HEALTH CARE PROVIDERS' TRAINING,

PERCEPTIONS, AND PRACTICES REGARDING

STRESS AND HEALTH OUTCOMES
Holly Avey, MPH; Kenneth B. Matheny, PhD; Anna Robbins, MPH; and Terry A. Jacobson, MD
Atlanta, Georgia and Portland, Oregon

In order to assess health care providers' training, perceptions, and practices regarding stress
and health outcomes, a survey was administered to primary care providers in the outpatient med-
ical clinics of a southeastern urban hospital serving a predominantly African-American indigent
population. One-hundred-fifty-one of 210 providers (72%) responded. Forty-two percent of respon-
dents reported receiving no instruction regarding stress and health outcomes during their med-
ical/professional education. While 90% believed stress management was "very" or "somewhat"
effective in improving health outcomes, 45% "rarely" or "never" discussed stress management with
their patients. Respondents were twice as likely to believe that counseling patients about smoking,
nutrition, or exercise was more important than counseling them about stress. Seventy-six percent
lacked confidence in their ability to counsel patients about stress. The majority of respondents
(57%) "rarely" or "never" practiced stress reduction techniques themselves. Belief in the importance
of stress counseling, its effectiveness in improving health, and confidence in one's ability to teach
relaxation techniques were all related to the probability that providers would counsel patients
regarding stress. There is a need for curriculum reform that emphasizes new knowledge about
stress and disease, new skills in stress reduction, and more positive beliefs about mind/body med-
icine and its integration into the existing health care structure. (J Nati Med Assoc. 2003;95:833-845.)

Key words: stress * health care providers persons at risk for disease"'. Stress has been asso-
+ health outcomes + training * practice ciated with greater severity and duration of infec-
tious diseases,2 as well as immune-mediated dis-
INTRODUCTION eases such as asthma3, rheumatoid arthritis4,
Stress has been defined as, "the process in inflammatory bowel disease5, and progression from
which environmental demands tax or exceed the HIV to AIDS6. There is now significant research
adaptive capacity of an organism, resulting in psy- showing that psychological stress can down-regu-
chological and biological changes that may place late various aspects of the cellular immune
response and disrupt the bidirectional communica-
tion links between the central nervous system and
© 2003. From the Grady Health System, Atlanta, GA (Avey); the immune system7-8.
Georgia State University (Matheny); Department of Human Stress has also been linked to hypertension9. An
Services, Health Services, Portland, OR (Robbins); and Emory
University, Atlanta, GA (Jacobson). Send correspondence and extensive body of evidence from animal models,
reprint requests for J Natl Med Assoc. 2003;95:833-845 to: Holly which has been summarized in a review by
Avey, MPH, Grady Health System, Office of Health Promotion, 80 Rozanski et al. (1999), reveals that chronic psy-
Jesse Hill Jr. Drive SE, P.O. Box 26101, Atlanta, GA 30303; phone: chosocial stress can lead to exacerbation and accel-
(404) 616-7561; fax: (404) 880-9464; e-mail: [email protected] eration of coronary artery atherosclerosis, as well

833 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003
 TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

as to hypercortisolemia'°. The authors also report Table 1. Health Care Provider Demographics
that acute stress has been shown to trigger myocar- (n=151)
dial ischemia, promote arrhythmogenesis, stimu-
late platelet function, increase blood viscosity, and Age
cause coronary vasoconstriction in the presence of Average age 30
underlying atherosclerosis. The review concludes Age range 24-58
by suggesting that present studies "provide clear Gender
and convincing evidence that psychosocial factors Male 64%
contribute significantly to the pathogenesis and Female 36%
expression of coronary artery disease"10. Recent Race or ethnic group
Caucasian 65%
research has also clarified the relationship between Asian/Pacific Islander 15%
stress, inflammation, and cardiovascular disease". African-American 10%
Further evidence suggests that the cardiovascular, Other 8%
immune, and endocrine systems react simultane- Hispanic 3%
ously to a psychological stressor, creating a coordi- Year of medical/professional school graduation
nated and consistent response'2. In addition, stress 1995-1999 80%
has also been implicated in the development and 1990-1994 11%
control of diabetes'3-'4, the perception and tolerance 1967-1989 10%
of acute and chronic pain conditions,'5-'7 and the Current or future specialty
Primary care 44%
expression of somatization complaints'8. Subspecialty 40%
Given its relationship with such a wide variety Undecided 16%
of disease states, it is not surprising to learn that Country of citizenship (17 countries represented)
stress has been considered an underlying factor for U.S. 84%
a wide variety, and perhaps even the majority, of Other 16%
health care provider visits. In 1981, a landmark Current position
study reviewing the charts of Kaiser-Permanente Resident physician 78%
patients concluded that 60-90% of physician visits Attending physician 18%
reflect emotional distress and somatization'9. A Nurse practitioner 3%
Physician's assistant 1%
similar study published in 1989 reviewed diag-
noses and treatments for the 14 most common
complaints of 1,000 patients followed in an internal research has established the relationship between
medicine clinic over a three-year period. Although stress management or relaxation and the improve-
diagnostic testing was performed in two-thirds of ment of these disease states. Cardiologist Herbert
the cases, an organic etiology was demonstrated in Benson (1975) pioneered the term "relaxation
only 16% of the complaints. The authors report that response" to describe the physiological effects of
many of the symptoms of unknown etiology were relaxation which reverse the trends of stress and
probably related to "psychosocial factors"20. disease22-23. Building on his ground-breaking work,
McEwen's (1998) contribution to the stress lit- researchers have established a wide variety of ben-
erature through the development of allostatic load efits of stress management, relaxation, and coping
theory helps to explain the development of these interventions in treating asthma24, arthritis24-26,
disease processes through a breakdown in the allo- hypertension27-33, diabetes34-37, pain'6' 38-39, and somati-
static process, which brings an acute stress reaction zation40, as well as general health and well-being4".
back to homeostatic balance2'. This theory suggests Despite proven scientific research on the physi-
that when individuals are exposed to acute and/or ological risks of stress and the benefits of stress
chronic stressors over an extended period of time, management, little is known about health care
the stress response becomes sustained (as in hyper- providers' training, perceptions, and practices
tension and diabetes), with some body systems regarding stress and health outcomes. Previous
experiencing tissue fatigue and suppression (such research on physician counseling behavior, utilizing
as immune and pain systems). the theoretical framework of Bandura's Social
While the body of evidence linking stress to dis- Cognitive Theory (SCT)42, indicates that physicians
ease continues to mount, a concurrent body of who expect that they will be successful in changing

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003 836
TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Table 2. Prior Training in Stress and METHODS

Health Outcomes
Participants
Received instruction (during medical/ Participants were internal medicine residents,
professional school or residency)* (n=140) attending physicians, nurse practitioners, and physi-
Stress management 62 (44%) cian's assistants working in the outpatient medical
Mind/body medicine 37 (26%) clinics of an academic training program serving a
Other training on stress & health 27 (19%)
predominantly African-American indigent popula-
Psychoneuroimmunology 14 (10%)
None of the above 59 (42%) tion in an urban hospital in the southeast.
How much instruction: (n=76**)
One lecture 20 (26%) Study Design
2-3 lectures 44 (57%) Between August and November 1999, the sur-
4-5 lectures 11 (14%) vey was administered to 210 providers prior to
Entire course 1 (1%) their weekly ambulatory care clinics. Providers
Content of instruction (check all that apply)* who did not initially respond to the survey were
(n=83) solicited for participation two more times, for an
Heart disease and stress 53 (64%)
Health outcomes and stress 49 (59%) overall response rate of 72% (n=15 1).
Blood pressure and stress 45 (54%) The survey measured the stress counseling prac-
Pain tolerance and stress 32 (39%) tices ofproviders and their level oftraining in stress
Immune functioning and stress 30 (36%) management, mind/body medicine, and/or psy-
Other 5 (6%) choneuroimmunology (these terms were not mutu-
Instruction on relaxation techniques (n=150) ally exclusive and respondents could check more
Yes 51 (34%) than one). Building on Bandura's Social Cognitive
No 99 (66%) Theory (SCT)42, we also investigated whether the
Type of relaxation techniques (n=50) tendency for providers to offer stress counseling to
Progressive muscle relaxation 38 (76%)
Meditation 30 (60%) patients was associated with a) perceived effective-
Diaphragmatic breathing 28 (56%) ness of stress management and stress counseling
Imagery/visualization 26 (52%) (outcome expectations), b) perceived importance of
Mindfulness' 7 (14%) stress counseling (outcome evaluation), or c) per-
Other 1 (2%) ceived success in changing behavior (self-efficacy).
In addition, health care providers' personal practice
Not mutually exclusive terms, may equal of stress management techniques was evaluated to
more than 100%. determine if their personal practice might affect
5 missing
Focusing awareness on the present moment their perceptions of the importance and effective-
ness of such techniques and the frequency that they
would provide stress counseling to patients (See
a patient's behavior (outcome expectation) are more Appendix A for specific survey questions).
likely to ask their patients about the behavior and
counsel them about it3. Physician self-perceived Statistical Analysis
effectiveness (self-efficacy) in changing patient The data were entered into EPI Info 6.04b and
behavior is also associated with greater efforts to analyzed using SAS and SPSS 11.0. Simple frequen-
counsel patients on adult preventive care"4. Finally, cies and means were calculated for all categorical
continuing medical education sessions that are and ordinal variables. Spearman correlations were
interactive and allow participants the opportunity to calculated for all of the ranked variables46. Multiple
practice their skills (role modeling, self practice) linear regression was used to determine relationships
are the most effective in developing appropriate between the dependent variable the frequency with
counseling skills45. The study below reflects our which providers counsel patients about stress-and
attempts to investigate health care providers' train- the independent variables mentioned above, which
ing, perceptions, and practices regarding stress and might influence the probability that such counseling
health outcomes in a primary care setting, utilizing would be provided47. For each analysis, a p value of
a Social Cognitive Theory approach. <0.05 was considered to be significant.

837 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003
 TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Table 3. Importance of Counseling Patients on Health Behaviors

In general, how important do you think it is for providers to counsel patients about ... ? (n=1 50)
Very Somewhat Not Very Not at All Don't Know
Important Important Important Important Not Sure
Smoking 96%a 4% 0% 0% 0%
Nutrition 93%b 7% 1% 0% 0%
Exercise 90%C 9% 1% 0% 0%
Stress 47% 51% 1% 0% 0%
a statistically significant difference from same category for stress
(X2=62.23, p<0.001)
b statistically significant difference from same category for stress (X2=66.06, p<0.001)
c
statistically significant difference from same category for stress (X2=66.06, p<0.001)

The survey items relating to variables in the Outcome Expectations

regression models were arranged in a four-point Perceived Effectiveness of Stress Management.
Likert scale, with two exceptions. Respondents When asked, "How effective is stress management
who reported "none of the above" for training in in improving health outcomes?," 24% responded
stress management, mind/body medicine, etc. were "very effective" and 66% said "somewhat effec-
assigned a "5" on a Likert scale, with the remain- tive." No respondents felt that stress was "not very"
ing four variables representing the amount of or "not at all effective," but 10% indicated that they
instruction (4=one lecture, 1=entire course). The were "not sure" (n=1 50).
survey item measuring the degree to which respon- Perceived Effectiveness of Stress Counseling.
dents felt stress management was effective in Twenty-nine percent ofproviders reported that they
improving health outcomes was collapsed into a were not confident that their patients would actual-
dichotomous variable because there were no ly follow their stress management recommenda-
respondents who felt that stress management was tions, if offered them (n=143).
"not very" or "not at all" effective.
Outcome Evaluation: Perceived
RESULTS Importance of Counseling
Although 90-96% of respondents thought it was
Subjects and Response Rate very important to counsel patients about diet, exer-
One-hundred-fifty-one participants completed cise, and smoking, only 47% thought it was very
the survey, resulting in a 72% response rate. Specific important to counsel them about stress (n=150).
respondent demographics are reported in Table 1. Chi-square analyses showed that these differences
were statistically significant. Table 3 reports the
Training of Providers opinions of respondents regarding the importance
A large proportion of the respondents (42%) of counseling patients on different health behaviors
received no instruction in stress management, and the results of the chi-square analyses (See
mind/body medicine, or psychoneuroimmunology Table 3).
(n=140) during their professional education, and
66% received no training in specific relaxation Self-Efficacy: Perceived Success in
techniques (n=150). Of those who did receive Changing Behavior
instruction (n=78), 83% had received only 1-3 lec- When asked. "How confident are you in your
tures. Less than 40% of those who received train- ability to teach relaxation techniques (such as med-
ing (approximately 20% of all respondents) itation, deep breathing, etc.) to your patients?,"
received specific instruction on the relationship 76% responded that they were "not very" or "not at
between stress and immune functioning or stress all confident." Fifteen percent reported that they
and pain. Table 2 summarizes the health care were "somewhat" confident, while only 3% report-
providers' training regarding stress and relaxation. ed that they were "very" confident (n=15 1).

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003 838
TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Table 4. Correlations of Social Cognitive Theory Variables with Stress Counseling

Training Important to Personal Confidence Stress mgmt. Discuss stress
counsel about practice of teaching improves management
stress relaxation relaxation health with patients
Training 1.00 0.290" 0.234" 0.297** 0.1 71 * 0.154
Importance 1.00 0.438** 0.262** 0.430" 0.405"
Personal practice 1.00 0.397** 0.369** 0.393**
Confidence teaching 1.00 0.310** 0.345**
Stress mgmt. 1.00 0.394**
improves health
Discuss stress mgmt. 1.00
"Significant at the 0.01 level 'Significant at the 0.05 level

Stress Practices groups. Frequencies for training, personal practice,

Stress Counseling. When asked, "How often do outcome evaluation, outcome expectation, and
you discuss stress management with your self-efficacy were calculated for each subgroup.
patients?," 45% of respondents replied "rarely" or Any difference in frequencies was then examined
"never." Fifty percent reported that they discuss using Pearson chi-square analysis46.
stress management "sometimes" and 6% discuss it Physicians vs. Nurse Practitioners and
"routinely" (n=151). Physician's Assistants. Although there were not
Personal Practice. The majority of respondents enough physician's assistants and nurse practition-
(57%) reported that they "rarely" or "never" practice ers (n=7) to compare to physicians (n=144), analy-
stress reduction techniques themselves (n=149). Of ses were repeated with nurse practitioners and
those who do (43%), exercise was the most common physician's assistants removed. This did not result
technique used (70%), with a lesser amount report- in statistically significant changes for any of the
ing meditation (32%), imagery (26%), diaphragmat- findings. Given the fact that the results of this
ic breathing (21 %), mindfulness (16%), and pro- study reflect the training, perceptions, and prac-
gressive muscle relaxation (16%). tices of a group of health care providers working
together in one health care setting, we believe it is
Barriers to Counseling Patients important to represent this reality fully by includ-
About Stress ing all providers in our representation of the data.
Providers reported many barriers in counseling Primary Care vs. Subspecialty. Analyses were
patients about stress (n=149). Nearly three-quarters performed to determine if there were any differ-
(73%) of respondents felt that they were too busy or ences between health care providers who declared
didn't have enough time to counsel patients about an interest in primary care (n=65) versus those who
stress management. Sixty-three percent reported declared intentions to specialize (n=85). There
their lack of training in stress management was a were no statistically significant differences detect-
barrier, 54% felt a lack of confidence in teaching ed between these two groups.
relaxation techniques, and 49% said a lack of refer- Caucasian Providers vs. Providers of Color
ral sources was a barrier. Very few (7%) reported Analyses were performed to determine if there
lack of interest in stress management as a barrier were any differences between caucasian health care
and none reported doubt that stress affects health. providers (n=97) versus providers of color (n=53).
To prepare for this analysis, providers who selected
Post-hoc Analyses for Potential African-American, Hispanic, Asian/Pacific Islander,
Group Differences Native American, or Other for race were collapsed
Several post-hoc analyses were performed to into a "Providers of Color" category, creating a
determine if there were any differences between dichotomous variable with caucasians. The only

839 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003
 TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Table 5. Summary of Multiple Linear Regression Analysis for Variables

Predicting Frequency of Discussing Stress Management with Patients
Variable F value t-test p value R2adj
* 11.058 * <0.001 0.238
Importance of counseling about stress 0.248 * 2.106 0.037
Personal practice of relaxation techniques 0.118 * 1.763 0.080
Confidence in teaching stress mgmt 0.144 * 1.935 0.055
Stress management is effective in improving health outcomes 0.209 * 2.400 0.018
Training in stress and health outcomes'
n=151, df=4
n Using backward variable elimination, training was eliminated as a variable for this model.

finding which was statistically significant between Cognitive Theory, we assessed the degree to which
the two groups was the degree to which each group a select group of independent variables measuring
ranked the importance of counseling patients about training, personal practice, outcome evaluation
stress (X2=8.1 1, p=0.004), with providers of color (perception that stress counseling is important),
more likely than caucasian providers to view stress outcome expectation (perception that stress man-
as a very important patient counseling topic. agement is effective in improving health out-
comes), and self-efficacy (confidence in teaching
Correlations Among Variables relaxation techniques to patients), were associated
Single Variables Predicting Stress Counseling. with the frequency with which providers would
The frequency with which providers discussed stress offer stress counseling to their patients. The result-
management with their patients was not significant- ing model eliminated training as a variable. The
ly correlated with the amount of training a provider relationship between the remaining variables was
received regarding the effects of stress on health found to be significant (p<0.001). The resulting
(Spearman=0. 154; p=0.08). However, training, was R2adj indicates that 24% of the variability in the
significantly correlated with the outcome evaluation frequency with which providers counsel their
that it is very important to counsel patients about patients about stress was explained by the combi-
stress (Spearman=0.290; p=0.001) and confidence nation of personal practice, outcome evaluation,
in teaching relaxation (Spearman=0.297; p < .00 1). outcome expectation, and self-efficacy.
The frequency with which providers discussed
stress management with their patients was signifi- DISCUSSION
cantly correlated with the provider's personal use With a sample of health care providers in an aca-
of relaxation techniques (Spearman=0.393; p= demic training program at an urban hospital for the
<0.001). The provider's personal practice of relax- indigent, we have begun to document the training,
ation techniques also was significantly correlated perceptions, and practices of health care providers
with self-efficacy or confidence in teaching relax- regarding stress and health outcomes. Our results
ation techniques (Spearman=0.397; p=<0.00 1), indicated that many providers received no instruction
and with the outcome expectation that stress man- in topics related to stress and health, and the majori-
agement is likely to be effective in improving ty did not receive instruction in relaxation techniques
health outcomes (Spearman=0.369; p=<0.001) during their medical/professional training. Among
(See Table 4 for correlation matrix). those who did receive instruction, only limited train-
Combined Variables Predicting Stress Counseling. ing was provided (1-3 days) for most, and less atten-
We used multiple linear regression with backward tion was directed to the relationship of stress to pain
variable elimination to investigate the degree to tolerance and to immune functioning than to heart
which certain independent variables were associat- disease, blood pressure, or other health outcomes.
ed with the dependent variable of frequency with Although 90% of providers judged stress man-
which health care providers discuss stress manage- agement to be "very" or "somewhat" effective in
ment with their patients. Informed by the Social improving a patient's health, this belief often did not

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003 840
TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Appendix A-Relevant Sections of the Health Care Provider Survey

Demographics
1. Country of citizenship
2. Race or ethnic group (if biracial, check all that apply)
O Caucasian O African-American O Hispanic O Asian/Pacific Islander O Native American
Other
3. Sex O Female O Male
4. Age
5. Which of the following best describes your current position?
O PGY-1 O PGY-2 O PGY-3 O PGY-4 O PGY-5 O Attending Physician O Nurse Practitioner
O Physician's Assistant O Other
6. What is your current or future career specialty?
O Primary care O Subspecialty LI Undecided
7. Year of graduation from medical/professional school:
Professional Training
8. Please check the areas in which you received any instruction during medical school, residency, or
professional school:
O Stress Management
O Mind/Body Medicine-that thoughts and behaviors influence health and well-being
O Psychoneuroimmunology (PNI)-the links between the mind, the brain, and the immune system
O Other training on the relationship between stress and health
O None of the above-Skip to number 10.
If you received training in stress management, mind/body medicine, or psychoneuroimmunology,
please check how much training you received:
O 1 lecture O 2-3 lectures O 4-5 lectures O Entire course O Other
9. Please check which specific topics you received instruction in:
O Immune functioning and stress O Blood pressure and stress
O Pain tolerance and stress O Other:
O Health outcomes and stress O None of the above
LI Heart disease and stress
10. Did you receive any instruction/training on relaxation techniques in medical school, residency, or
professional school? O Yes O No
If yes, please check which techniques: (check all that apply)
O Diaphragmatic breathing O Progressive muscle relaxation
O Meditation O Imagery/visualization
O Mindfulness-focusing your awareness O Other:
on the present moment
Clinical Practice
1 1. In general, how important do you think it is for providers to counsel patients about ...?
Very Somewhat Not Very Not at All Don't Know Not Sure
Important Important Important Important
Diet / Nutrition
Exercise
Smoking
Stress
12. How effective do you believe stress management is in improving health outcomes?
O Very effective O Somewhat effective O Not effective O Don't Know/Not Sure
13. How often do you discuss stress management with your patients?
O Routinely O Sometimes O Rarely O Never
14. How confident are you in your ability to teach relaxation techniques (such as meditation, deep
breathing, etc.) to your patients? O Very Confident O Somewhat Confident O Not Very Confident
O Not At All Confident O Not Sure
15. Which of the following are barriers to you in counseling patients about stress management? (Check all
that apply.) O I'm too busy/not enough time U I'm not very interested in stress management O don't
believe that stress affects health O received little training in stress management O am not confident in
my ability to teach relaxation techniques O am not confident my patients will follow my stress
management recommendations O Lack of referral sources for stress management
Personal Health Practices
16. How often do you practice relaxation or stress reduction techniques?
O Routinely O Sometimes O Rarely O Never-Skip to comments
17. Which techniques do you use? O Diaphragmatic breathing O Meditation O Mindfulness
O Progressive muscle relaxation O Imagery/visualization O Exercise O Other:

841 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003
 TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

translate to discussing stress management with their patients. However, training that emphasizes the
patients. One reason may be that providers were importance of stress management to health, encour-
twice as likely to believe that counseling patients ages the personal practice of stress reduction tech-
about smoking, nutrition, or exercise was more niques, and enhances confidence in teaching stress
important than counseling them about stress. When management may prove more effective. Thus inter-
providers are faced with limited time to interact with vening in several areas of provider training, percep-
their patients, a topic that is deemed only "some- tions, and skills may be required to improve the fre-
what" effective in improving their health is quite quency of provider stress counseling.
likely to never be mentioned. This belief, when cou-
pled with the lack of confidence many providers Study Limitations
reported in their ability to teach relaxation tech- There are several limitations to our study. The
niques to their patients or convince their patients to instrument was intended for use as a needs assess-
follow these recommendations, may reinforce the ment tool, so it was not tested for reliability or
low priority of stress counseling for many providers. validity. Like all surveys, recall bias may affect
The fact that providers of color were more like- respondents' ability to remember prior training on
ly to view stress as a very important patient coun- specific topics, especially if the question addresses
seling topic than caucasian providers is interesting, something that is not easily quantifiable like stress
although we cannot explain this finding fully in the training. We did not document training or self-
context of our study. However, we may hypothesize instruction in stress reduction techniques obtained
that providers of color may be more likely to expe- outside medical/professional school. Also, the
rience personal stress as a result of racism in our design of our study involved mainly resident physi-
society48, and therefore may be more sensitive to its cians in one academic training program, so our
biopsychosocial implications. It is also possible sample may not be representative of other residen-
that African-American providers (who make up cy programs or health care providers in general.
28% of our providers of color) may be more sensi- However, it is our hope that by providing a suffi-
tive to issues related to stress and hypertension, cient description of the demographic characteris-
given the higher rates of hypertension in the tics of our study population, readers may make an
African-American community49. informed decision regarding the generalizability of
Although training in stress-related topics was our findings to other similar groups.
not significantly correlated with the frequency of Even though we did not obtain a random sample
providing stress counseling to patients, the personal of all house officers, our response rate was high at
practice of stress management by providers (previ- 72%, and our results do reflect recent trends in
ous experience) was significantly associated with medical education. These findings suggest that the
the perception that stress counseling is important primary components ofthe Social Cognitive Theory
(outcome evaluation), confidence in teaching relax- -specifically skill-building personal practice, pos-
ation techniques (self-efficacy), and the perception itive outcome evaluation and outcome expectation,
that stress management is effective in improving and high self-efficacy, are all constructs which sup-
health outcomes (outcome expectation). These per- port physician counseling practices. We believe
ceptions were then significantly correlated with the these findings are also relevant to the training of
frequency with which providers discussed stress future nurse practitioners and physician's assistants.
management with patients. Thus, while training It is possible that the results of our study might
may increase the provider's perception ofthe impor- have been different if the survey had been adminis-
tance of stress counseling for patients, it may tered in a setting with providers who were not in an
require the personal practice of stress reduction academic training setting, or providers who were see-
techniques to increase the frequency with which ing patients with a higher SES. Providers with more
providers will actually provide such counseling. experience may be more likely to discuss psychoso-
Although the cross-sectional design of our study cial or lifestyle issues with their patients, but younger
limits our ability to predict provider stress counsel- providers may be more likely to have the most up-to-
ing, it does appear that current training strategies date training on stress topics. Providers who serve
have been insufficient in increasing the frequency patients with a higher socioeconomic status may feel
with which providers will offer stress counseling to more comfortable recommending techniques such as

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003 842
TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

meditation which their patients can obtain privately sel their patients about stress. The creation of such
through out-of-pocket expenses. On the other hand, guidelines should be a high priority for these organi-
providers who serve patients with health insurance zations, as well as for credentialing organizations-
may be more hesitant to recommend a treatment that such as the American College of Physicians, the
is not currently reimbursable through most health American Academy of Family Practice, and others.
insurance companies. Only further research can
determine the influence of these variables. CONCLUSIONS
Much of the research on stress and health out-
Implications for Medical/ comes challenges the traditional "medical model".
Professional Education The body and mind can no longer be considered
In 1998, Dr. Herbert Benson testified to a U.S. separate entities without interactive effects. A
Senate appropriations subcommittee that the patient's individual perception of stress can be
biggest barriers to integration of mind/body thera- related to adverse outcomes in multiple disease
pies (such as stress management) are "the lack of states, thus a better understanding of how to ame-
knowledge of the existing scientific data among liorate a patient's stress reactions to emotional,
health care providers" and "a bias against social, psychological, and environmental stimuli
mind/body interventions in medical care as being will enable health care providers to be more effec-
'soft' science." He further stated, "the full integra- tive with their patients. There is a need for curricu-
tion of mind/body, self-care medicine is completely lum reform that teaches health care providers from
compatible with existing health care approaches"50. the "biopsychosocial" perspective5, emphasizing
There is obviously a need for more medical/pro- not only new knowledge about stress and disease
fessional education on the relationship between but also new skills in stress reduction and more pos-
stress and health outcomes. Since the ultimate goal itive beliefs about mind/body medicine and its inte-
is not just to increase knowledge, but to increase the gration into the existing health care structure.
frequency with which health care providers discuss
stress management with their patients, providing REFERENCES
training on the most recent research may not be suf- 1. Cohen S, Kessler R, Gordon L. Measuring stress: a guide
ficient. Perhaps the key is to combine training in the for health and social scientists. New York, Oxford University
latest stress-related research and teaching actual Press. 1995.
2. Cohen S, Tyrrell D, Smith A. Psychological stress and sus-
stress reduction techniques. Having providers adopt ceptibility to the common cold. New England Journal of
their own stress reduction regimen may facilitate Medicine. 1991; 325: 606-12.
acceptance and confidence. This knowledge and 3. Sandberg S, Paton JY, Ahola S, McCann DC, McGuinness
skills acquisition might then result in attitude D, Hillary CR, et al. The role of acute and chronic stress in asth-
changes which affect the motivation (outcome eval- ma attacks in children. [see comments]. Lancet. 2000; 356: 982-7.
uation), self-efficacy, and outcome expectations of 4. Cutolo M, Villaggio B, Foppiani L, Briata M, Sulli A,
the provider, resulting in a much higher incidence of Pizzorni C, et al. The hypothalamic-pituitary-adrenal and gonadal
axes in rheumatoid arthritis. Annals of the New York Academy of
discussion of stress management with their patients. Sciences. 2000; 917: 835-43.
One barrier to increasing the frequency with 5. Dancey CP, Taghavi M, Fox RJ. The relationship between
which providers offer stress counseling is the lack of daily stress and symptoms of irritable bowel: A time-series
clinical practice guidelines. Many professional approach. Journal of Psychosomatic Research. 1998; 44: 537-45.
organizations, including the American Medical 6. Leserman J, Petitto JM, Golden RN, Gaynes BN, Gu H,
Association, the Agency for Health Care Research Perkins DO, et al. Impact of stressful life events, depression, social
and Quality, and the American College of Sports support, coping, and cortisol on progression to AIDS. American
Journal of Psychiatry. 2000; 157: 1221-8.
Medicine, provide clinical practice guidelines to help 7. Glaser R, Rabin B, Chesney M, Cohen S, Natelson B.
health care providers counsel their patients on smok- Stress-induced immunomodulation: implications for infectious
ing cessation, diet, and exercise. But the American diseases? JAMA. 1999; 281: 2268-70.
Medical Association, the American Institute of 8. Kiecolt-Glaser J, McGuire L, Robles T, Glaser R. Emotions,
Stress, the International Stress Management Associ- morbidity, and mortality: New perspectives from psychoneuroim-
ation and the American Psychological Association all munology. Annual Review of Psychology. 2002; 53: 83-107.
confirm that there are currently no professional 9. Carels RA, Blumenthal JA, Sherwood A. Emotional respon-
sivity during daily life: Relationship to psychosocial functioning
guidelines for health care providers on how to coun- and ambulatory blood pressure. International Journal of

843 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003
 TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

PsIychophv'siology. 2000; 36: 25-33. 29. Irvine MJ, Johnston DW, Jenner DA, Marie GV Relaxation
10. Rozanski A, Blumenthal J, Kaplan J. Impact of psycho- and stress management in the treatment of essential hypertension.
logical factors on the pathogenesis of cardiovascular disease and Jou-rnal of Psyvchosomatic Resear-ch. 1986; 30: 437-50.
implications for therapy. Circulation. 1999; 99: 2192-217. 30. Linden W, Chambers L. Clinical effectiveness of non-drug
11. Black P, Garbutt L. Stress, inflammation and cardiovascu- treatment for hypertension: A meta-analysis. Annals of Behavioral
lar disease. Journal of Psychosomatic Research. 2002; 52: 1-23. Medicine. 1994; 16: 35-45.
12. Cohen S, Hamrick N, Rodriguez M, Feldman P, Rabin B, 31. Linden W, Lenz JW, Con AH. Individualized stress man-
Manuck S. The stability of and intercorrelations among cardio- agement for primary hypertension: a randomized trial. Archives of
vascular, immune, endocrine, and psychological reactivity. Annals Internal Medicine. 2001; 161: 1071-80.
of Behavioral Medicine. 2000; 22: 171-79. 32. Schneider R, Staggers F, Alexander C, Sheppard W,
13. Gaskill SP, Williams K, Stern MP, Hazuda HP. Marital Rainforth M, Kondwani K, et al. A randomized controlled trial of
Stress Predicts the Incidence of Type 2 Diabetes in Mexican stress reduction for hypertension in older African Americans.
Americans and Non-Hispanic Whites. 2000. Hyper-tension1. 1995; 26: 820-27.
14. Moody-Ayers SY, Mutran EJ, Inouye SK. The impact of 33. Zurawski R, Smith T, Houston B. Stress management for
stress and perceived racism on diabetic control in older African essential hypertension: comparison with a minimally effective
Americans. 1999. treatment, predictors of response to treatment, and effects on reac-
15. Clark WC, Yang JC, Janal MN. Altered pain and visual tivity. Journal of Psychosoinatic Research. 1987; 31: 453-62.
sensitivity in humans: the effects of acute and chronic stress. 34. Boardway RH, Delamater AM, Tomakowsky J, Gutai JP
Annals of the New York Academv of Sciences. 1986; 467: 116-29. Stress management training for adolescents with diabetes. Journal
16. DePalma MT, Weisse CS. Psychological influences on of Pediatric Psychology. 1993; 18: 29-45.
pain perception and non-pharmacologic approaches to the treat- 35. McGrady A, Gerstenmaier L. Effect of biofeedback assist-
ment of pain. Journal of Hand Therapj. 1997; 10: 183-91. ed relaxation training on blood glucose levels in a type I insulin
17. Pertovaara A, Kemppainen P, Huopaniemi T, Johansson G. dependent diabetic. A case report. Jouri-nal of Therapeutic &
Pain and stress: correlation of stress hormone release to pain mod- Exrper imental Psychiatry. 1990; 21: 69-75.
ulation in man. Annals oJfClinical Research. 1987; 19: 83-6. 36. McGrady A, Graham G, Bailey B. Biofeedback-assisted
18. Servan-Schreiber D, Kolb R, Tabas G. The somatizing relaxation in insulin-dependent diabetes: a replication and exten-
patient. Primary Care; Clinics in Office Practice. 1999; 26: 225-42. sion study. Annals of Behavioral Medicine. 1996; 18: 185-89.
19. Cummings N, VandenBos G. The twenty years Kaiser- 37. Rosenbaum L. Biofeedback-assisted stress management
Permanente experience with psychotherapy and medical utiliza- for insulin-treated diabetes mellitus. Biofeedback and Self-
tion: implications for national health policy and national health Regutlation. 1983; 8: 519-32.
insurance. Health Policy Quarterly. 1981; 1: 159-75. 38. Bogaards M, Kuile Mt. Treatment of recurrent tension
20. Kroenke K, Mangelsdorff A. Common symptoms in headache: a meta-analytic review. Clinical Joutrnal of Pain. 1994;
ambulatory care: incidence, evaluation, therapy, and outcome. 10: 174-90.
American Journal ofJMedicine. 1989; 86: 262-6. 39. NIH. NIH Technology Assessment Panel on Integration of
21. McEwen B. Protective and damaging effects of stress Behavioral and Relaxation Approaches into the Treatment of
mediators. Nes England Jouri-nal oqfMedicine. 1998; 338: 171-9. Chronic Pain and Insomnia. Integration of behavioral and relax-
22. Benson H. The Relaxation Response. New York, Morrow. ation approaches into the treatment of chronic pain and insomnia.
1975. JAMA. 1996; 276: 313-8.
23. Hoffman J, Benson H, Arns P, al. e. Reduced sympathetic 40. Hellman C, Budd M, Borysenko J, McClelland D, Benson
nervous system responsivity associated with the relaxation H. A study of the effectiveness of two group behavioral medicine
response. Science. 1982; 215: 190-2. interventions for patients with psychosomatic complaints.
24. Smyth JM, Stone AA, Hurewitz A, Kaell A. Effects of Behavioral Medicine. 1990; 16: 165-73.
writing about stressful experiences on symptom reduction in 41. Francis ME, Pennebaker JW Putting stress into words: the
patients with asthma or rheumatoid arthritis: a randomized trial. impact of writing on physiological, absentee, and self-reported
JAMA. 1999; 281: 1304-9. emotional well-being measures. American Journal of Health
25. Bradley L, Young L, Anderson K, al. e. Effects of psycho- Promiiotioni. 1992; 6: 280-7.
logical therapy on pain behavior of rheumatoid arthritis patients. 42. Bandura A. Social foundations of thought and action: A social
Treatment outcome and six-month followup. Ar-thr-itis & cognitive theory. Englewood Cliffs, NJ, Prentice-Hall, Inc. 1986.
Rheumatism. 1987; 30: 1105-14. 43. Walsh J, Swangard D, Davis T, McPhee S. Exercise coun-
26. Parker JC, Smarr KL, Buckelew SP, Stucky-Ropp RC, seling by primary care physicians in the era of managed care.
Hewett JE, Johnson JC, et al. Effects of stress management on Amer-ican Jouirnal of Preventive Medicine. 1999; 16: 307-13.
clinical outcomes in rheumatoid arthritis. Ar-thr-itis & Rheumatism. 44. Ely J, Goerdt C, Bergus G, West C, Dawson J, Doebbeling
1995; 38: 1807-18. B. The effect of physician characteristics on compliance with adult
27. Amigo 1, Gonzalez A. Comparison of physical exercise preventive care guidelines. Family Medicine. 1998; 30: 34-9.
and muscle relaxation training in the treatment of mild essential 45. Davis D, O'Brien M, Freemantle N, Wolf F, Mazmanian PR
hypertension. Stress Medicine. 1997; 13: 59-65. Taylor-Vaisey A. Impact of formal continuing medical education:
28. Garcia-Vera MP, Labrador FJ, Sanz J. Stress-management do conferences, workshops, rounds, and other traditional continu-
training for essential hypertension: a controlled study. Applied ing education activities change physician behavior or health care
Psychophysiology and Biofeedback. 1997; 22: 261-83. outcomes'? JAMA. 1999; 282: 867-74.

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003 844
TRAINING, PERCEPTIONS, & PRACTICES FOR STRESS

Discover your passion

for excellence. Agaiw.
46. Huck S. Reading Statistics and Research, Third Edition.
New York, Longman. 2000.
47. Pedhazur E. Multiple regression in behavioral research.
Fort Worth, TX, Harcourt Brace. 1997. The
48. Carroll G. Mundane extreme environmental stress and
communities
we serve expect
African-American families: a case for recognizing different reali-
ties. Journal of Comparative Family Studies. 1998; 29: 271-84. the best care.
49. Anderson N, McNeilly M, Myers H. Autonomic reactivity Nothing less
and hypertension in blacks: a review and proposed model. will do.
Ethnicity & Disease. 1991; 1: 154-70. That's why
we look to
50. Benson H. Prepared statement by Herbert Benson. 1998.
51. Engel GL. The need for a new medical model: a challenge health care
for biomedicine. Science. 1977; 196: 129-36. professionals
like you.
Lovelace Sandia
We Welcome Your Comments Health System
The Journal of the National Medical provides.
Association welcomes your Letters to the Dr. Marilynne McKay outstanding
Lovelace Sandia Health System health care to
Editor about articles that appear in the JNMA or
issues relevant to minority health care. the region.
Our secret? 1)edicated, talented people
Address correspondence to Editor-in-Chief, who share our values: to provide care with
JNMA, 1012 Tenth St. NW, Washington, compassion and hope,- while practicing
DC 20001; fax (202) 371-1162; or e-mail superior teamwork. It's a recipe for success at
[email protected]. Lovelace Sandia Health System, a nationally
recognized regional health care organization
located in beautiful Albuquerque, NM.
We invite you to rediscover your passion
lllUNIVERSY OF for excellence in an
=~~~~~~qa environment of
OpotntyEpoe
PENNSYLVANIA uncomnpromi-sing innovation, with an
SCHOOL OF MEDICINE unyielding spirit of care. After all, our
patients are expecting the very best.
The Department of Pathology and Laboratory Medicine Currently, we offer opportunities in the
at the University of Pennsylvania's School of Medicine following a-reas:
seeks candidates for several Assistant, Associate and/or
Full Professor positions in the non-tenure clinician-educator Anesthesiology,* Cardiology
track. Rank will be commensurate with experience. Applicants Rheumatology * Pulmonology
must have an M.D. or M.D./Ph.D. degree and have
demonstrated excellent qualifications in education, research, Dermatology * Hospitalist
and clinical care. Eligibility/certification in Anatomic Lovelace Sandia Health System offers
Pathology from the American Board of Pathology is necessary.
competitive salary and benefits. To apply for a
Eligibility for an unrestricted medical license in the position, please forward your resume to:
Commonwealth of Pennsylvania is required. Dynamic Human Resources Department
individuals with a strong background in general surgical 1258 Ortiz Drive, S.E.
pathology and subspecialty interest/expertise including Albuquerque, NM 87108
but not limited to genitourinary/kidney pathology, Deborah Baca, Physician Recruiter
gastrointestinal/liver pathology, molecular diagnostics,
tissue imaging, or infectious disease related pathology I)avid Thomas, Physician Recruiter
preferred. Surgical pathology publication record and direct [email protected]
experience with modern diagnostic techniques required. (505) 262-7700
www.lovelace.com
Please submit curriculum vitae, a brief statement of research
interests, and three reference letters to:
John E. Tomaszewski, M.D.
Surgical Pathology Section
Hospital of the University of Pennsylvania ILOVELACE
Founders Pavilion, Room 6.042
3400 Spruce St., Phila, PA 19104/4283
SANDIA
HEATHSYSTEM
www.uphs.upenn.edu/path/JobOpps.html
Equal Opportunity/Affirmative Action Employer

845 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 95, NO. 9, SEPTEMBER 2003