Domestic violence: a national simulation-based educational program to improve physicians’ knowledge, skills and detection rates

Abstract

Background: Although physicians are in a unique position to identify and report domestic violence (DV), detection rates are poor.

Aim: To develop a national DV experiential training program, based on standardized patients (SPs), to improve knowledge, skills and detection rates among physicians.

Methods: The program was initiated by the Israeli Ministry of Health and took place at the Israel Center of Medical Simulation (MSR). Three one-day workshops for physicians were developed, each focusing on intimate partner violence, elder abuse or child abuse. Outcome measures were perceived capabilities, reported case management, and perceived intervention barriers, as obtained by self-assessment questionnaires at baseline and within a follow-up period of six months.

Results: A total of 150 participants took part in 15 workshops. Perception of knowledge and skills, routine screening frequency and reported case management all demonstrated significant improvement. A clear trend to elevation in detection, evaluation and referral rates was found. Ranking of intervention barriers was compared with baseline values and lack of knowledge, lack of skills and psychological difficulties diminished significantly.

Conclusions: An SP-based experiential DV training program for physicians improved perceived capabilities and overall management of DV cases and reduced intervention barriers in a follow-up period of six months.

Practice points

• Domestic violence (DV) educational programs for medical staff are essential to improving currently low detection and reporting rates.

• We present a national DV experiential training program for physicians, based on standardized patients (SPs), re-creating common and relevant clinical scenarios.

• Perceived capabilities, reported case management, and perceived intervention barriers were highly improved in a follow-up period of six months.

Introduction

Domestic violence (DV) is a common clinical and social entity with substantial morbidity and mortality. It is frequently encountered in all healthcare settings, including primary care, emergency medicine (one in four women seeking care), gynecology, pediatrics and psychiatry (Eisenstat & Bancroft 1999). Physicians, worldwide, are uniquely positioned to play a key role in the prevention, detection and management of DV (AMA 1992); nevertheless, diagnosis and report rates are scanty (Rosenblatt et al. 1996). Israel is no different: in a 2001 survey, 11.2% of women reported past physical abuse and 7.6% reported past sexual abuse. According to this survey, 5.8% of the women were victims of moderate to severe physical violence that year. In 2004, 32,737 appeals to sexual assault help centers were registered, compared with 10,842 appeals in 1997. The largest age group seeking help was those aged 13 to 18 years. Again in 2004, in a survey among elders in Haifa, 18.4% reported abuse or neglect in the previous 12 months. In a population of over 6.7 million, these percentages were expected to be translated to many thousands of healthcare visits, yet the Israel Ministry of Health reported only 5106 cases of DV injuries treated in 2004, of which 823 were sexual abuse and merely 200 were elders (Israel Ministry of Health—internal publications). The enormous and inconceivable gap between the abundance of abuse and its miniature detection by the healthcare systems proves the need for an immediate and large-scale national educational program to increase detection rates of DV.

Common intervention barriers reported by physicians are lack of education and training in the field, time and reimbursement constraints and psychological barriers (Sugg & Inui 1992; Lachs 1995; Reid & Glasser 1997). No data from Israel regarding intervention barriers has been published to date. Educational programs targeted to medical staff have proved effective in improving knowledge and skills and increasing detection rates (Jonassen et al. 1999; Coonrod et al. 2000; Berger et al. 2002; Richardson et al. 2002). In past decades, programs based on standardized patients (SP) have increasingly been employed in the teaching and assessment of medical skills (Borrows 1993). Such a model of experiential training befits the subject of DV, due to the complexity and sensitivity required of the physician facing DV victims. Encountering emotion-filled scenarios in a protected setting may help the physician to overcome psychological barriers and obtain the imperative communication skills unique to this field. Indeed, a small number of small-scale SP-based programs for medical students and residents have recently been successfully implemented in this field (Haist et al. 2003; Elman et al. 2004). A national, far-reaching SP-based training pilot program aimed at improving knowledge, skills and detection rates of DV among physicians is presented and assessed.

Methods

The program was initiated by the Israeli Ministry of Health, Social Work Division, and took place at the Israel Center of Medical Simulation (MSR) (Ziv et al. 2005; http://www.msr.co.il/). The target population was physicians, including general practitioners, residents and specialists in relevant primary care fields, from both outpatient and inpatient settings. Physicians were recruited nationwide by all HMOs, mostly on a voluntary basis. Participants were designated as a pilot group for the program, as well as future DV didactic ambassadors in their precincts. Heterogeneity in specialty, HMO affiliation and geographical area were all a prerequisite of the Ministry of Health.

The program included three branches: intimate partner violence, child abuse and elder abuse. All branches shared common educational goals, and differed in unique emphases related to each. Each branch developed an eight-hour workshop, based on SPs. Each workshop was developed by a national committee of DV experts and included eight scenarios reflecting common DV-related encounters with patients and/or family members and care takers. Each physician encountered two scenarios, and actively viewed, via a one-way mirror, four others. All encounters were audio-visually recorded. Encounters lasted 12 minutes each, after which four minutes were allotted to documentation and comments, and another four minutes for a private, undocumented oral feedback by the actor. At two points during the workshop—halfway through and at the end—the participants assembled in a debriefing room and viewed selected segments of recordings from each encounter. Key points from each of the scenarios (content and/or communication skills) were discussed under the instruction of both a physician and a social worker specializing in DV. The instructors were pre-trained in a workshop held at MSR, where they experienced all the encounters and discussed the main educational messages of the workshop in general and of each scenario in particular. Complementary material containing articles, relevant resources list and a PowerPoint presentation of the main points, was handed out to participants and/or made available on the MSR intranet website (MSR.ORG.IL).

Encounter example

An elderly woman presents to your clinic with a superficial laceration, resulting from an attack by her husband. The woman admits to being a victim of longstanding and escalating domestic violence, but belittles the situation and refuses to file a complaint. Points of discussion in this scenario include conveying legislative knowledge (duty to inform aid resources to victims; duty to document DV in medical chart; respect of patient's autonomy if not a minor or legally debilitated), and conveying conversation skills with victims refusing assistance (patient empowerment, condemnation of violence, and providing information regarding DV such as prevalence, escalation with time, morbidity and mortality).

The outcome measures were self-perception of knowledge and skills, reported case management (routine screening rates, actions taken upon encountering suspected DV, reported detection and referral rates), and perceived intervention barriers. These were obtained by two similar self-assessment questionnaires, on the morning of the workshop (baseline questionnaire) and six months after (follow-up questionnaire). Questionnaires were mainly based on a previous questionnaire, which was developed by the second author (DWH) for a regional DV training program of family medicine residents and primary care staffs, and has been in use since 2001. Follow-up questionnaires were obtained by fax or e-mail. Participants who failed to respond were contacted by phone and urged to do so up to three times, as needed. In addition, participants completed a satisfaction survey immediately following the workshop.

Results

Fifteen workshops (equally divided between the three branches) were held between November 2004 and January 2005, including 150 participants. A total of 141 participants completed the baseline questionnaire, of which 74 (52.48%) also completed the follow-up questionnaire approximately 6 months later.

Demographic data

Demographic data were obtained at baseline. The mean age of participants was 46 years (SD 7.7), 63.2% were female, 87% were married, and all but one was a parent. Some 33% of the participants were in the field of pediatrics (all participants in the child-abuse workshops), 26.2% in family medicine, 14.2% in geriatric medicine (all participants in the elder-abuse workshops), 12% general practitioners (GP), 9.2% in internal medicine, 2.2% in gynecology and the rest in other fields; 69.2% were specialists and the rest were residents or general practitioners; 71.6% worked mainly in outpatient settings, while 22% worked mainly in hospitals.

Previous education related to DV

In all, 41.4% (58) physicians reported participation in previous DV educational programs. The extent of those programs was estimated at around two hours in 30% of cases and up to eight hours in 90% of cases. All programs had taken place in the previous five years.

Satisfaction feedback

A satisfaction feedback form was distributed immediately after the workshop. Participants were asked to what degree they were satisfied with key elements of the workshop, on a scale of 1 (not at all) to 4 (very much). The scenario reality, SPs’ quality, and the workshop as a learning experience were rated at 3.81, 3.75 and 3.87, respectively (SD 0.46, 0.54 and 0.36). On the same scale, the mean grade for the question ‘Would you recommend a colleague to participate in the workshop?’ was 3.78 out of 4.00 (SD 0.48), while ‘Would you recommend implementing the workshop in your residency's curriculum?’ ranked 3.79 out of 4.00 (SD 0.49).

Comparison between baseline and follow-up

In order to dismiss alternative interpretations for the potential results, it was essential to assure that there was no preliminary difference between the group that submitted the follow-up questionnaire and the group that did not submit it. Several key parameters were used for comparison: The mean age and all perceived capabilities were submitted to a t-test, while gender, stage of training, main work setting and place of birth were compared by a χ² test of homogeneity. No significant difference between the groups was found in any of these tests. Members of the group that returned the follow-up questionnaires tended to participate more in previous DV educational programs than members of the group that did not return questionnaires (45.9% vs. 36.3%, χ²[1] = 3.18, p = 0.07). Although this difference did not reach statistical significance, it might indicate that members of the group who returned the follow-up questionnaire were generally more exposed to the issue of abuse and therefore collaborated more with the research. However, we do not believe that this difference on its own could compromise any conclusion regarding the before-and-after comparison we planned to carry out. As the group which answered the questionnaire was found on the whole to be representative, comparison between self-assessment questionnaires before and six months after the workshop was analyzed for the 74 participants who answered both.

Perceived capability in diagnostic skills, communication skills, knowledge of favorable intervention, graded on a scale of 1 (not at all capable) to 4 (capable to a large extent), had increased by 0.29 to 0.6. All increments were statistically significant (p < 0.05, Table 1).

Table 1.  Perceived capabilities

In the following questions, please circle what you think to be the appropriate answer on a scale of 1 (not at all) to 4 (to a large extent)
Question	t	n	1	2	3	4	M	Δm	p
To what extent are you certain of your ability to identify a patient as a DV victim?	0 6	73 74	0% 0%	42.3% 18.9%	50.5% 68.9%	7.1% 12.2%	2.6 2.9	0.29	0.005
To what extent do you feel you know enough about DV to enable your intervention in favor of the patient?	0 6	73 74	5.7% 0%	53.1% 32.4%	36.8% 46%	4.4% 21.6%	2.4 2.9	0.49	0.0001>
To what extent do you feel able to discuss DV with your patients?	0 6	73 74	4.4% 1.4%	38.1% 13.5%	40.9% 44.6%	16.6% 40.5%	2.7 3.2	0.56	>0.0001
When encountering DV, to what extent do you know what to do?	0 6	73 74	4.4% 0%	40.9% 10.8%	42.2% 55.4%	12.5% 33.8%	2.6 3.2	0.60	>0.0001

Notes: DV = domestic violence; t = time: 0 = baseline, 6 = follow-up 6 months; M = mean; Δm = difference between time 6 and time 0; p = p-value.

Knowledge of legislation was tested by four true/false questions, pertaining mainly to discrimination between DV cases in which report is mandatory by (Israeli) law and those in which it is not. A mean score for all questions was calculated, allotting one point for each correct answer, range 0–4 points. There was no difference between baseline (1.93) and follow-up score (1.96, t = 0.22, p = 0.82).

Reported case management

Frequency of routine screening of DV (on a scale of 1 = always to 4 = never) has increased (mean score decreased by 0.19, p = 0.03).

Reported actions: Participants were given a list of nine actions, and were asked how often, upon encountering a case suspicious of DV, they take these actions (on a scale of 1 = never, to 4 = always, Table 2). All frequencies of reported actions taken were increased, including documentation of the violence in the medical chart, empowering the patient, providing the patient with relevant information and referring him/her to relevant agencies for treatment. All but one increment were of statistical significance (Table 2). There was no statistical difference in the frequency of scheduling an appointment with the alleged aggressor, which is not currently a recommended action.

Table 2.  Reported actions

When encountering a case suspicious of DV, do you take the following actions? Please answer on a scale of 1 (never) to 4 (always)
Action	t	n	1	2	3	4	M	Δm	p
Document the patient's remarks in his/her own words in the medical chart	0	73	9.6%	28.8%	21.9%	39.7%	2.92	0.52	0.0001
	6	74	4.1%	10.8%	21.6%	63.5%	3.45
Document suspicion of violence even if the patient denies it	0	73	11.0%	38.4%	21.9%	28.8%	2.68	0.55	0.0001
	6	73	4.1%	13.7%	35.6%	46.6%	3.25
Explain to the patient that violence against him/her is unjustified	0	70	10.0%	12.9%	30.0%	47.1%	3.14	0.43	0.0001>
	6	73	2.7%	8.2%	23.3%	65.8%	3.52
Provide information regarding DV (such as prevalence, escalation with time)	0	71	11.3%	45.1%	28.2%	15.5%	2.48	0.67	0.0396
Invite the alleged aggressor for a meeting	0	69	44.9%	24.6%	17.4%	13.1%	1.99	0.12	0.32
	6	72	38.9%	27.8%	18.1%	15.8%	2.1
Inquire about the danger of violence against other family members	0	70	10.0%	32.9%	34.3%	22.8%	2.70	0.36	0.0071
	6	73	9.6%	15.1%	34.3%	41.1%	3.07
	6	70	8.6%	30.0%	31.4%	30.0%	3.11
Check immediate risk and/or work with the patient on a safety plan	0	70	15.3%	36.1%	25.0%	23.6%	2.57	0.58	>0.0001
	6	73	6.8%	16.4%	28.8%	46%	3.18
Offer the patient appropriate help (such as social hospitalization,	0	73	4.1%	19.2%	31.5%	45.2%	3.18	0.42	0.0005
legal aid, social services)	6	72	2.8%	2.8%	22.2%	72.2%	3.64
Inform another professional (staff member/social agent/legal agent)	0	73	4.1%	11.0%	34.2%	50.7%	3.32	0.42	0.0013
	6	73	8.2%	15.1%	75.3%	1.4%	3.75

Diagnosis and referral

At baseline, 74 participants reported having encountered altogether 134 cases suspicious of DV in the preceding six months. In 102 cases (74%), the physician further evaluated the suspicion with a detailed history taking and/or physical exam. Seventy-four cases (54%) were referred to a relevant agency (social worker, police, legal aid, etc.). At follow-up, the same 74 participants reported having suspected 174 cases of DV in the period of six months between the workshop and the follow-up questionnaire. In 155 cases (89%) evaluation was continued, and 116 cases (66%) were further referred to relevant agencies as described above. Some 20% of participants at baseline and 25% at follow-up did not come across a single case that aroused their suspicion of DV.

Perceived intervention barriers

The participants were given a list of potential perceived barriers to interventions in previous cases of DV, and were asked to grade them from 1 (did not prevent my intervention at all) to 4 (had strongly prevented my intervention). At baseline, the three highest barriers cited were lack of knowledge (2.2), time and privacy constraints (2.0) and lack of communication skills (‘I don’t know how to ask’, 1.83). Attitude (‘it's none of my business’) was the least perceived barrier (1.26). This ranking matches the data obtained from the entire baseline group (n = 138). At follow-up, lack of knowledge and lack of communication skills, as well as unfamiliarity with support systems (‘I don’t know where to refer’) and psychological difficulties (‘I am afraid it will find it difficult to cope emotionally’) all received significantly lower scores (Table 3), which indicates an improvement in the physicians’ attitudes regarding these barriers.

Table 3.  Perceived intervention barriers

To what extent each of the following has prevented (or may prevent) your intervention in cases suspicious of DV? Please grade on a scale of 1 (will not prevent my intervention at all) to 4 (will strongly prevent my intervention*)
Barrier	t	n	1	2	3	4	mean	Δm	p
I don’t have enough knowledge on the subject	0	70	18.6%	52.9%	18.6%	10.0%	2.2
	6	72	41.7%	31.9%	15.3%	11.1%	1.96	−0.32	0.027
I don’t have the appropriate conditions (time, privacy)	0	72	38.9%	30.6%	22.2%	8.3%	2.00
	6	71	49.3%	31%	14.1%	5.6%	1.76	−0.23	0.142
I’m afraid to hurt the patient's or his family's feelings if I ask	0	72	34.7%	47.2%	13.9%	4.2%	1.86
	6	72	40.3%	37.5%	16.7%	5.6%	1.88	−0.06	0.641
It's a private matter—it's none of my business	0	71	74.3%	16.2%	5.4%	4.1%	1.28
	6	72	81.9%	12.5%	2.8%	2.8%	1.26	−0.06	0.541
One can’t help in these cases anyway	0	72	65.3%	23.6%	8.3%	2.8%	1.49
	6	71	77.5%	12.7%	8.5%	1.4%	1.34	−0.17	0.06
I don’t know where to refer the patient	0	72	58.3%	23.6%	15.3%	2.8%	1.63
	6	72	79.2%	11.1%	4.2%	5.5%	1.36	−0.32	0.008
I don’t know how to ask about it	0	72	37.5%	43.1%	18.1%	1.4%	1.83
	6	71	71.8%	16.9%	5.6%	5.6%	1.45	−0.36	0.006
I’m afraid I’ll find it difficult to cope emotionally	0	72	58.3%	26.4%	15.3%	0%	1.57
	6	70	72.9%	18.6%	5.7%	2.8%	1.39	−0.24	0.007
I’m afraid of violence against me	0	72	55.6%	29.2%	15.3%	0%	1.60
	6	70	61.4%	24.3%	10%	4.3%	1.57	−0.10	0.33
Negative professional experience in this field	0	68	60.9%	33.3%	1.4%	4.3%	1.49
(previous intervention has led to unwanted results)	6	69	62.3%	26.1%	7.2%	4.4%	1.54	−0.05	0.704
Cultural barriers (language, mentality)	0	71	43.7%	36.6%	15.5%	4.2%	1.80
	6	69	52.2	26.1%	14.5%	7.2%	1.76	−0.07	0.545

Note: *Lower scores indicate improvement (see text).

Discussion

Perception of knowledge, perception of skills, frequency of routine screening and nearly all required actions reportedly taken upon encountering DV cases have all improved with statistical significance. A clear trend to elevation in detection, evaluation and referral rates was found. Grading of intervention barriers was obtained at baseline, of which knowledge, skills and psychological difficulties have significantly improved following the workshops. Feedback obtained immediately after the workshop demonstrated high satisfaction rates.

Self-assessment as the main tool of evaluation is subjective and may be contaminated by recall bias. The ministry's prerequisite of a nationwide project, encompassing all HMOs and regions, made objective data collection highly complicated. We tried to minimize this limitation by choosing a follow-up time far enough on to show a difference, yet short enough for recollection. An extensive search of the literature did not provide an appropriate formally validated questionnaire, but we did base our questionnaire on one that is relevant to our program and has been operational for over three years.

A relatively low response rate to the follow-up questionnaire (52.48%) was disappointing. However, we had shown the 74 participants from whom full data was obtained to be similar to the large baseline group in all parameters except for participation in previous DV education programs. This parameter may indicate enthusiasm and interest in the subject, as reflected in the higher response rate. Still, this difference did not reach statistical significance. Furthermore, prior knowledge or skills alone are unlikely to produce a significant divergence in the learning curve. Accordingly, we considered the group sufficiently representative to allow analysis.

This study analyzes the results of three separate workshops in combination. We feel able to unite results, as workshops were developed in parallel, using similar format, resources (place, actors and administrative team) and emphasis points, and thus shared an identical evaluation tool.

The baseline questionnaire provides valuable information on physicians’ perceived skills, reported case management and perceived intervention barriers when confronting patients suffering DV. Fear of opening Pandora's Box (Sugg & Inui 1992) as well as lack of knowledge, skills and time constraints have proven again to be significant barriers to detection of DV victims. Thus, the workshop, which targeted knowledge and skills, had aimed correctly. A selection bias affecting results might have occurred for several reasons: participation was partially voluntary, and some participants were specifically selected by their HMOs as they were thought to be qualified to lead DV programs in their organizations in the future. This bias may have manifested in the intervention barriers, where negative attitude (‘none of my business’) was ranked the lowest obstacle to intervention. Such bias might also account for relatively high perceived baseline capabilities. However, even in this selected population, knowledge and skills are still unsatisfactory. In any case, it does not diminish the achievements of this workshop, as this bias cannot account for the improvement found in all parameters at follow-up.

While perception of knowledge has improved and lack of knowledge as a barrier has decreased, participants had failed to improve their knowledge of legislation. It is possible the questions pertaining to this topic, phrased in true/false form, were inadequate for analysis. This specific issue will be corrected and reassessed in future workshops.

The increase in detection, evaluation and referral rates is impressive. However, the question posed to the participants does not allow distinction between general referral to various agencies and mandatory report in designated cases. This distinction will be refined in future questionnaires, to allow accurate analysis.

Conclusions

A national pilot SP-based training program for physicians relating to DV was developed and implemented. The program proved highly effective in raising self-perceived preparedness and competences in this challenging social and clinical field during a follow-up period of six months. Due to the success of the program, it is currently being adapted and tested as a complementary educational tool for other health professionals (such as nurses and social workers) who face DV. Future plans include implementation of the program nationally in the presented format, as well as developing a combined workshop (intimate partner violence, child abuse and elder abuse) for family physicians, orthopedic surgeons and emergency-room teams, who encounter all three such populations.

Acknowledgments

The authors thank all physicians, social workers, SPs and administrative personnel, who took an active and invaluable part in this program. They are especially grateful to the experts who developed the workshops with the authors: Drs Ran Michalak, Shmuel Zur, Amir Vardi, Diana Fletcher and Yehudit Antonelli, Nurse Inbal Levin and Social Workers Malca Prager, Shula Ilnay, Wendy Chen and Sara Boxenboim. Thanks are also due to Ms Noa Froylich for her contribution and dedication.

This work was funded by grants from the Israel Ministry of Health (involved in design and conduct of study and approval of manuscript) and the United Jewish Agency (UJA)—New York (involved in design of study).

Conflicts of interest: None.

Additional information

Notes on contributors

Daphna Shefet

DAPHNA SHEFET is an internal medicine specialist and a psychiatry resident. ORIT RUBIN, PhD, is a psychometrician—an expert in assessment & evaluation in the field of medical education.

Hagit Dascal-Weichhendler

HAGIT DASCAL-WEICHHENDLER is a Family Medicine Specialist, coordinates domestic violence programs in the Technion and is chairperson of the Domestic Violence Committee in Clalit Health Services, Haifa and West Galilee District.

Orit Rubin

DAPHNA SHEFET is an internal medicine specialist and a psychiatry resident. ORIT RUBIN, PhD, is a psychometrician—an expert in assessment & evaluation in the field of medical education.

Nirit Pessach

NIRIT PESACH is a senior social worker, formerly chief coordinator of domestic abuse programs in the Israel Ministry of Health.

Dvora Itzik

DVORA ITZIK is a social worker and the current chief coordinator of domestic abuse programs in the Israel Ministry of Health.

Shuli Benita

SHULI BENITA, RN MA, is the director of simulated patient training in the Israel Center for Medical Simulation (MSR).

Amitai Ziv

AMITAI ZIV, MD MHA, is deputy director of the Sheba Medical Center in Israel, and founder and director of the Israel Center for Medical Simulation (MSR).