to the control providers, the intervention group showed a 3.8 fold increase over baseline rates in asking about DV (absolute change 14 percent) and a 1.4 fold increase in DV case finding (absolute change 6.4/1000). The percent of providers who had well documented management plans for DV cases increased somewhat more in intervention than control clinics.

Thompson concludes that "as a result of the intervention, providers came to believe that DV is a problem commonly seen in medical care, felt more comfortable asking patients about abuse, and believed themselves more competent in knowing what to do". However, he points out that there is more work to be done. Although rates of asking about DV increased dramatically among intervention providers, post-intervention asking rates were still only 20 percent in these sites.

Health care utilization: Yvonne Ulrich, a collaborator of Thompson's, found that annual visit rates among female DV cases in the two years prior to a DV diagnosis were approximately three times those of age-matched controls (21.0 vs 6.7 visits/year). Ulrich concludes that more effective identification and management of DV, in addition to improving victims' quality of life, might have the added benefit of reducing health care costs.

Dissemination: Intervention providers rated most components of the intervention very highly. Based on this positive reception and the encouraging results of the intervention, Thompson and colleagues developed a training package (including a video) to disseminate the message. This "lite" training (2 hours as opposed to the original 8) can be conducted by a member of any primary care team after a brief review of the package. (If you are interested in these materials, contact Lori Flemming at 206-326-2916.) At Group Health, several features of the intervention are likely to see widespread dissemination. Potential next steps are to: 1)add two DV-related questions to all Group Health health questionnaires; 2)hang posters in clinic waiting areas to indicate a willingness to talk; and 3)provide information in clinic bathroom related to community resources and safety plans.

CHILDHOOD MALTREATMENT

In order to examine the effects of childhood maltreatment (sexual, physical or emotional) on adult physical health problems, Ed Walker and colleagues surveyed a random sample of 1,963 Group Health female enrollees aged 18–65. The 1,225 women (62 percent) completing the survey were put into three mutually exclusive groups based on their responses to the Childhood Trauma Questionnaire. Eighteen percent had a childhood history of sexual maltreatment (they could have experienced physical or emotional maltreatment, as well); 25 percent had a childhood history of emotional or physical maltreatment but NOT sexual maltreatment; and the remaining 57 percent had below threshold levels of all forms of childhood maltreatment. In this random sample of adult Group health female enrollees, almost half (43 percent) had a history of some form of childhood maltreatment.

The researchers found that a history of childhood maltreatment was significantly associated with a perception of poorer overall health and greater functional disability. Patients with a history of childhood maltreatment reported higher mean numbers of distressing physical symptoms such as back pain, insomnia and nausea. They were also more likely to have experienced adverse sexual and reproductive health outcomes, such as abortions and having had many lifetime sex partners. Health risk behaviors, including smoking, a sedentary lifestyle, and driving while intoxicated were also more common among women with a history of childhood maltreatment. Women who experienced childhood sexual abuse tended to have worse outcomes than women who reported forms of childhood maltreatment not including sexual abuse. The researchers caution that their study of the relationship between childhood maltreatment and adult physical health status does not prove causation. Other experiences, such as adult rape or domestic violence, could also explain these patterns.

Using Group Health's automated data, Walker and colleagues examined the health care costs incurred by the three groups of women in the five years (1992–1996) before they completed the survey. This was the first study to directly assess the health care costs associated with a history of childhood maltreatment. Women reporting any form of childhood maltreatment had median annual health care costs $92 greater than women with below-threshold levels of abuse or neglect. This annual difference increased to $206 when comparing women with a history of sexual abuse to women with no such history. Women with a history of sexual abuse also had significantly higher primary care and outpatient costs, and emergency room utilization, compared with the non-maltreatment group. Although the absolute cost per woman per year is relatively modest, the economic burden to the HMO is large given the high prevalence of women with these experiences.

Childhood maltreatment, like domestic violence, often goes unmentioned in primary care encounters. Studies have shown that fewer than 10 percent of women have shared these experiences with their physicians, even though primary care physicians provide 60 percent of the mental health care in the U.S. Walker's results show that physical health variables could serve as clues to providers of the existence of early childhood maltreatment. As is the case with domestic violence, enhanced awareness of the prevalence and physical manifestation could lead to increased recognition, better management, and ultimately, improved health status and quality of life for victims.

Top