287Vol. 34, No. 4
There is ample documentation that domestic (or inti-
mate partner) violence (DV) is a frequent, yet under-
recognized cause of visits to health care profession-
als.1 The societal cost of failing to address DV issues is
enormous. Victims of abuse are at increased risk of in-
jury, death, and a number of somatic and mental health
conditions that impair their lives and increase their use
of health care resources.1-4 Moreover, children who wit-
ness violent acts are more prone to develop long-term
physical and mental health problems.5
Unfortunately, physicians and other health profes-
sionals do not screen for DV in their patients6 and may
avoid dealing with issues raised by DV, even when it is
recognized.7 This reluctance is often based on concerns
of retaliation, police involvement, time pressures, cul-
tural differences, lack of trust, and, importantly, lack
of knowledge and confidence in managing these diffi-
We believe, as do many others, that a successful strat-
egy for addressing the needs of DV patients requires a
comprehensive multidisciplinary systems approach, not
a piecemeal solution. Improvements in health profes-
sional education are one necessary component of such
a systems-based strategy.6,11 Ideally, the DV educational
component of a larger DV strategy should enable health
professionals to address the needs of DV patients with
Methods for Continuing Medical Education
Can Internet-based Education Improve Physician
Confidence in Dealing With Domestic Violence?
J ohn M. Harris J r, MD, MBA; Randa M. Kutob, MD; Zita J. Surprenant, MD, MPH;
Roland D. Maiuro, PhD; Thomas A. Delate, MS
From the Department of Medicine (Dr Harris) and the Department of Fam-
ily and Community Medicine (Dr Kutob), University of Arizona Health
Sciences Center; Medical Directions, Inc, Tucson (Dr Harris and Mr De-
late); the Department of Preventive Medicine, Kansas University (Dr
Surprenant); and the Department of Psychiatry and Behavioral Sciences,
University of Washington (Dr Maiuro).
Background and Objectives: Domestic violence (DV) is a common, under-recognized source of visits to
health care professionals. Even when recognized, physicians are reluctant to deal with DV, citing a lack
of education and lack of confidence in addressing issues presented by DV patients. Only a small num-
ber of DV education programs have been shown to lead to improvements in professional knowledge
and confidence, and these are intensive, multi-day courses. We sought to develop an on-line DV educa-
tion program that could achieve improvements in physician confidence and attitudes in managing DV
patients comparable to classroom-based courses. Methods: We created an interactive, case-based DV
education program targeted to physicians caring for DV patients. We tested the effectiveness of this
program in changing attitudes and beliefs in a randomized, controlled trial of Kansas physicians who
volunteered to participate in a study of on-line continuing medical education. We measured program
effectiveness with an externally developed and validated pretest/posttest instrument. Results: Sixty-five
physicians completed the pretest/posttest, 28 of whom were assigned to receive the on-line DV pro-
gram. We found a +17.8% mean change in confidence (self efficacy) for physicians who took the DV
program versus a -.6% change for physicians who did not take the program. We also found improve-
ments in other important areas associated with poor management of DV patients. These changes were
similar or greater in magnitude to those reported by others who have used the same survey tool to
evaluate an intensive, multi-hour classroom approach to DV education. User satisfaction with the on-
line program was high. Conclusions: An interactive, case-based, on-line DV education program that
teaches problem-solving skills improves physician confidence and beliefs in managing DV patients as
effectively as an intensive classroom-based approach. Such programs may be of benefit to those seek-
ing to improve their personal skills or their health care delivery system’s response to DV.
(Fam Med 2002;34(4):287-92.)
greater confidence and an improved knowledge of re-
sources and medical-legal issues associated with DV.
A useful program should also be easily distributed
throughout an organization, recognizing the conflict-
ing demands on professional (or student) time as well
as organizational resources.
Medical schools and academic centers are paying
increasing attention to the need for violence and abuse
training in their curricula.12 Family practice residen-
cies are also providing more DV education, approxi-
mately 4–5 hours per year.13 Unfortunately, few DV edu-
cation programs have undergone any type of evalua-
tion. It is unlikely that brief didactic programs, which
may do little harm, actually do any good.14 In contrast,
extensive, multi-day DV education programs have been
shown to increase knowledge and confidence in medi-
cal students15 and practicing health professionals.16
However, such programs are resource intensive and
difficult to reproduce.
We have previously shown that an interactive, com-
puter-based education program distributed via the
Internet can improve the confidence, knowledge, and
clinical skills of practicing physicians and physicians
in training in managing pigmented skin lesions.17,18 This
skin cancer program, based on well-established con-
cepts of adult learning theory, emphasized a user-driven,
problem-solving approach rather than a didactic pre-
sentation of skin cancer facts and tests of memory re-
call. We wished to determine whether a similarly de-
signed program could improve the confidence of prac-
ticing physicians in managing DV patients. If a brief,
inexpensive, on-line DV education program can pro-
vide benefits comparable to extensive multi-day
“clerkships” in DV, then the use of this type of educa-
tional approach may enable academic centers, residency
programs, and health care delivery organizations to
improve their educational strategies and overall sys-
tems for managing DV patients without substantial in-
vestments in new programs or in student time.
Educational Program Development
We developed our DV educational program around
a series of case studies in which different aspects of
DV were gradually revealed as the user worked through
the case. The interactive case-based scenarios empha-
sized clinical aspects of DV as it might appear in a fam-
ily practice office: DV during pregnancy, DV in older
persons, DV within a family presenting with somatic
complaints in a child, and dealing with the DV perpe-
trator. Users were asked to respond to typical clinical
questions, ranging from “What tests would you now
order during this routine pregnancy?” to “Would you
recommend family counseling for the victim and her
batterer?” Users were provided discussions of correct
and incorrect answers, with no penalty for selecting an
incorrect response. The program also provided links to
on-line DV resources and electronically transferable
materials (“enablers”) that physicians could use in their
practices. Additionally, the program provided geo-
graphically specific information on the locations of lo-
cal shelters and support groups and the state-mandated
DV reporting requirements for health providers.
The specific objectives of the on-line DV program
were to (1) increase physicians’ ability to recognize
victims of DV by increasing their use of office-based
screening tools, (2) improve physicians’ knowledge of
risk assessment techniques, (3) improve medical record
documentation of recognized cases of DV, (4) increase
physicians’ self efficacy in managing DV victims, and
(5) increase physicians’ self efficacy in managing DV
We created the on-line DV program using standard
Web-authoring tools, including Macromedia Cold Fu-
sion® and Microsoft SQL Server.® By using a data-
base-driven (versus a static hypertext mark-up language
[HTML]) approach, we were able to provide custom-
ized responses to user input.
The program was presented to the study subjects via
a private Web site, which was maintained by a com-
mercial continuing medical education (CME) publish-
ing company. Although available via the Internet, ac-
cess to the test program was restricted to study partici-
pants who were assigned to the intervention group by
use of predefined user ID/password combination. Once
subjects entered the identification/password, they were
instructed to enter additional identifying information
and create a unique individual identification/password.
Study subjects were practicing Kansas physicians
who had not received more than 1 hour of CME in-
struction in DV during the prior year and who responded
to an invitation to participate in a study of on-line DV
education sent by the Johnson/Wyandotte Counties
(Kansas) Medical Society. In total, we sent invitations
via fax to 1,887 physicians (distributed throughout the
state) who were members of the Kansas Medical Asso-
ciation and were practitioners of a primary care spe-
cialty or members of certain specialties likely to care
for DV patients (eg, emergency medicine and orthope-
dics). Those who agreed to participate and met the eli-
gibility criteria were provided a $50 honorarium for
completing two surveys of DV attitudes and beliefs. If
the participants were randomized to the intervention
(on-line education) group, they were offered an addi-
tional $50 honorarium and 2 hours of American Acad-
emy of Family Physicians Prescribed CME Credit for
completing the on-line education program. All eligible
participants who completed both surveys comprised the
289Vol. 34, No. 4
Our primary study hypothesis was that physicians
who were assigned to receive the on-line DV program
would have a greater change in their confidence in
managing DV patients (self efficacy), as measured by
a self-reported survey instrument, than would a control
group that did not receive the program. We also hy-
pothesized that the on-line program would reduce in-
hibiting beliefs and attitudes (eg, fear of offending the
patient, blaming the victim, and safety concerns) pre-
viously documented as obstacles to physician interven-
tion with DV patients.7
We measured subjects’ DV attitudes and beliefs with
a survey instrument developed at the University of
Washington19 that has been demonstrated to be a reli-
able and valid measure in studies of DV educational
programs.16 The initial survey instrument contained 39
items that measured six DV domains: provider self ef-
ficacy, perceived systems support, victim blaming, fear
of offense, safety concerns, and perceived frequency
of asking about DV. Cronbach alphas for the instru-
ment ranged from .73–.91, indicating excellent inter-
item reliability for the six domains.19 For the present
study, the instrument’s designer (Dr Maiuro) added two
new items to the victim/provider safety domain and one
to the frequency of DV inquiry domain, for a total of
42 items. We also added 14 questions to the survey in-
strument to assess DV knowledge and eight questions
assessing provider performance expectancy. All items
were presented as questions and scored based on a 5-
point Likert-type scale. Sample items from each do-
main of the survey instrument are shown in Table 1.
We developed a summary score to represent change
across each of the eight domains and also examined
results for key items individually for descriptive and
qualitative purposes. The survey instrument also in-
cluded basic demographic information, including age,
gender, year of graduation from medical school, pri-
mary practice field, previous hours of training in DV,
and current practices in managing DV patients.
At the conclusion of the on-line program, we also
asked course users to evaluate the quality of the pro-
gram, using a similar survey instrument as we have used
for evaluating other on-line educational programs. Spe-
cifically, users were asked to rate (on a 1–5 scale) how
well the learning objectives were met, how relevant the
program was to their practice, and the overall quality
of the program.
Participating physicians who met the eligibility cri-
teria were randomly assigned to either the intervention
or control group. Both groups were asked to complete
the written survey at the initiation and conclusion of
the study. When physicians in the intervention group
returned the baseline survey, they were given written
information on how to access the on-line program and
asked to complete the education program within 2
weeks. Approximately 3 weeks after completing the
on-line program, physicians in the intervention group
were mailed a second (posttest) survey (approximately
6 weeks after completing the first survey). Physicians
in the control group also received the second survey
approximately 6 weeks after completing the first survey.
The study endpoints were changes in mean Likert-
type scores on the eight domains measured by the sur-
vey instrument. We used independent sample t tests to
compare group means for the changes in domain scores
Sample Items From Domestic Violence Survey Instrument
Perceived self efficacy
Fear of offense
Frequency of asking about DV
I feel confident that I can make the appropriate referrals for abused patients.
I have ready access to mental health services should our patients need referrals.
The victim’s passive-dependent personality often leads to abuse.
It is not my place to interfere with how a couple chooses to resolve conflicts.
I feel there are ways of asking about battering behavior without placing myself at risk.
In the past 3 months, when seeing patients with headaches, how often have you asked about the possibility of domestic
How much do you feel you now know about how to document DV or suspected DV in the medical chart?
The number of times I document DV in the medical chart will . . .
* See Maiuro et al19 for additional information and examples.
Items were scored 1 to 5: “strongly disagree” to “strongly agree,” “never” to “always,” “nothing” to “very much,” or “decrease” to “increase a lot,” as
appropriate for item. Improvement in two items, “victim blaming” and “fear of offense,” were represented by lower scores. Improvement in all other items
were represented by higher scores.
Methods for Continuing Medical Education
between the intervention and control groups, as well as
continuous demographic variables. The Pearson chi-
square test was used to test for differences in propor-
tions between the intervention and control groups. To
control for possible confounding, we performed hier-
archical multivariate regression analyses, with domain
change score as the dependent variable, intervention
group as the independent variable, and, adjusted for
gender, hours of previous DV training, and years since
A total of 121 physicians agreed to participate in the
study. Twenty-two of these were excluded for not meet-
ing the study criterion of having ≤1 hour of DV educa-
tion in the prior year. Sixty-five (66%) of the 99 eli-
gible physicians completed both the pretest and posttest
survey. This group was the study population. The
progress through the study is shown in Figure 1.
Study subjects were predominantly
male (70%) and white, non-Hispanic
(81%). Demographic and other baseline
characteristics were evenly balanced
between the intervention (28 physi-
cians) and control (37 physicians)
groups (Table 2).
The primary study hypothesis, that
use of the on-line DV program would
lead to increased self efficacy, was sup-
ported by the study findings (Table 3).
There was a positive 17.8% mean
change in the self-efficacy domain score
for the intervention group versus a nega-
tive .6% change for the control group
(P<.001). In addition, we found signifi-
cant changes in five other domain scores
that reflected decreases in victim blam-
ing and fear of offense (note: negative
or decreased mean scores in these do-
mains are desirable) and increases in the
provider’s composite sense of personal
safety as well as the safety of the vic-
tim, provider knowledge, and provider
performance expectancy. Although the
scores changed in a positive direction
for perceived system support and per-
ceived frequency of asking about DV,
the differences between the intervention
and control groups did not reach the
level of P<.05 for these domains. Mul-
tivariate analyses to control for possible
confounding of these results by differ-
ences in age, gender, or levels of previ-
ous DV training in the two groups did
not alter the interpretation of the results
(data not shown).
A key item of the on-line DV program was an expla-
nation of mandatory abuse reporting laws for health
care providers. One question in the DV knowledge do-
main was “How much do you feel you know about your
legal reporting requirements for DV?” Mean changes
in scores for this question showed a 39.4% increase for
the intervention group versus a 5.1% decrease for the
control group (P<.005).
Self-reported user satisfaction with the program was
high. On a 1–5 scale, where 5 represented “very well,”
“very relevant,” or “excellent,” the mean response to
our satisfaction questions “How well were the learning
objectives of this program met?” was 4.5, “How rel-
evant was the information in this program to your clini-
cal practice?” was 4.1, and “How would you rate this
program overall?” was 4.5.
Flow Diagram of Progress Through the Study
291Vol. 34, No. 4
We demonstrated that an on-line DV
education program can improve physician
confidence (as measured by self efficacy),
attitudes, and self-reported knowledge in
managing DV patients. This finding should
be of considerable interest to academic
centers and residency programs wishing
to incorporate less resource-intensive, ef-
fective DV education programs into their
curricula or to organizations seeking to
develop systems-based solutions to DV
issues. Our results suggest that a relatively
brief, easily distributed, interactive, on-line
DV education program can be as effective
as much more-intensive curriculum-based
programs. The provider survey results in
this study were similar or greater in mag-
nitude to those reported by Thompson et
al,16 who relied on multiple expert instruc-
tors and an intensive, multi-hour classroom
approach with role-play exercises.
Based on this study and our earlier work
with on-line skin cancer CME, we believe
that the elements associated with the
program’s success were its interactive for-
mat, its focus on realistic problem-solv-
ing exercises, the presence of user-acces-
sible enablers, and its flexibility. This is
consistent with long-standing recommen-
dations for changes in medical education.20
The use of low-cost computer-based in-
struction permits the development of indi-
vidualized, problem-solving educational
programs. The presence of the Internet per-
mits their inexpensive distribution. Read-
ers who wish to view this program via the
Internet can follow instructions for doing
so in the Acknowledgments section.
Our study had several strengths. We used a broad
base of Kansas physicians, not physicians in training
or members of a single organization, to enhance the
generalizability of the results. We used a randomized,
controlled trial design with a psychometrically vali-
dated, externally derived survey instrument. These ele-
ments enhance the reliability and validity of the results.
We allowed approximately 3 weeks (mean 25.7 days)
between completion of the on-line program and admin-
istration of the second survey to the intervention phy-
sicians, thus partially addressing the issue of effect de-
lay. Few DV education programs have been formally
evaluated in this rigorous fashion.
An important study limitation was that we did not
directly evaluate the program’s effect on actual physi-
cian behaviors or health outcomes. We did note impor-
tant changes in physician self efficacy and five other
key attitudes and beliefs. Self efficacy, which implies
more than “self-confidence,” may be the most critical
and influential link between beliefs and behavior.21 It
can predict a variety of health-related behaviors.22 We
are hopeful that these types of changes will improve
individual physician’s and the health care delivery
system’s approach to the problems raised by DV.
We do not know if others can duplicate our success
with this type of on-line education or whether the posi-
tive effects we observed will be durable. These are top-
ics of ongoing research. Lastly, if a program such as
this can produce desirable benefits, will every academic
center and health delivery system still need to develop
its own on-line education programs or will successful
models emerge that can be shared by all?
Characteristics of Intervention and Control Groups
Completed Survey #1
Completed Survey #2*
Sample size44 37
Male (%)69.270.4 64.373.0
White, non-Hispanic (%)84.7 80.0 92.983.8
Age** in years
Years** since graduation
Hours** of prior non-CME
DV cases seen/year (%):
Primary practice (%)
CME—continuing medical education
* Data from subjects who completed both surveys were used to test the study hypotheses.
All comparisons between intervention and control groups were P>.05.
Mean values shown
*** Includes emergency medicine, OB-GYN, and a small number of other specialists likely
to care for DV patients.
Methods for Continuing Medical Education
292 Download full-text
Acknowledgments and Access to the On-line DV Program for Readers: This
CME program and its evaluation were developed under a Small Business
Innovation and Research grant from the National Institute of Mental Health
to Medical Directions, Inc, a private company. Dr Harris and Mr Delate are
employees of Medical Directions, Inc. Medical Directions has developed
an Internet site that will allow readers of Family Medicine to view the on-
line DV program described in this paper, without charge or special access
codes at http://dv.lecturehall.com. Instructions for access are provided on
Financial Support: This work was supported by grant 1R43-MH62233 from
the US National Institute of Mental Health (NIMH). The opinions and as-
sertions contained herein represent those of the authors and not the NIMH.
Corresponding Author: Address correspondence to Dr Harris, Medical Di-
rections, Inc, 6101 E. Grant Road, Tucson, AZ 85712. 520-722-1970. Fax:
520-722-2670. E-mail: email@example.com.
1.Eisenstat SA, Bancroft L. Domestic violence. N Engl J Med 1999;
Coker AL, Smith PH, Bethea L, King MR, McKeown RE. Physical
health consequences of physical and psychological intimate partner
violence. Arch Fam Med 2000;9:451-7.
Wisner CL, Gilmer TP, Saltzman LE, Zink TM. Intimate partner vio-
lence against women: do victims cost health plans more? J Fam Pract
Kernic MA, Wolf ME, Holt VL. Rates and relative risk of hospital ad-
mission among women in violent intimate partner relationships. Am J
Public Health 2000;90:1416-20.
Straus M. Children as witnesses to marital violence: a risk factor for
lifelong problems among a nationally representative sample of Ameri-
can men and women. In: Schwartz DF, ed. Children and violence: re-
port of the 23rd Ross Roundtable on critical approaches to common
pediatric problems. Columbus, Ohio: Ross Laboratories, 1992:98-104.
Waalen J, Goodwin MM, Spitz AM, Peterson R, Saltzman LE. Screen-
ing for intimate partner violence by health care providers: barriers and
interventions. Am J Prev Med 2000;19:230-7.
Sugg NK, Inui T. Primary care physicians’ response to domestic vio-
lence: opening Pandora’s box. JAMA 1992;267:3157-60.
8. Sugg NK, Thompson RS, Thompson DC, Maiuro R, Rivara FP.
Domestic violence and primary care: attitudes, practices, and beliefs.
Arch Fam Med 1999;8:301-6.
Rodriguez MA, Bauer HM, McLoughlin E, Grumbach K. Screening
and intervention for intimate partner abuse: practices and attitudes of
primary care physicians. JAMA 1999;282:468-74.
10. Sherin KM, Sinacore JM, Li X, Zitter RE, Shakil A. HITS: a short
domestic violence screening tool for use in a family practice setting.
Fam Med 1998;30:508-12.
11. Cole TB. Case management for domestic violence. JAMA 1999;282:
12. Alperts EJ, Tonkin AE, Seeberman AM, Holtz HA. Family violence
curricula in US medical schools. Am J Prev Med 1998;14:273-82.
13. Rovi S, Mouton CP. Domestic violence education in family practice
residencies. Fam Med 1999;31(6):398-403.
14. Coonrod DV, Bay C, Rowley BD, et al. A randomized controlled study
of brief interventions to teach residents about domestic violence. Acad
15. Jonassen JA, Pugnaire MP, Mazor K, et al. The effect of a domestic
violence inter clerkship on the knowledge, attitudes, and skills of third-
year medical students. Acad Med 1999;74:821-8.
16. Thompson RS, Rivara FP, Thompson DC, et al. Identification and man-
agement of domestic violence—a randomized trial. Am J Prev Med
17. Harris JM, Salasche SJ, Harris RB. Can Internet-based continuing medi-
cal education improve physicians’ skin cancer knowledge and
skills? J Gen Intern Med 2001;16:50-6.
18. Harris JM, Salasche SJ, Harris RB. Using the Internet to teach mela-
noma management guidelines to primary care physicians. J Eval Clin
19. Maiuro RD, Vitaliano PP, Sugg NK, Thompson DC, Rivara FP, Thomp-
son RS. Development of a health care provider survey for domestic
violence—psychometric properties. Am J Prev Med 2000;19:245-52.
20. Vollan DD. Postgraduate medical education in the United States. Re-
port of a survey by the Council on Medical Education and Hospitals,
1952–1955. Chicago: American Medical Association, 1955.
21. Bandura A. Self efficacy: the exercise of control. New York: W.H. Free-
22 Maddux JE, ed. Self efficacy, adaptation, and adjustment: theory, re-
search, and application. New York: Plenum, 1995.
Key Study Findings
BASELINE MEAN SCORE* POSTTEST MEAN SCORE* MEAN SCORE CHANGE
Fear of offending
Asking about DV
* Scores ranged from 1–5, with 5 indicating maximum amount; see text for discussion.
**P value for independent t test comparison between intervention and control groups