Job Stressors among Female Physicians

Job Stressors among Female Physicians Relation to Having a Clinical Diagnosis of Hypertension ´, KAREN BELKIC ´, DANKA FILIPOVIC ´, NEDA JOCIC ´ OLESJA NEDIC Thirty-five female physicians with, and 74 without clinically-diagnosed hypertension completed the physicianspecific Occupational Stress Index (OSI) questionnaire in Novi Sad. Adjusting for covariates including BMI, an OSI high-demand score above the mean yielded an odds ratio (OR) of 3.14 (95% confidence interval [95%CI], 1.05–9.43) for hypertension. Among those with BMI > 26, long workhours and hazardous task performance yielded significantly elevated adjusted OR’s for hypertension. Overweight physicians without diagnosed hypertension were more often smokers. The strongest multivariate model for the favorable risk profile (FRP) was: non-smoker without diagnosed-hypertension, having a hobby and lower BMI, with total threat avoidant vigilance score below the mean showing the most significant OR (0.30, 95%CI, 0.12–0.78). Disturbances from other people and listening to emotionally disturbing occurrences also showed significant inverse multivariate associations with FRP. Diminishing work stressor burden should be part of hypertension and other disease prevention strategies for female physicians. Key words: hypertension; physicians; job stressors; women; Occupational Stress Index; obesity; smoking; threat avoidance vigilance; long work hours; Novi Sad, Serbia. INT J OCCUP ENVIRON HEALTH 2010;16:330–340

INTRODUCTION Social-historical patterns suggest that workplace conditions play an important role in hypertension.1,2 The empirical evidence is strongest for job strain, that is, the combination of high psychological demands and low decision-making latitude,3 as a risk-factor for hypertension.2,4,5 Women are more likely than men to have low

Received from: Ambulatory Health Care Center, Division for Occupational Health Protection, Novi Sad, Serbia (ON); Department of Oncology and Pathology, Karolinska Institute, Stockholm, Sweden (KB); Institute for Prevention Research, Keck School of Medicine, University of Southern California, Los Angeles, USA (KB); Center for Physiology, Faculty of Medicine, Novi Sad, Serbia DF); Center for Occupational Medicine, Faculty of Medicine, Novi Sad, Serbia (NJ). An abstract of some of the results from this paper has been presented to the XI Congress of Occupational Medicine of the Serbian Medical Society held October 2009 in Kopaonik. Send correspondence to: Dr. Karen Belkic´, MD, PhD, Physician Specialist in Internal Medicine, Adjunct Professor of Preventive Medicine, Department of Oncology and Pathology, Karolinska Institute, PO Box 260, Stockholm, SE171776, Sweden; email: . Disclosure: The authors declare no conflicts of interest.

330

decision-making latitude and high-strain jobs.6,7 Further, studies have shown that women are more likely than men to have low decision-making latitude and high-strain jobs.6,7 However, the effect of exposure to job strain upon blood pressure (BP), while observed in some studies,8 seems generally less pronounced among women.9–11 There have also been single-occupation studies among women in which job strain assessed using standard instruments was not found to impact on blood pressure. These null studies were carried-out predominantly among female health-professionals.12,13 On the other hand, among women there appears to be an association between number of years as a health professional (independent of age) and elevated BP (diastolic BP at night during “off-days”).14 Among physicians of both genders, BP elevations have been reported during particularly arduous working conditions such as 24-hour- or night-shifts in the emergency room, compared to less strenuous periods.15,16 Elevated BP and/or risk of hypertension have been reported in other stressful occupations, such as among air-traffic controllers,17 sea pilots,18 and professional drivers.19,20 The above-enumerated occupations have not been consistently categorized as high-strain,21 nor have standard assessment tools consistently detected withinoccupation relationships between exposure to high demands/low decision-making latitude and elevated BP. This is in part because many of the heaviest demands of these occupations are not detected by asking about “working fast” and “working hard,” as in the usual inquiries about job strain. One of the most abiding characteristics of air-traffic control, sea piloting, and mass-transit and other types of professional driving associated with increased hypertension risk is “threat avoidant vigilance” (TAV). Namely, a key aspect of these occupations is the need to very carefully process information, in order to be prepared to rapidly respond, since even the shortest lapse of attention, the smallest error, or the slightest delay in responding could potentially have serious or even fatal consequences.22,23 It is clear that TAV is a fundamental component of health professionals’ work. The concept of TAV emerges from the field of cognitive ergonomics, wherein work is viewed in terms of demands on mental resources and how these are controlled by the individual.24 For survival reasons, the nervous system selectively directs mental resources to threatening stimuli25 and for this reason, TAV represents an especially heavy burden upon the nervous system.

The Occupational Stress Index (OSI)26 is a model derived from this cognitive ergonomic perspective, and unlike job strain and other sociological models, incorporates dimensions such as TAV. The OSI takes into account the work environment as a whole, including task-level issues and work schedules, as well as physical conditions, chemical exposures, and broad organizational factors, all of which can all contribute to total stressor burden. The OSI model (Table 1) is arranged as a two-dimensional matrix: the vertical axis is composed of levels of information transmission27 and the stressor aspects are shown along the horizontal axis. The individual elements are summed into the OSI aspects, that are then summed into the total OSI score, reflecting the overall burden from work stressors. The OSI bridges a gap between two different approaches in occupational psychosocial research. One of these is the theory-based, generic approach, which is usually remote from work experiences, and often less helpful for assessing the within-occupation variance needed for developing intervention strategies in practice. The other, occupation-specific approach yields detailed information to identify areas for intervention. However, since this latter approach is focused on a single occupation, more generalizable conclusions from between-group analyses are often missed. Here, the OSI offers a potential solution. Occupation-specific questionnaires that are mutually compatible within the OSI theoretical framework allow for between-occupation comparisons. These more operationalized instruments also yield results that are more relevant to a given occupation. A physician-specific OSI questionnaire has been created and validated among various clinical specialties, and is “by-physicians-for-physicians.”26,28 We have used the physician-specific OSI questionnaire in studies of over 200 clinicians in Novi Sad, the capital of the Vojvodina region of Serbia (part of the former Yugoslavia), a region with high hypertension rates.29–31 Among the female physicians, scores for total OSI and TAV, as well as other aspects and individual elements of the OSI, showed significant multivariate associations with obesity and sedentary lifestyle,32 well-recognized risk factors for hypertension. Severe consequences of hypertension for health professionals of both genders in that region were seen in a seven-year follow-up study33 comparing 160 physicians and nurses to 122 hospital employees without clinical duties; the former had a higher risk (RR, 3.7; 95%CI, 1.6–8.6) for developing cardiovascular/cerebrovascular complications. These findings underscored the special etiological importance of occupational stressors in the progression from hypertension to ischemic heart disease (IHD)34,35 and highlight the urgency of finding more effective strategies to prevent hypertension among male and female health professionals. This is the overall goal of the present study. We focussed on female physicians, upon whom society is becoming increasingly dependent

VOL 16/NO 3, JUL/SEP 2010 • www.ijoeh.com

and who appear to be at increased risk for other workstressor related disorders, such as burnout.36 In this study, the physician-specific OSI questionnaire was used in order to compare female physicians with and without diagnosed hypertension in Novi Sad. The specific objective of this study was to identify associations between work stressors and hypertensive status. We hypothesized that individual stressors contributing to threat avoidant vigilance, high demand, and conflict would show a strong relation to hypertension. We also considered additive burden by examining the relation between OSI stressor aspects and hypertensive status. Our emphasis was upon modifiable workplace stressors that could be a focus for interventions.

METHODS Study Population This study was part of a larger investigation carried out between 2002 and 2004 among physicians with and without IHD or hypertension employed at the Novi Sad Clinical Center (NSCC).29 Physicians employed at the NSCC receive primary care at the Occupational Health Center located within the NSCC premises. Blood pressure measurement was obligatory for all visits to the Occupational Health Center. This was performed by the first author with a standardized protocol, using a calibrated mercury sphygmomanometer whose size was appropriate to right arm circumference, with the patient sitting after ≥ 5 minutes rest. Whenever systolic BP was > 130 mmHg or diastolic BP was ≥ 90, the patient was scheduled for two follow-up visits. Hypertension was diagnosed when mean levels of the three separate clinic measurements exceeded 130 mmHg systolic or ≥ 90 diastolic. In order to select participants for this current study, we retrospectively reviewed the Occupational Health Center records and identified all physicians with hypertension who were employed at the NSCC in 2002. There were forty female physicians aged 35 to 60 with diagnosed hypertension without secondary cause (secondary causes included pregnancy-induced hypertension; no participant was pregnant during the study) and without IHD or other structural cardiovascular disease (CVD). A random sample of 78 female physicians—also employed at the NSCC in 2002—within the same age group and without diagnosed hypertension or other CVD were identified as potential referents. The study was approved by Novi Sad’s Medical School Ethics Committee. Each eligible physician was contacted by the first author and invited to participate in the study. They were told the following: (1) the aim of the study was to assess working conditions and health for our profession in a study “by-physicians-for-physicians”; (2) the study was comprised of a questionnaire about working conditions, lifestyle, sociodemograph-

Job Stressors among Female Physicians • 331

ics, and the Minnesota Multiphasic Personality Inventory (the latter results reported elsewhere29) and included consent to verify questionnaire data with medical records; (3) there was complete freedom to withdraw from the study at any time with no consequences whatsoever; and (4) all data would be handled confidentially. Informed consent was obtained from all participants. Measures. Sociodemographic information including age, marital status, number of children, and homeownership were noted. We asked the participants whether they had a hobby and whether they had major family problem(s). Working years as a physician and specialty were noted, the latter categorized as: (1) surgeons/anesthesiologists, (2) non-surgical clinical-care, (3) diagnostic/preventive (such as radiologists and pathologists). To evaluate workplace stressors, the validated Yugoslav version26,28 of the physician-specific OSI questionnaire was used. Participants were asked to answer questions about working conditions over the last 10 years, or for as long as they had worked as a physician if it was less than 10 years. The scoring for each element of the generic OSI is from 0 to 2 (0: “not-present”; 2: “strongly-present”). However, for some elements of specific OSIs, the unchanging characteristics of the occupation are assigned fixed scores or the range of possible scores is narrowed. In the physician-specific OSI there are 12 fixed-score elements, indicating that these are unchanging characteristics of physicians’ work. For example, since potential human injury or fatality from an error is an essential feature of physicians’ work (TAV on the central decision-making level), this element was always scored maximally (2 points). There are 27 elements with narrowed-score ranges in the OSI for physicians.26 One example is hazardous task performance indicating threat avoidance on the task-performance level. The maximum score of 2 points is given if there are threats of violence from psychotic patients, risk of infection from close contact with blood/other bodily fluids, or work entailing exposure to ionizing radiation. This element is assigned a minimum score of 1 since even without these clearly dangerous situations, there are always some potentially hazardous tasks related to the work of physicians. Summated scores for the aspects and the total OSI were calculated, according to the OSI model shown as Table 1. The full OSI questionnaires and score sheets, as well as detailed information about the validity and reliability of the OSI instruments are found in References.26,32 Although the OSI is questionnaire-based, worksite measurements and other data can be incorporated to improve accuracy/precision.26 For the present study, information was available from expert-observations, worksite measurements and medical-records, about work hours, nightshifts, vacations, moonlighting, physical/chemical exposures, performance of invasive pro-

332 • Nedic´ et al.

cedures, and emergency work.29 Each OSI questionnaire was cross-validated with these other sources. Each participant was asked about current and evercigarette smoking. Current smokers were asked about the number of cigarettes smoked per day. Smoking data were cross-validated with medical records. We calculated BMI from reported height and weight. Each participant was asked when she last weighed herself; medical records were reviewed regarding obesity. We inquired about regular recreational physical activity (defined as one or more times per week) and daily alcohol consumption; medical records were reviewed regarding the latter.

STATISTICAL ANALYSIS Univariate and bivariate analysis of continuous/semicontinuous variables was performed using Pearson and Spearman correlations, 2-sample t-tests, and Mann-Whitney tests. Differences in dichotomous variables were examined with Yates-corrected 2 tests; if expected cellsize was less than five, dichotomous variables were examined by Fisher’s exact test. Multiple logistic regression was used to identify parsimonious sets of independent variables that explained the largest amount of variance for the outcome variables. When two or more OSI elements within a given aspect were each significant or near significant (p < 0.1) for the outcome variable, we tested the total aspect to determine whether the latter provided a more powerful multivariate model.

RESULTS Thirty-five (87.5%) physicians with hypertension agreed to participate. Thirty of these physicians were treated with antihypertensive medication and the other five only with recommendations for diet and other lifestyle changes. Altogether 95% (n = 74) of the physicians without hypertension agreed to participate. Three physicians refused because they declined to complete the Minnesota Multiphasic Personality Inventory. Another six did not return the questionnaire or did so incompletely, and so were not included. The five nonparticipants with diagnosed hypertension included one surgeon, two non-surgical clinical-care specialists, and two preventive/diagnostic specialists. Their mean age was 52.6 ± 4.8 (t = 0.76, compared to participant cases). The four non-participants without diagnosed hypertension had a mean age of 50.2 ± 1.5 (t = 0.96, compared to participant referents) and included one surgeon, one non-surgical clinical-care specialist, and two preventive/ diagnostic specialists. We compared sociodemographic and lifestyle risk factor data among physicians with and without diagnosed hypertension (Table 2). There were significant differences between the groups: physicians with hypertension were older, less often recreationally physically active, and had a higher mean BMI; more physicians

www.ijoeh.com • INT J OCCUP ENVIRON HEALTH

TABLE 2 Sociodemography and Lifestyle-Related Risk Factors among Female Physicians with and without Diagnosed Hypertension With Hypertension (n = 35) ___________________ Continuous/ Semi-continuous Variables Age BMI (kg/m2) Number of children Number of cigarettes smoked/day Coding 0: None 1: 1–10 2: 11–20 3: over 20 Cups of coffee consumed/day Coding 0: None 1: 1–5 2: 6–10 3: > 10 Number of institutions for work Coding 1: 1 2: 2 3: ≥ 3 Number years worked as a physician Coding 1: < 10 y 2: 10–19 3: 20–29 4: ≥ 30 Categorical Variables Physician category Surgical/anesthesia Non-surgical clinical care Diagnostic/preventive Married Homeowner Major family problem Has a hobby Regular recreational physical activity Smoking status Current smoker Heavy smoker (> 20 cigarettes/day) Former smoker Never smoker Overweight BMI > 26 Consumes alcohol daily

Without Hypertension (n = 74) ______________________

Mean ± SD (Range)

Level of Significance

Mean ± SD (Range)

50.6 ± 6.2 (35–60) 27.4 ± 4.5 (18–37.2) 1.3 ± 0.8 (0–3) 0.6 ± 1.0 (0–3)

*

47.8 ± 7.0 (35–60) 23.6 ± 2.9 (18–31.9) 1.3 ± 0.9 (0–3) 0.7 ± 1.0 (0–3)

***

0.9 ± 0.3 (0–1)

1.0 ± 0.4 (0–3)

1.3 ± 0.4 (1–2)

1.2 ± 0.5 (1–3)

2.9 ± 0.8 (1–4)

2.7 ± 0.8 (1–4)

n (%)

n (%)

10 (29%) 20 (57%) 5 (14%) 22 (63%) 29 (83%) 4 (11%) 13 (37%) 3 (9%) 10 3 4 21

*

(29%) (9%) (11%) (60%)

22 (63%) 3 (9%)

22 35 17 58 59 12 38 22

(30%) (47%) (23%) (78%) (80%) (16%) (51%) (30%)

25 (34%) 7 (9.5%) 8 (11%) 41 (55%) ***

14 (19%) 3 (4%)

Note: The statistical tests used and the level of significance by which the two groups differed were as follows: 2-sample t-test for continuous/semi-continuous variables, 2 test or Fisher’s exact test for categorical variables. All 2 tests are with 1 degree of freedom and Yates correction. All significance levels are 2-sided: *p < 0.05, **p < 0.01, ***p < 0.001.

with hypertension were overweight (had a BMI over 26). Altogether 94% of the participating physicians had weighed themselves within the last two months.


The mean total high-demand scores were significantly higher among the physicians with hypertension (Table 3). The physicians with hypertension had significantly


TABLE 3 Total Occupational Stress Index (OSI), High Demands, and Components of OSI High Demands that Differ Significantly between Female Physicians with and without Diagnosed Hypertension With Hypertension (n = 35) ___________________

Without Hypertension (n = 74) ______________________

Mean ± SD (Range)

Level of Significance

Mean ± SD (Range)

Total input level OSI high demands

9.4 ± 1.2 (6–11)

*

8.7 ± 1.4 (6.3–11)

Heterogeneous information Coding Range = 1.5 to 2, scored based on heterogeneity of of clinical work and whether the physician also does teaching, administration, or research

1.7 ± 0.2 (1.5–2)

*

1.6 ± 0.2 (1.5–2)

Burden on the visual system Coding 0.5 = Primary care clinician or psychiatrist and does not perform many procedures 1 = Surgeon, radiologist, pathologist, or dermatologist or performs many procedures

0.9 ± 0.2 (0.5–1)

*

0.8 ± 0.3 (0.5–1)

High frequency of incoming signals Coding Range = 1 to 2, scored based on working in the emergency room, patient load (outpatient, inpatient, intensive care), supervisory, and teaching load

1.7 ± 0.3 (1–2)

*

1.5 ± 0.4 (1–2)

Total task performance level OSI high demands

5.9 ± 1.5 (2.5–8)

*

5.0 ± 1.9 (2.5–8)

Heterogeneous tasks Coding Range = 1 to 2, scored based on the number of different diagnostic procedures performed as well as teaching, administration, and/or other tasks outside the realm of clinical work

1.5 ± 0.4 (1–2)

*

1.3 ± 0.4 (1–2)

Complex tasks Coding Range = 1 to 2, scored based on performance of non-invasive and invasive diagnostic procedures, surgical and other invasive therapeutic interventions

1.5 ± 0.4 (1–2)

*

1.3 ± 0.4 (1–2)

Total general level high demands

5.0 ± 2.3 (0 –9.5)

*

4.0 ± 2.5 (0–9)

Long work hours Coding 0: ≤ 40 hours/week 2: Frequently ≥ 48 hours/week or occasionally ≥ 60 hours/week

1.6 ± 0.7 (0 –2)

**

1.1 ± 0.9 (0 –2)

OSI total high demands

27.9 ± 4.3 (16–34)

*

25.3 ± 5.2 (15.8–34)

Variables

Total OSI

79.7. ± 11.1 (46.8–101.3)

75.8 ± 12.0 (50.8–103.3)

Note: Two-sample t-test used; all significance levels 2-sided: *p < 0.05, **p < 0.01.

higher input high-demand scores (that is, they worked with more heterogeneous information, and had a heavier burden on the visual system and a greater frequency of incoming signals), higher output high-demand scores (that is, they performed more heterogeneous and complex tasks), and general high demand scores (that is, they had longer work hours). Although the mean total OSI scores were higher among the physicians with hypertension, this was not statistically significant.


Salient bivariate findings were: age and years as a physician were closely correlated (r = 0.90; p < 0.001); mean BMI was lower in current smokers (23.5 ± 3.3) than in those not currently smoking (25.5 ± 4.0) (twosample t-test, p = 0.01); and fewer than expected current smokers had a hobby (2 = 5.8; p = 0.02). Total avoidance scores were higher among current smokers (9.7 ± 2.2) than those not currently smoking (8.5 ± 2.4) (2-sample t-test, p = 0.009). There was no correlation


( = 0.07) between BMI and the date that the physician weighed herself. We conducted subgroup analysis among overweight physicians (that is, those with a BMI greater than 26). Overweight physicians without hypertension smoked more cigarettes than those with this diagnosis (p = 0.03). More than the expected number of overweight physicians without hypertension were current smokers (Fisher’s exact test, p = 0.06). In this subgroup, the mean OSI total high-demand scores were 27.5 ± 4.0 among those with hypertension versus 23.2 ± 5.4 without hypertension (p = 0.001). The input and output levels of high demands were significantly higher among those with hypertension, as were long-work-hour scores. The mean scores for an element of strictness (fixed body position) and of avoidance (hazardous task performance) were higher among overweight physicians with hypertension (p = 0.02). We also reported univariate findings for the favorable composite risk variable (that is, no diagnosed hypertension and non-smoker). Sociodemographic, life-style related risk factors, and OSI data are reported for the physicians with and without this favorable composite risk status (Table 4). The physicians who were currently non-smokers without diagnosed hypertension had a significantly lower mean BMI and borderline significantly more often had a hobby. The mean total OSI scores were significantly lower, as were the mean total high demand and avoidance scores among the non-smoking physicians without hypertension. The physicians with favorable risk status also had significantly lower mean scores for inadequate pay and fewer interruptions from other people during work. The results of multiple logistic regression analysis identifying sets of independent variables showing the strongest associations with diagnosed hypertension are shown in Table 5. For the entire sample, this included BMI and having an OSI total high-demand score above the mean, adjusting for non-significant covariates: age, marital status, regular recreational physical activity, and daily alcohol consumption. For physicians with BMI greater than 26, models yielding significant associations were: total output level high demands, hazardous task performance, total input level high demands, and long work hours, adjusting for age and number of cigarettes smoked daily. Three multiple logistic regression models for physicians who were non-smokers without diagnosed hypertension (that is, those with favorable risk status) are also shown in Table 5. Having a total avoidance score below the mean, low BMI, and having a hobby yielded the most powerful model for this favorable status, after adjusting for non-significant covariates: age, marital status, regular recreational physical activity, and daily alcohol consumption. Disturbances from other people also generated a significant model, when the same nonoccupational factors were included. Within the avoid-


ance aspect the individual element, “listening to emotionally-disturbing occurrences,” was also significant when the same non-occupational factors were included in the multiple logistic regression model.

DISCUSSION High BMI was the most powerful risk factor for having hypertension among the physicians in this study. As previously reported,32 the total burden of work stressors assessed by the OSI showed a significant independent association with high BMI among Novi Sad female physicians. Thus, in the pathway to hypertension, work-stressors play an important role through their relation to obesity. The present results suggest that workplace factors also act independently of this pathway, since the association between OSI total highdemand scores showed a significant association with hypertension when BMI was included in the multivariate model. Subgroup analysis among overweight physicians provides further insights. The long-work-hours score was singled out as a factor showing multivariate association with hypertension. This is consistent with populationbased data linking long work hours to prevalence of self-reported hypertension37 and data showing increased worksite blood pressure among physicians who worked long hours.16 Our results suggest that once a female physician becomes overweight, her chances of being diagnosed with hypertension increase if she works long hours. Hazardous task performance from threats of violence from psychotic patients, risk of infection from contact with blood/other body-fluids, or work entailing exposure to ionizing radiation also showed a multivariate association with hypertension among the overweight physicians. It seems plausible that the heightened arousal associated with these hazards, as well as prolonged exposure to the exigencies of work accelerated the development of hypertension among these overweight physicians. However, compared to those with this combined morbidity, overweight physicians without diagnosed hypertension tended to be current smokers and smoked more heavily. In other words, overweight physicians who were smokers were less likely to be diagnosed with hypertension. We therefore constructed the composite variable reflecting an overall more favorable risk profile: being a non-smoker without diagnosed hypertension. Fewer than half of the participants belonged to this category. Together with having a hobby and lower BMI, a lower total avoidance score provided the strongest multivariate model for this favorable risk profile. Having a hobby can be considered to be a buffer, providing a healthy release from the pressures of work, and therefore preventing harmful activity, such as smoking. We previously reported that among female physicians, total avoidance scores were associated with


TABLE 4 Female Physicians with the Favorable Composite Risk Variablea Compared with the Other Participants Regarding Sociodemography, Lifestyle-Related Risk Factors, and the Occupational Stress Index (OSI) Hypertension and/or Currently Smoking (n = 60) ___________________ Continuous/ Semi-continuous Variables

Mean ± SD (Range)

Age

48.6± 6.3 (35–60)

BMI (kg/m2)

25.7 ± 4.6 (18–37.2)

Number of children

1.3 ± 0.8 (0–3)

OSI Inadequate pay (element of underload)

No Hypertension and Not Currently Smoking (n = 49) ______________________ Level of Significance

Mean ± SD (Range) 48.9 ± 7.4 (35–60)

**

23.9 ± 2.6 (19.9–30.8) 1.3 ± 0.9 (0–3)

1.5 ± 0.5 (0–2)

**

1.3 ± 0.6 (0–2)

1.6 ± 0.6 (0–2)

**

1.3 ± 0.6 (0.5–2)

Total high demands

27.1 ± 4.6 (16–34)

**

25.1 ± 5.3 (15.8–33.5)

Total avoidance

9.5 ± 2.2 (4–16)

***

8.1 ± 2.5 (4–13.5)

79.1 ± 10.2 (46.8–101.3)

**

74.6 ± 13.2 (50.8–103.3)

Interruptions from other people (element of conflict)

Total OSI Categorical Variables

n (%)

n (%)

Physician Category: Surgical/anesthesia Non-surgical clinical care Diagnostic/preventive

18 (30%) 31 (52%) 11 (18%)

14 (29%) 24 (49%) 11 (22%)

Married Major family problem Has a hobby Regular recreational physical activity Consumes alcohol daily

42 (70%) 10 (17%) 23 (38%) 10 (17%) 5 (8%)

38 (78%) 6 (12%) 28 (57%) 15 (31%) 1 (2%)

*

Note: The statistical tests used and the level of significance by which the two groups differ are as follows: 2-sample t-test for continuous/semi-continuous variables; for categorical variables: all 2 test are with 1 degree of freedom and with Yates correction; Fisher’s exact test was used if there were cells with expected values < 5. All significance levels are 2-sided: *p < 0.08, **p < 0.05, ***p < 0.01. a The favorable composite risk variable: no diagnosed hypertension and not a current smoker

current and heavy smoking.32 Notwithstanding the fact that for all physicians a wrong decision can lead to injury or fatality, and vigilance is required to avoid these consequences, there is considerable variability in the weight of the overall TAV burden that physicians encounter, such as how often they listen to emotionallydisturbing accounts, encounter visually-disturbing scenes, observe or suffer work-related physical injury, have had patients commit suicide, or have testified in court in relation to work. Many of these exposures, while routine parts of medical practice, can be broadly considered traumatic events of varying intensities. Namely, they entail experiencing or witnessing events involving actual or threatened death or serious injury or a threat to the integrity of oneself or others. Smoking is often used to regulate arousal and counteract unpleasant emotional states,38 and tobacco


use increases with exposure to threatening situations.39 We now find the chances that a female physician has hypertension and/or is a current smoker are related to the aversive burden she faces. Notably, a higher score on listening to emotional-disturbing occurrences showed an independent association with the unfavorable risk profile. A separate multivariate model identified disturbances from other people as a significant factor that lowers the chances for a physician being a non-smoker and without diagnosed hypertension. The burden from the above-mentioned factors could be lowered if more attention were paid to protecting the physician from unnecessary interruptions, as well as realizing the need for limits on listening to emotionallydisturbing occurrences and exposure to other aversive situations. There should be further research on, and


TABLE 5 Multiple Logistic Regression Models among Female Physicians for Factors Associated with Diagnosed Hypertension and for the Favorable Composite Risk Variablea Number of Physicians Included and Endpoint

Model 2 (p level)

Significant Independent Variables

Adjusted Odds Ratio 95%CI

p-value

Entire sample of the present study n = 35 with hypertension n = 74 without

38.7b (p = 0.0000008) OSI total high demand Score > 26.2 (> mean)

3.14

1.05–9.43

0.04

BMI ≥ 26

10.32

3.42–31.15

0.0003

Long work hourse (High demand aspect, general level)

2.76

1.03–7.40

0.04

12.1c (p = 0.007)

Hazardous task performancee (Avoidance aspect, task performance level)

12.0

1.4–103

0.02

27.8b (p = 0.0002)

OSI total avoidance score > 8.9 (> mean) BMI ≥ 26

0.30

0.12–0.78

0.01

0.17

0.06–0.49

0.0009

Endpoint: Diagnosed Hypertension Subgroup analyses: BMI > 26 10.5c (p = 0.01) n = 22 with n = 14 without Endpoint: Diagnosed Hypertension

Entire sample of the present study n = 49 (no hypertension and non-smoker) n = 60 (hypertension and/or smoker) Endpoint: Favorable composite risk variablea

Has a hobby

3.10

1.22–7.87

0.02

26.3d (p = 0.0004)

Interruptions from other peoplee (Conflict aspect, task performance level)

0.45

0.22–0.92

0.03

25.6d (p = 0.0006)

Frequently listens to emotionally-disturbing occurrencese (Avoidance aspect, input level)

0.54

0.30–0.97

0.04

a

The favorable composite risk variable: no diagnosed hypertension and not a current smoker. Besides the significant independent variables, the model includes age, marital status, regular recreational physical activity, and daily alcohol consumption as covariates that were not statistically significant. c Besides the significant independent variable, the model includes age and number of cigarettes smoked/day as covariates that were not statistically significant. d The model includes age, marital status, regular recreational physical activity, and daily alcohol consumption as covariates that were not statistically significant. BMI ≥ 26 and has a hobby were included in the model and yielded significant ORs that were not listed to avoid repetitive information. e The OR is given for a unit change in the value of the independent variable. b

exploration of preventive measures to attenuate exposure to such burdens. One of the particular strengths of the present study was that the participation rate was very high, attesting to a strong motivation to improve the health and working conditions for female physicians in Novi Sad. Since only actively-employed physicians were included, there was likely some healthy worker selection, which could bias results to the null. To minimize information bias, physicians with hypertension were not explicitly told they were “cases.” Cross-validation of self-reported data with external


sources further helped guard against report bias. Moreover, the OSI queries are phrased in a concrete, neutral way, which minimized subjectivity.26 The number of children did not differ between physicians with and without hypertension; we included marital status as a covariate, thereby striving to account for home and family obligations. However, this burden could be assessed in more detail. On the other hand, having a hobby could indirectly indicate a lower burden of work at home. Body mass index was calculated from self-reported height and weight, but nearly all the physicians had


weighed themselves recently and there was no correlation with weight date, suggesting that these data are reliable. Nevertheless, objective measures would be preferred. A major endpoint of the present study was clinically diagnosed hypertension. Systematic surveillance was ensured by measurement of blood pressure at all clinic visits with a standardized protocol. Requiring three separate visits before this diagnosis was made would tend to minimize the white-coat effect.40,41 Since the Occupational Health Center is physically located within the physicians’ workplace, the blood pressure measurements might have been more akin to workplace pointestimates42 than to casual clinic readings. Nevertheless, hidden hypertension2,41 among the referents cannot be ruled out. The disproportionately adverse consequences of hypertension among clinicians,33 and the relation between working conditions and hypertension among female physicians found here, argue in favor of performing ambulatory blood pressures monitoring at work for earlier detection of hypertension. Existing effects may not have been detected due to power limitations; larger investigations are needed. Without longitudinal data, one must be cautious in making inferences about the temporal nature of observed associations (that is, whether exposure to work stressors preceded and therefore caused hypertension). Randomized intervention trials in which the implicated work stressors were ameliorated would be a practical way to further examine their etiologic association with hypertension and the evaluated risk behaviors, and at the same time test the efficacy of prevention strategies.

CONCLUSIONS Physicians are on the frontlines of the battle against hypertension; female physicians are increasingly involved in providing surveillance diagnostics and care of this disorder. Female physicians are by no means spared from this epidemic. Stressors in female physicians’ working environments appear to play an important role in the development of hypertension. Lowering cognitive demands, and especially avoiding long work hours, is an intervention strategy that should be examined. Our findings also suggest that efforts to help female physicians maintain normal blood pressure levels and body weight need to be coupled with efforts to prevent smoking. Diminishing the total threat avoidance burden, in particular, limiting exposure to emotionally-disturbing situations and to hazardous tasks, could be important for maintaining normal blood pressure and non-smoking status. Minimizing interruptions during work is a feasible intervention that might be also effective, according to the results of this study. Insights gleaned by addressing how our working conditions impact upon hypertension and IHD risk could help improve not only our own health, but that of our patients as well.


We are grateful to our colleagues who set aside their own free time to participate in this study, providing concrete suggestions as to how to improve health protection in our profession. We would also like to thank Boris Nedic´ for his help during the preparation of the data sets. Dr. Belkic´ would like to thank the Signe and Olof Wallenius Stiftelse, Cancerfonden, and the King Gustav the Fifth’s Jubilee Foundation for support of her research activity.

References 1. Schnall P, Belkic´ K, Landsbergis PA, Baker D. Why the workplace and cardiovascular disease? Occup Med. 2000; 15:1-5. 2. Landsbergis P, Schnall P, Belkic´ K, Schwartz J, Baker D, Pickering T. Working conditions and masked (hidden) hypertension: Insights into the global epidemic of hypertension. Scand J Work Environ Health 2008. (Suppl 6): 41-51. 3. Karasek R. Job demands, job decision latitude and mental strain: Implications for job redesign. Adm Sci Q. 1979; 24: 285308. 4. Schnall PL, Schwartz J, Landsbergis PA, Warren K, Pickering TG. A longitudinal study of job strain and ambulatory blood pressure: Results of a three-year follow-up. Psychosom Med. 1998; 60: 697-706 5. Markovitz JH, Matthews KA, Whooley M, Lewis CE, Greenlund KJ. Increases in job strain are associated with incident hypertension in the CARDIA Study. Ann Behav Med. 2004; 28: 4-9. 6. Hall EM. Gender, work control, and stress: A theoretical discussion and an empirical test. Int J Health Serv. 1989; 19:725-745. 7. Karasek RA, Theorell T. The Demand-Control-Support Model and CVD. Occup Med. 2000; 15: 78-83. 8. Brisson C. Women, work and cardiovascular disease. Occup Med. 2000; 15: 49-57. 9. Cesana G, Sega R, Ferrario M, Chiodini P, Corrao G, Mancia G. Job strain and blood pressure in employed men and women: A pooled analysis of four northern Italian population samples. Psychosom Med. 2003; 65: 558-63. 10. Guimont C, Brisson C, Dagenais GR, et al. Effects of job strain on blood pressure: A prospective study of male and female white-collar workers. Am J Public Health. 2006; 96: 1436-43. 11. Öhlin B, Berglund G, Rosvall M, Nilsson PM. Job strain in men, but not in women, predicts a significant rise in blood pressure after 6.5 years of follow-up. J Hypertens. 2007; 25: 525-31. 12. Brown DE, James GD, Nordloh L, Jones AA. Job strain and physiological stress responses in nurses and nurse’s aides: Predictors of daily blood pressure variability. Blood Press Monit. 2003; 8: 237-42. 13. Riese H, Van Doornen LJ, Houtman IL, De Geus EJ. Job strain in relation to ambulatory blood pressure, heart rate, and heart rate variability among female nurses. Scand J Work Environ Health. 2004; 30: 477-85. 14. Goldstein IB, Shapiro D, Chicz-DeMet A, Guthrie D. Ambulatory blood pressure, heart rate and neuroendocrine responses in women nurses during work and off work days. Psychosom Med. 1999; 61: 387-396. 15. Adams SL, Roxe DM, Weiss J, Zhang F, Rosenthal JE. Ambulatory blood pressure and Holter monitoring of emergency physicians before, during, and after a night shift. Acad Emerg Med. 1998; 5: 871-877. 16. Fialho G, Cavichio L, Povoa R, Pimenta J. Effects of 24-h shift work in the emergency room on ambulatory blood pressure monitoring values of medical residents. Am J Hypertens. 2006; 19: 1005-9. 17. Cobb S, Rose RM. Hypertension, peptic ulcer disease and diabetes in air traffic controllers. JAMA. 1973; 224: 489-492. 18. Erikssen J, Johansen AH, Rodahl K. Coronary heart disease in Norwegian sea-pilots: Part of the occupational hazard? Acta Med Scand. Suppl. 1981; 645: 79-83. 19. Belkic´ K, Emdad R, Theorell T. Occupational profile and cardiac risk: Possible mechanisms and implications for professional drivers. Int J Occup Environ Health 1998; 11: 37-57. 20. Greiner BA, Krause N, Ragland D, Fisher JM. Occupational stressors and hypertension: A multi-method study using observer-based job analysis and self-reports in urban transit operators. Soc Sci Med. 2004; 59: 1081-1094.


21. Karasek RA, Theorell T. Healthy work: Stress, productivity and the reconstruction of working life. New York; Basic Books; 1990. 22. Fuller R. A conceptualization of driving behaviour as threat avoidance. Ergon. 1984; 27: 1139-1155. 23. Belkic´ K, Savic´ C, Djordjevic´ M, Ugljesˇic´ M, Mickovic´ LJ. Eventrelated potentials in professional drivers: Heightened sensitivity to cognitively relevant visual signals. Physiol Behav. 1992; 52: 423-427. 24. Singleton WT. Introduction to ergonomics. Geneva: World Health Organization; 1974. 25. Friedman MJ, Charney DS, Deutch AY. Neurobiological and clinical consequences of stress: From normal adaptation to PostTraumatic Stress Disorder. Lippincott-Raven: Philadelphia; 1995. 26. Belkic´ K. The Occupational Stress Index: An approach derived from cognitive ergonomics and brain research for clinical practice. Cambridge: Cambridge International Science Publishing; 2003. 27. Welford AT. The measurement of sensory-motor performance: Survey and reappraisal of twelve years’ progress. Ergon. 1960; 3: 189-230. 28. Savic´ Cˇ. The physician and stress. Proceedings of the 17th Medical Meeting of the Medical Society of Vojvodina; 2002. p. 52-54. [Serbian]. 29. Nedic´ O. Occupational Stressors and Physician Health. [Dissertation.] University of Novi Sad, Novi Sad, Serbia; 2006. 228p. [Serbian]. 30. National Health Survey of Serbia. Serbia: Ministry of Health; 2006. 31. Jakovljevic´ D, Maksimovic´ J, Petrovic´ V, Planojevic´ M. Health and health risks among women of Novi Sad in the MONICA Project. Report to the Institute for Cardiovascular Diseases. Sremska Kamenica; 2005. 32. Belkic´ K, Nedic´ O. Workplace stressors and lifestyle-related


33. 34.

35.

36.

37. 38. 39. 40. 41.

42.

cancer risk factors among female physicians: Assessment using the Occupational Stress Index. J Occup Health 2007; 49:61-71. Nedic´ O, Filipovic´ D, Solak Z. Job stress and cardiovascular diseases among health workers. Med Pregl. 2001; 423-431. Schwartz J, Pickering T, Landsbergis P. Work-related stress and blood pressure: Current theoretical models and considerations from a behavioral medicine perspective. J Occup Health Psychol. 1996; 1: 287-310. Uchiyama S, Kurasawa T, Sekizawa T, Nakatsuka H. Job strain and risk of cardiovascular events in treated hypertensive Japanese workers: Hypertension follow-up group study. J Occup Health 2005; 47: 102-11. McMurray JE, Linzer M, Konrad TR, Douglas J, Shugerman R, Nelson K. The work lives of women physicians: Results from the physician work life study. The SGIM Career Satisfaction Study Group. J Gen Intern Med. 2000; 15: 372-80. Yang H, Schnall P, Jauregui M, Su T, Baker D. Work hours and self-reported hypertension among working people in California. Hypertension. 2006; 48: 744-750. Landsbergis PA, Schnall PL, Dietz DK, Warren K, Pickering TG, Schwartz JE. Job strain and health behaviors: Results of a prospective study. Am J Health Promot. 1998; 12: 237-245. Smith D, Devine S, Leggat P, Ishitake T. Alcohol and tobacco consumption among police officers. Kurume Med J. 2005; 52: 63-65. Jula A, Puukka P, Karanko H. Multiple clinic and home blood pressure measurements versus ambulatory blood pressure monitoring. Hypertension 1999; 34: 261-266. Belkic´ K, Schnall PL, Landsbergis PA, Schwartz JE, Gerber L, Baker D, Pickering TG. Hypertension at the workplace—an occult disease? The need for work site surveillance. Adv Psychosom Med. 2001; 22: 116-138. Schnall P, Belkic´ K. Point estimates of blood pressure at the worksite. Occup Med. 2000; 15:203-208.
