Secondary Prevention of Domestic Violence in HMO Primary Care: Evaluation of Alternative Implementation Strategies

, , ,
The American Journal of Managed Care, October 2004, Volume 10, Issue 10

Objectives: To compare the effectiveness of 2 system-level multifaceted quality improvement approaches to enhancing the secondary prevention of domestic violence (DV) in health maintenance organization (HMO) primary care settings.

Study Design: A "2 new interventions" pre/post design that compared a basic implementation strategy using current staff levels with an augmented basic strategy that added a paid part-time role for medical office social workers as DV social change agents.

Methods: Comparison of pre-/postimplementation changes in clinicians' DV-related knowledge, attitudes, and self-reported practices (KAP) and rates of inquiry about DV with female patients. Study outcomes were measured in surveys of primary care clinicians and women aged 18 to 45 years who had recent appointments with them.

Results: Both strategies resulted in significant improvements in rates of clinician inquiry about DV and in most clinician KAP measures, but only 2 clinician KAP measures showed significantly greater improvement in the augmented strategy arm.

Conclusions: Our finding that substantial sustainable improvement in secondary prevention of DV in primary care is feasible in an HMO primary care setting using existing organizational mechanisms and current staff levels should be encouraging to other such organizations considering undertaking similar initiatives. Our findings support viewing clinician behavior change as a process of organizational cultural change to be undertaken over the long term using multifaceted implementation strategies that employ techniques consistent with accepted theories of behavior change. They also emphasize the need to investigate and address the special barriers to screening that apparently exist in the pediatric setting.

(Am J Manag Care. 2004;10:706-716)

Domestic violence (DV) is a major public health problem1,2 with high healthcare and social costs.3-12 The full cost of DV is difficult to estimate, because even when women seek healthcare for the sequelae of abuse they rarely disclose the abuse unless directly questioned, and clinicians frequently fail to do so.13-15 Nevertheless, most women (including abuse victims) concur that doctors should ask about abuse.13,16,17

A growing understanding of the dynamics of abuse18 and of its prevalence and health costs has led medical organizations to develop clinical practice guidelines for dealing with DV problems. Some guidelines adopt a flexible approach based on "routine inquiry" in specific populations or when patients present with certain symptoms. Others recommend universal screening with standardized screening questions.19-23 Recent initiatives in managed care and other healthcare settings indicate that system-level, multifaceted, clinician behavior-change strategies can significantly increase DV secondary prevention activities.24-26 When healthcare organizations consider such initiatives, they face policy decisions about the level of resource investment required to achieve their goals. Currently, research findings are not available to inform these decisions.

The background for this study was an initiative by a large health maintenance organization' (HMO) to enhance secondary prevention of DV among members. The HMO recruited medical and administrative representatives to a task force on family violence that defined DV-related quality improvement goals, recommended system-level changes to support them, and undertook an organizational cultural change process. In this context, we evaluated 2 multifaceted implementation strategies, each representing a different level of HMO resource investment.

METHODS

Study Site

The study setting was a large, not-for-profit group practice HMO in the northwestern United States.

Study Design

The HMO's decision to undertake its system-level initiative predated the planning of our research project, ruling out a study design with a control group unexposed to the planned initiative. We therefore chose a "2 new interventions" pre/post design27 and compared the basic HMO implementation strategy with a second strategy that augmented the basic strategy with additional resources. The 11 HMO medical offices with departments of internal medicine (IM), family practice (FP), obstetrics/gynecology (OB/gyn), or pediatrics were purposively divided into 2 study arms with approximately equal total numbers of clinicians and relatively balanced numbers from each specialty. Clinicians' exposure to the basic or augmented strategy was determined by the study arm of their medical office. Female patients' exposure status was determined by that of their clinicians.

Implementation Strategies

Basic Implementation Strategy (BIS).

The task force's implementation strategy included writing and disseminating a DV guideline, traditional continuing medical education (CME), and clinical and environmental supports and cues to increase clinician inquiry and patient self-disclosure of DV exposure. The guideline adopted a "routine inquiry" rather than a universal screening approach,28 recommending that primary care clinicians routinely ask about DV exposure of female patients and mothers of pediatric patients at "health maintenance visits" (eg, visits for nonacute care including routine check-ups, routine pregnancy visits, and "well-baby" care) and of all patients whose symptoms suggest abuse. The task force charged local medical office managers with setting up DV response teams (consisting usually of nurses, medical assistants, social workers, and occasionally a female physician) to intervene with identified DV-exposed persons. The task force organized a half-day conference to train DV response team members and other clinicians. Two primary care clinician task force members wrote an article describing the clinician's role in response to DV for the HMO's local medical journal.29 An article describing the signs and dynamics of DV and encouraging HMO members to discuss DV problems with their primary care clinicians appeared in the HMO's member newsletter.30 Plastic dispensers containing "calling cards" with information about community resources for DV victims were placed in all the HMO's restrooms. Printed materials were developed and distributed, including patient brochures and pocket reminders for clinicians about screening, safety assessment, safety planning, and community referral resources. The HMO allotted 4 hours per month to the pediatrician co-chair of the task force to oversee implementation. Medical office managers received modest earmarked funds for materials to support staff training and to purchase Polaroid cameras for documenting abuse. However, the BIS allotted no HMO funds to support medical office staff time dedicated to implementation activities.

Augmented Basic Implementation Strategy (ABIS)

. The ABIS augmented the BIS by giving medical office social workers paid time, funded by the research project, to assume a structured role as DV social change agents31 (5.2 months of full-time employment for the 18-month study period for all the ABIS social workers together). The social workers' role involved (1) conveying information to clinicians about DV prevalence and risk markers, dynamics of abusive relationships, etc; (2) advocating an active primary care clinician role in secondary prevention; (3) elucidating the appropriate goals of screening and intervention activities; and (4) modeling secondary prevention skills (ie, asking patients about DV, danger assessment, documenting abuse, etc). They undertook these activities in department meetings and in individual "academic detailing"-style contacts32 with clinicians. A full description of the ABIS activities can be found at http://www.kpchr.org/public/dv/append1.pdf.

Study Subjects

Clinician Subjects

. Clinician subjects included all IM, FP, health appraisal (HAP), pediatric, and OB/gyn physicians, physician assistants, and nurse practitioners in the HMO's main metropolitan area (n = 273). These clinicians received the baseline survey questionnaire; clinicians in this group who remained in their original study arm during the 10-month implementation period and worked at least 6 months in a single medical office (n = 238) also received the postintervention questionnaire. Pre- and postintervention survey response rates were 74.4% and 68.9%, respectively. In all, 50.2% of the preintervention sample responded to both surveys. Data on these 137 clinicians were used to assess changes in clinician-level outcomes.

Female Patient Subjects

. The initial sampling frame for the independent pre- and postintervention IM, FP, and OB/gyn female patient samples was "health maintenance visits" by women aged 18 to 45 years to clinicians in these specialties during the pre- or postintervention evaluation period (January 29 through April 4, 1996, and February 2 through April 20, 1997, respectively) identified through the HMO's appointment and encounter databases. To create the pediatric visit sampling frames, we identified "well-baby" visits by infants aged 0 to 24 months to the pediatrics department during the same time periods and used the HMO's hospital admissions and discharge database to link these infants to the women aged 18 to 45 years who had given birth to them. We then converted the visit sampling frames to sampling frames of unique women and randomly sampled them according to their most recent visit (one third from IM, FP, and HAP; one third from OB/gyn; and one third from pediatrics). Final sample sizes were n = 1925 and n = 1979 for the pre- and postintervention samples, respectively. Response rates for the pre- and postintervention female patient surveys were 85.8% (n = 1652) and 80.7% (n = 1598), respectively.

Data Sources

Clinician Self-Administered Questionnaires

. Structured questionnaires were mailed to clinician subjects in April through May 1996 (preintervention) and April through May 1997 (postintervention). They elicited data on DV-related knowledge, attitudes, and reported practices (KAP).

Patient Telephone Surveys

. Specially trained interviewers conducted 8- to 10-minute structured telephone interviews with female patient subjects in April through May 1996 (preintervention) and April through May 1997 (postintervention) to obtain data on sociodemographic characteristics, rates of clinician inquiry about DV, rates of DV disclosure, and satisfaction with the HMO's efforts to address DV problems. All study procedures were approved by the organization's human subjects review committee. The content of both questionnaires can be viewed at http://www.kpchr.org/public/dv/append2.pdf.

Study Measures

Primary Outcome Measures

  1. Proportion of women asked about DV exposure at their (or their infant's) recent health maintenance visit. After providing a definition of DV, interviewers asked each respondent whether, at her recent appointment with a specific clinician, anyone had asked her about current or past DV exposure.
  2. Clinicians' DV-related KAP. We based these measures on Prochaska and DiClemente's "transtheoretical model" for modeling lay people's readiness to change their health behaviors,33-35 which has also been adapted to model physicians' readiness to adopt new clinical behaviors,36,37 and on relevant questionnaires.15,38,39 Readiness to change: Clinicians were asked where they fell on Prochaska and DiClemente's "readiness to change" scale33 as adapted by us for DV screening. (See Table 1 for scale categories.) Process of change: a scale of 10 items representing each step in the process of change identified by Prochaska et al.35 Cronbach's alpha40 was .79 and .80 for the pre- and postintervention surveys, respectively. Decisional balance: a 4-item scale measuring advocacy of the "pros" of routine inquiry and a 3-item scale measuring advocacy of its "cons." Alpha for the "pros" measure was .62 for both the pre- and postintervention surveys. For the "cons" measure, alphas were .66 and .68 for the pre- and postintervention surveys, respectively. Self-efficacy: an 8-item scale measuring clinicians' perceptions about their self-efficacy in DV detection and intervention. Alphas were .81 and .83 for the pre- and postintervention surveys, respectively. Frequency of inquiry for "red-flag" symptoms: a 6-item scale measuring the likelihood that clinicians would ask about DV exposure of patients presenting with injuries, chronic pelvic pain, irritable bowel syndrome, headaches, depression/anxiety, and fibromyalgia. Alphas were .90 and .87 for the pre- and postintervention surveys, respectively. General knowledge about DV: Eight questions about general DV knowledge. Knowledge about mandated reporting laws: Four questions about state-mandated reporting laws.
  3. Clinicians' perceptions about medical office social workers as DV experts. This 2-item scale measured clinicians' view of social workers as a resource for help in dealing with DV problems. Pre- and postintervention alphas were .52 and .68, respectively.

Independent Variables

. The primary independent variable was exposure status (BIS or ABIS). Covariates for the patient samples included age, education and income levels, race/ethnicity, and marital, pregnancy, and employment status. For the clinician samples, covariates included age, sex, profession, specialty, years practicing in the specialty, and years working in the HMO.

Statistical Analysis

The analysis focused on detecting whether exposure to the ABIS was associated with significantly greater improvement in study outcomes than exposure to the BIS. For the clinician-level outcomes, we developed multivariate linear regression models to examine, among clinicians who responded to both the pre- and postintervention surveys, the effect of study arm on postintervention KAP scores, while controlling for preintervention scores and relevant covariates.

For patient-level outcomes, we used logistic regression to model postintervention rates of clinician inquiry while controlling for relevant covariates. We merged the pre- and postintervention female patient sample data and created dummy variables that expressed the effect of time (time 1 = preimplementation, time 2 = postimplementation), study arm, and the interaction between the 2. For both clinician- and patient-level outcomes, the final trimmed models maximized the explained variance for each outcome.

The above analyses did not account for the intraclass correlation possible in our nested study design (female patients nested within clinicians nested within specialty departments). Because the presence of significant intraclass correlation would lead to overstatement of the magnitude of any significant intervention effects found in separate analyses of patient outcomes, we verified our finding about the effects of study arm on inquiry rates using multilevel analysis41 among the subsample of 70 clinicians who responded to both the pre- and postintervention surveys and also had visits by female patient respondents to both the pre- and postintervention patient surveys.

RESULTS

Clinician Findings

Personal and professional traits were similar among survey respondents and nonrespondents, with 1 exception: the response rates of female clinicians were significantly higher than those of the male clinicians. Clinicians who responded to both surveys were similar to those who responded to only 1. At baseline, clinicians in the 2 study arms had similar personal and professional characteristics and DV-related KAP.

P

P

Table 1 shows the distribution of study variables in the pre- and postintervention clinician samples. In the preintervention sample three fourths of the clinicians supported the task force's "routine inquiry" approach to DV detection. The proportion of clinicians who were in either the "action" or "maintenance phase" stage of readiness to routinely inquire about DV exposure rose from 16.7% at baseline to 35.9% at follow-up. All mean KAP scores except for general knowledge about DV improved between baseline and follow-up. The improvement in mean scores on readiness to change, process of change, self-efficacy, inquiry for red-flag symptoms, and perceptions about social workers as DV experts was statistically significant at <.001. Improvement in mean scores on knowledge about pros and cons of routine inquiry and mandated reporting requirements was not statistically significant ( <.10). Self-reported DV CME and detection and intervention activities also increased.

Table 2 summarizes the trimmed multivariate models predicting, among clinicians who responded to both surveys, postintervention scores on each KAP measure. Exposure to the ABIS was associated with significantly greater improvement only on process of change and pros of routine inquiry. For all KAP measures, the strongest predictors of postintervention scores were baseline score values. Another noteworthy finding was the effect of clinician specialty on scores for the cons of routine inquiry, self-efficacy, and inquiry for red-flag symptoms: relative to IM clinicians, OB/gyn and HAP clinicians' postintervention scores showed greater improvement, whereas pediatric clinicians showed considerably less improvement.

P

Our multivariate model predicting individual clinicians' postintervention scores on perceptions of the medical office social workers as DV experts indicated that improvement on this measure was strongly associated with exposure to the social workers' social change agent role in the ABIS arm (beta for exposure to the ABIS [vs the BIS] = .85, = .0001 for ABIS exposure and the model).

P

A related finding regarding the social workers' role as social change agents was that, at follow-up, 78.8% of ABIS arm clinicians knew about their medical office's DV response team, compared with only 48.8% of BIS arm clinicians ( = .0001). In addition, clinicians in the ABIS arm were much more likely to have learned about the DV guideline from their medical office social worker than were clinicians in the BIS arm (data not shown).

Female Patient Findings

P

Respondents to the pre- and postintervention surveys were significantly older than nonrespondents ( = .004). In both samples, women in the ABIS arm were significantly older and more educated than women in the BIS arm and had slightly higher income levels. In the preintervention sample, ABIS female patients were more likely to be pregnant and of minority race/ethnicity. At baseline, patient-reported levels of clinician inquiry about DV were similar in the BIS and ABIS study arms.

P

Table 3 shows the distribution of study variables for the pre- and postintervention samples for female patients. At baseline, 2.9% of the respondents reported being asked about DV exposure at their recent health maintenance visit; this figure rose to 9.5% at follow-up ( = .001). The proportion of women who reported discussing a DV problem with a clinician during the medical visit increased from 0.7% (n = 12) at baseline to 1.2% (n = 19) at follow-up, but this was not statistically significant. Women's satisfaction with the HMO's efforts to address DV problems rose dramatically during the study period; the proportion of respondents who were "very" or "somewhat" satisfied rose from 38.5% at baseline to 64.2% at follow-up (Table 3).

Table 4 shows the multivariate logistic regression model predicting the effect of study arm on inquiry rates. The odds ratio for time 2 reflects the highly statistically significant overall increase in inquiry rates during the study period. The odds ratios for the ABIS arm and for the interaction term Time 2 × Study Arm indicate that the ABIS had no greater effect on inquiry rates than the BIS. Rather, inquiry rates were a function of patient characteristics and clinician specialty. Clinician inquiry was more likely among pregnant than nonpregnant women, less likely among college graduates than among less educated patients, and apparently unrelated to women's actual lifetime experience of DV. Internal medicine and pediatric clinicians had particularly poor showings relative to those in other departments. The visibility of printed materials about DV in the medical office and the examination room was strongly associated with higher inquiry rates. Our multilevel analysis of this patient-level outcome variable confirmed the absence of a significant effect of study arm on inquiry rates (results not shown).

DISCUSSION

The "routine inquiry" approach to DV detection and intervention adopted by this HMO is consistent with other DV guidelines and was supported by the clinicians who were expected to carry it out. The task force's overall approach to clinician behavior change—a multifaceted, system-level quality improvement process—has proven effectiveness.24,31 The special contribution of our study is its comparison of the effectiveness of 2 system-level implementation strategies involving different levels of HMO resource investment. Our findings thus have implications for policy decisions of HMOs about the level of resources needed to achieve DV-related organizational goals.

The task force delegated responsibility for implementation to local medical office managers to minimize additions to existing operating budgets. The strengths of such a decentralized approach are the resulting flexibility for managers and medical staff to develop implementation strategies that suit their particular clinical settings and the potential such flexibility creates for the emergence of local initiatives and opinion leaders. Its weakness is that if barriers to the desired change are substantial, the change simply may not happen.

The ABIS addressed potential barriers—especially insufficient staff/resources and absence of staff or patient KAP needed to support the change—by creating a structured, paid role for medical office social workers as DV social change agents. In this instance, the research budget paid the cost of the social workers' role; in principle, however, this additional cost was a substantial HMO resource investment compared with that of the BIS.

Our finding that the BIS alone led to significant improvements in clinicians' DVrelated KAP and inquiry rates over a relatively short period should be encouraging to other healthcare organizations considering undertaking such initiatives—especially because the endeavor occurred in a pressured primary care environment and addressed an issue as challenging as dealing with DV.15 Yet our finding that the added investment in the ABIS did not result in an additional increment in inquiry rates requires explanation.

added

The most likely explanation for the ABIS's lack of effect on most outcomes involves the high priority given to implementing the HMO's DV quality improvement initiative in medical offices in both study arms. The fact that nearly all the medical office administrators, social workers, and DV response team members were women is also probably important. Once given the mandate to address the issue and a suggested implementation approach, these female health professionals energetically did so. In practice, the implementation activities adopted by the BIS medical offices included many elements of the program we developed for the ABIS. However, in carrying out planned DV social change activities, ABIS social workers ran up against the constraints of "real life" in a busy primary care setting. Urgent non-DV problems sometimes diverted time from their paid DV social change role. In particular, the planned academic-detailing—style contacts with clinicians were only partially implemented. Thus, the most likely explanation for the similar level of improvement in outcomes in the 2 study arms is that the actual level of implementation was similar in each.

Given the lack of difference in outcomes between the 2 study arms, our key findings are a more than 3-fold increase in screening rates (from 2.9%-9.5%) and significant improvement in most clinician KAP measures during the 10-month implementation period—a substantial achievement for such a short time. In the absence of other initiatives within the HMO or broader concurrent secular trends during that period to account for these improvements, these changes appear attributable to the intervention activities. However, whether these improvements have been sustained is unknown.

Nevertheless, the task force continues to provide CME conferences on DV problems and has broadened the scope of its in-service training to include emergency room and urgent care providers, dental clinicians, mammography technicians, and newly hired clinicians. The DV response teams receive periodic enrichment training. The Task Force on Family Violence has become part of the HMO's Preventive Care Steering Committee and is presently collaborating with the Depression, Substance Abuse, and Stress Illness Task Forces to develop a coordinated behavioral health-screening tool that clinicians can use to screen simultaneously for all these conditions. When completed, this tool will be incorporated into the HMO's computerized medical record system. The pediatric urgent care clinic in the HMO's hospital has already incorporated routine screening for exposure to all types of violence, including DV.

Thus our study findings and subsequent DV-related developments in this HMO are consistent with those of other researchers who conceive clinician behavior change as a process of organizational cultural change to be undertaken over the long term using multifaceted implementation strategies consistent with accepted theories of behavior change.31 The HMO's continuing commitment to increasing clinicians' sensitivity to and skills for dealing with DV is clear and—in light of evidence that even when organizations undertake intensive efforts to increase DV detection, most women exposed "only" to moderate or low levels of DV will not be detected42—seems as appropriate now as when initially formulated. Furthermore, our findings as well as past experience in this HMO suggest that, over time, the process of organizational cultural change initiated by the task force will increasingly yield positive results. Further progress in DV secondary prevention might be achieved through greater involvement of nonclinician staff (eg, medical assistants) in routine inquiry, and through increasing environmental cues that encourage women to self-disclose abuse.

Screening for DV was especially infrequent among pediatric clinicians, consistent with another recent study that found pediatricians poorly motivated to deal with DV exposure among their pediatric patients' mothers.43 If the presence of children inhibits busy pediatric clinicians from asking mothers about DV exposure, practice routines could be modified to enlist more help from appropriately trained nonclinician staff. In-service training could address the special problems of screening this population for DV exposure.

Our study shared 2 limitations with other similar studies of DV interventions based in healthcare settings.44,45 First, our study focused on process measures of clinician attitude change and rates of routine inquiry about DV exposure rather than on the ultimate goal of detection and effective intervention with abused women. This limitation stemmed from budget and time frame limitations imposed by funding availability, as well as from the small absolute number of women who reported disclosing DV exposure to a clinician. Future evaluations of implementation approaches should use outcome measures in addition to process measures. Second, we did not address the issue of possible harm to women as a result of increased DV detection and intervention efforts. Our finding that female patients' satisfaction with the HMO's efforts to address DV problems increased dramatically during the study offers some reassurance in this regard, but future research needs to address this question explicitly.

Our findings are probably generalizable to other large, highly integrated group- and staff-model HMOs such as the one we studied. McCaw et al24 implemented a similar intervention in another large group-model HMO, as did Thompson et al25 in a large staff-model HMO, with similar results in both cases (although neither of these studies addressed the issue of the resource levels required for effective implementation). Our findings may generalize less well to more loosely integrated managed care organizations.

CONCLUSIONS

This study compared, in the same HMO, alternative, multifaceted, system-level quality improvement approaches to enhancing secondary prevention of DV that involved differing levels of resource investment. We found that, in an HMO with highly motivated medical staff, substantial progress toward improved secondary prevention of DV in the primary care setting can be achieved quickly without major organizational changes or staffing increases. Subsequent DV-related developments in this HMO suggest that this progress is sustainable over time. Our findings confirmed a view of clinician behavior change as a process of organizational cultural change to be undertaken over the long term using a variety of strategies consistent with accepted theories of behavior change. The special barriers to clinician inquiry about DV that apparently exist in the pediatric setting need to be investigated and addressed. Finally, well-funded, long-term, large-scale studies are needed that measure outcomes of detection and intervention and assess their possible negative side effects.

Acknowledgments

We acknowledge with thanks the women HMO members and clinicians who responded to the study surveys. Special thanks go also to social workers Sherry Bohannen, MSW, Laurie Davis, MSW, Steve Jackson, MSW, Susan Kiley, MSW, Monica Kleier, MSW, Linda Knudson, MSV, Susan Pederson, MSW, and Linda Dreyer, MSW. The skilled participation of programmer Kelly Kirk; research analysts Charlotte Corelle, Cacie Cunningham, and Joanne Rizzo; recruitment coordinator Stephanie Hertert; and recruitment lead Cheryl Johnson was integral to the success of the project. Finally, we thank 3 anonymous reviewers of an earlier version of this article.

From the Department of Information and Evaluation, Mental Health Services, Israel Ministry of Health, Jerusalem, Israel (DS); the Department of Pediatrics, East Interstate Medical Office (VF), and the Center for Health Research (JPM), Kaiser Permanente Northwest Division, Portland, Ore; and the Department of Family Medicine, Cascade Park Medical Office, Kaiser Permanente Northwest Region, Vancouver, Wash (CSH).

Financial support for this study was provided by Kaiser Permanente Garfield Memorial Fund grant #101-9062.

Address correspondence to: Diana Shye, PhD, MPH, 7 Kibbutz Galuyot Street, Apt. 17, Jerusalem 93625 Israel. E-mail: dianas@actcom.net.il.

Violence in Families. Assessing Prevention and Treatment Programs

1. Chalk R, King P, eds. . Washington, DC: National Academy Press; 1998.

Full Report of the Prevalence, Incidence and Consequences of Violence Against Women: Research Report.

2. Tjaden P, Thoennes N. Washington, DC: National Institute of Justice; 2000. NIJ publication 183781.

Arch Fam Med.

3. Coker AL, Smith PH, Bethea L, King MR, McKeown RE. Physical health consequences of physical and psychological intimate partner violence. 2000;9:451-457.

Ann Intern Med.

4. Drossman DA, Alley NJ, Leserman J, Olden KW, Barreiro MA. Sexual and physical abuse and gastrointestinal illness. 1995;123:782-794.

Ann Intern Med.

5. McCauley J, Kern DE, Kolodner K, et al. The "battering syndrome": prevalence and clinical characteristics of domestic violence in primary care internal medicine practices. 1995;123:737-746.

Behav Med.

6. Resnick HS, Acierno R, Kilpatrick DG. Health impact of interpersonal violence. 2: medical and mental health outcomes. 1997;23(2):65-78.

JAMA.

7. American Medical Association Council on Scientific Affairs. Adolescents as victims of family violence. 1993;270:1850-1856.

Psychiatr Clin North Am.

8. Campbell JD, Lewandowski LA. Mental and physical health effects of intimate partner violence on women and children. 1997;20:353-374.

Women's Mental Health Services. A Pubic Health Perspective.

9. Newmann JP, Greenley D, Sweeney JK, Van Dien G. Abuse histories, severe mental illness, and the cost of care. In: Levine BL, Blanch AK, Jennings A, eds. Thousand Oaks, Calif: Sage; 1998:279-305.

10. Shye D. Domestic abuse, health status, and HMO health care use. Final report submitted to the US Agency for Healthcare Research and Quality (grant R03 HS0952501); 2001.

J Fam Pract.

11. Wisner CL, Gilmer TP, Saltzman LE, Zink TM. Intimate partner violence against women: do victims cost health plans more? 1999;48:439-443.

Innocenti Digest

12. United Nations Children's Fund (UNICEF). Domestic violence against women and girls. No. 6; 2000:1-29. Available at: http://www.crin.org/docs/resources/treaties/crc.28/Innocenti-E.pdf. Accessed April 28, 2004.

South Med J.

13. Caralis PV, Musialowski R. Women's experiences with domestic violence and their attitudes and expectations regarding medical care of abuse victims. 1997;90:1075-1080.

JAMA.

14. Sugg NK, Inui T. Primary care physicians' response to domestic violence. Opening Pandora's box. 1992;267:3157-3160.

Am J Public Health.

15. Tilden VP, Schmidt TA, Limandri BJ, Chiodo GT, Garland MJ, Loveless PA. Factors that influence clinicians' assessment and management of family violence. 1994;84:628-633.

Arch Intern Med.

16. Friedman LS, Samet JH, Roberts MS, Hudlin M, Hans P. Inquiry about victimization experiences. A survey of patient preferences and physician practices. 1992;152:1186-1190.

J Am Med Womens Assoc.

17. McNutt LA, Carlson BE, Gagen D, Winterbauer N. Reproductive violence screening in primary care: perspectives and experiences of patients and battered women. 1999;54:85-90.

Clin Issues Perinat Womens Health Nurs.

18. Landenburger KM. Exploration of women's identity: clinical approaches with abused women. 1993;4:378-384.

Diagnostic and Treatment Guidelines on Domestic Violence.

19. American Medical Association. Washington, DC: American Medical Association; 1992.

Am Fam Physician.

20. American College of Obstetricians and Gynecologists. ACOG issues technical bulletin on domestic violence. 1995;52:2387-2388, 2391.

Pediatrics.

21. American Academy of Pediatrics Committee on Child Abuse and Neglect. The role of the pediatrician in recognizing and intervening on behalf of abused women. 1998;101:1091-1092.

22. American College of Emergency Physicians. Domestic violence. ACEP policy statement 400286. Approved October 1999. Available at: http://www.acep.org/1,2194,0.html. Accessed February 9, 2004.

23. American Academy of Family Physicians. Violence (position paper). AAFP policy and advocacy statement. Available at: http://www.aafp.org/x7132.xml. Accessed February 9, 2004.

Am J Prev Med.

24. McCaw B, Berman WH, Syme SL, Hunkler EF. Beyond screening for domestic violence: a systems model approach in a managed care setting. 2001;21:170-176.

Am J Prev Med.

25. Thompson RS, Rivara FP, Thompson DC, et al. Identification and management of domestic violence: a randomized trial. 2000;19:253-263.

Am J Prev Med.

26. Waalen J, Goodwin MM, Spitz AM, Petersen R, Saltzman LE. Screening for intimate partner violence by health care providers. Barriers and interventions. 2000;19:230-237.

Health Services Research Methods.

27. Shi L. Albany, NY: Delmar Publishers; 1997.

28. Taket A, Nurse J, Smith K, et al. Routinely asking women about domestic violence in health settings. BMJ. 2003;327:673-676.

Northwest Permanente J Clin Pract.

29. Feldman V, Hokanson C. Asking the question. The clinician's role in domestic violence. 1996;3(2):2-10.

Planning for Health.

30. Kaiser Permanente Northwest Region. Domestic violence. Fall 1996:2-5.

31. Moulding NT, Silagy CA, Weller DP. A framework for effective management of change in clinical practice: dissemination and implementation of clinical practice guidelines. Qual Health Care. 1999;8:177-183.

JAMA.

32. Soumerai SB, Avorn J. Principles of educational outreach ("academic detailing") to improve clinical decision making. 1990;263:549-556.

The Transtheoretical Approach: Crossing Traditional Boundaries of Therapy.

33. Prochaska JO, DiClemente CC. Homewood, Ill: Dow Jones/Irwin; 1984.

Health Educ Q.

34. Prochaska JO, Redding CA, Harlow LL, Rossi JS, Velicer WF. The transtheoretical model of change and HIV prevention: a review. 1994;21:471-486.

35. Prochaska JO, Velicer WF, Rossi JS, et al. Stages of change and decisional balance for 12 problem behaviors. Health Psychol. 1994;13:39-46.

Prev Med.

36. Cohen SJ, Halvorson HW, Gosselink CA. Changing physician behavior to improve disease prevention. 1994;23:284-291.

Am J Prev Med.

37. Main DS, Cohen SJ, DiClemente CC. Measuring physician readiness to change cancer screening: preliminary results. 1995;11:54-58.

Med Care.

38. Ferris LE. Canadian family physicians' and general practitioners' perceptions of their effectiveness in identifying and treating wife abuse. 1994;32:1163-1172.

Violence Vict.

39. Thompson RS, Meyer BA, Smith-DiJulio K, et al. A training program to improve domestic violence identification and management in primary care: preliminary results. 1998;13:395-410.

BMJ.

40. Bland JM, Altman DG. Cronbach's alpha. 1997;314:572.

Introducing Multilevel Modeling.

41. Kreft I, de Leeuw J. Thousand Oaks, Calif: Sage; 1998.

Am J Prev Med.

42. McNutt LA, Carlson BE, Rose IM, Robinson DA. Partner violence intervention in the busy primary care environment. 2002;22:84-91.

Pediatrics.

43. Wright RJ, Wright RO, Isaac NE. Response to battered mothers in the pediatric emergency department: a call for an interdisciplinary approach to family violence. 1997;99:186-192.

JAMA.

44. Wathen CN, MacMillan HL. Interventions for violence against women: scientific review. 2003;289:589-600.

BMJ.

45. Ramsay J, Richardson J, Carter YH, Davison LL, Feder G. Should health professionals screen women for domestic violence? Systematic review. 2002;325:314-326.