Franke, Warren D. PhD; Ramey, Sandra L. PhD; Shelley, Mack C. II, PhD
Few researchers have studied cardiovascular disease (CVD) morbidity among law enforcement officers, 1,2 although CVD mortality has been assessed relatively frequently. 3–11 In assessing mortality, some 3–7 but not all 8–11 researchers have found increased prevalence among law enforcement cohorts compared to control groups. In the United States, mortality due to CVD is declining more rapidly than is CVD prevalence; therefore, assessments of the former may present a biased picture of the relationship between CVD and the law enforcement profession. Nevertheless, Sparrow and colleagues 1 found that morbidity among active duty law enforcement officers (LEOs) did not differ from morbidity in the general population, although others found retired LEOs to have increased CVD morbidity after considering conventional risk factors. 2
A heightened presence of CVD risk factors, such as hypertension, hypercholesterolemia, and obesity, may account for an increased prevalence of CVD in the law enforcement community. 12–22 However, unique aspects of this profession also may contribute to the increased risk. One potential risk factor is stress. 4,12,13,23–27 LEOs often are assumed to have higher stress levels than other occupations, 26 and stress has been associated with negative health-related effects, 4,12,13,23–27 including cardiovascular diseases. 28 Sources of job-related stress within the profession include life events, 13 shift work, 29 job demands, excessive and unnecessary amounts of paperwork, lack of input into decision-making, and communication issues within the law enforcement organization. 26 Regardless of its origins, perceived stress among LEOs can be an important and significant predictor of negative health consequences. 27
As with assessments of the relationship between CVD morbidity and the law enforcement profession in general, very little research has examined specifically the effects of stress on CVD morbidity or risk factors in the LEO population. Consequently, the objectives of this study were 2-fold. First, using a much larger group than heretofore examined, we wanted to determine whether CVD morbidity was higher among employed law enforcement officers than in the general population, and to identify what conventional risk factors affected this relationship. Second, in the law enforcement group only, we wanted to determine whether perceived stress was a significant contributor to the prevalence of CVD or CVD risk factors.
General Experimental Design
The experimental design of this study consisted of two parts. First, sworn LEOs in nine states completed a questionnaire evaluating the presence of CVD, CVD risk factors, and perceived stress. Second, these responses were compared to data obtained from the general public in the Behavioral Risk Factor Surveillance System (BRFSS) questionnaire administered in the respective states in 1999. The latter data were obtained from the BRFSS coordinators in the states in which LEOs were surveyed and from data available on CD-ROM from the Centers for Disease Control and Prevention (CDC).
The BRFSS was developed by the CDC and is administered under the auspices of state Departments of Public Health. The survey is designed to determine health risk behaviors of adult Americans using a standardized series of questions. Adults in each state are randomly selected to complete a telephone survey in which they are asked to respond to those questions. An optional series of questions specifically queries the presence of CVD; 4 of the 9 states involved in the present study also used this series in 1999 (see http://www.cdc.gov/brfss for complete information on the BRFSS).
The survey administered to the officers used questions from relevant BRFSS modules. For example, the BRFSS has 3 questions specifically related to hypertension awareness (“About how long has it been since you last had your blood pressure taken by a doctor, nurse, or other health professional?” “Have you ever been told by a doctor, nurse, or other health professional that you have high blood pressure?” “Have you been told on more than one occasion that your blood pressure was high, or have you been told this only once?”) These three questions were used in the survey sent to the LEOs, as were similar BRFSS questions inquiring about cholesterol awareness, tobacco use, diabetes, and the presence of CVD. Physical activity was assessed using four questions from the BRFSS designed to assess the frequency, intensity, duration, and type of the most common activity performed in the past month. Self-reported height and weight were used to calculate body mass index (BMI), with “overweight” defined as a BMI > 25 kg/m2. Age also was self-reported. The presence of CVD was defined as whether a physician had ever informed the subject that he had; (1) coronary heart disease or angina; (2) a myocardial infarction or; (3) a stroke. This definition was the same as that employed in the BRFSS.
As part of their survey, LEOs also completed the Perceived Stress Scale [(PSS) 30], which measured LEOs’ perception of stress by quantifying the degree to which environmental demands exceeded abilities to cope during the past month. 30 It is influenced by a wide variety of factors, and consequently works well as a general measure of perceived stress. 31 In a law enforcement group, it has been shown to be a significant predictor of negative consequences of stress. 27 The PSS is a 14-item scale using a 5-point Likert scoring system. Scores can range from 0 to 56, with higher scores denoting greater perceived stress.
Potential subjects were the 4627 sworn male LEOs currently employed by nine state Departments of Public Safety (Iowa, Minnesota, Nebraska, Michigan, Missouri, North Dakota, South Dakota, Ohio, and Oklahoma). While the population distribution and demographics vary among these states, the job tasks of these officers are similar; eg, a highway patrol officer has primary responsibility for enforcing motor vehicle laws on major thoroughfares, while a criminal investigator generally investigates crimes against persons or property. The control group initially consisted of 9650 male respondents from the general public who answered the 1999 BRFSS questionnaire in the same states as the available LEOs. To insure comparability to the LEO group, respondents in the control group were limited to males with incomes generally similar to those of the officers and in the same age range as the officers (21 to 65 years). Consequently, 8046 respondents to the BRFSS met the criteria for inclusion in the analyses. Four of the 9 states (Iowa, Ohio, Oklahoma, and North Dakota) included questions related to CVD prevalence on the 1999 BRFSS. In those 4 states, there were 2722 employed officers, while 2575 respondents to the BRFSS met the inclusion criteria for this study.
Subjects were considered to have provided informed consent by completing the written survey. The study was approved by the respective state Departments of Public Safety and was conducted in accordance with the ethical standards for research involving human subjects of the Institutional Review Board at Iowa State University.
Initial screening of the data was conducted for missing values and outliers. Outliers were recoded as missing data. All variables on the BRFSS and LEO surveys were recoded to treat “don’t know” or “refused” responses as missing data. LEO data from the 9 states were weighted to adjust for differences in response rates across states. The weight for each state was calculated with reference to the combined LEO response rate across all 9 states, with this value (0.609034) divided by the response rate in each state. Thus, states in which LEO response rates were below average had weights greater than one, to bring their representation in the sample up to that state’s proportional share of all LEOs for the 9 states.
Analyses of the Combined LEO and General Population Groups
Respondent data from all nine states were used in comparing CVD risk factors between LEOs and the general population. The risk factors of hypertension, hypercholesterolemia, diabetes, physical inactivity (no activity or exercise in the past month), recent tobacco use (daily smoking within past 5 years), and elevated body mass index (BMI > 25 kg/m 2) were treated as categorical independent variables coded 1 if the risk factor was present and 0 if not. The risk factor of age was treated as a continuous variable. Subsequently, simple percentages were crosstabulated to compare the prevalence of risk factors in the LEO cohort with that found in the general population. Multicollinearity was evaluated by examining the correlation matrix for evidence of high correlations among predictor variables. Similar procedures were performed to compare CVD prevalence between the LEO group and the general population. Because CVD prevalence data from the general population were available from only 4 states, this comparison was made using the general population and the officers from just those states.
Using the data from the 4 states where CVD prevalence was known, logistic regression was employed to generate the best subset model to distinguish those with and without CVD. Initially, forward selection logistic regression was used to predict group membership with seven independent variables. Stepwise selection and backward elimination methods were employed subsequently to select the model with the best ability to predict CVD group membership.
Analyses of the LEO Group Only
The following analyses were performed to determine whether perceived stress was a significant contributor to either the prevalence of CVD or CVD risk factors among LEOs only. The effect of perceived stress on the risk factors and on CVD prevalence was estimated using multivariate analysis of variance (MANOVA). Years of experience as a LEO and perception of stress were treated as continuous independent variables; these and the aforementioned categorical variables were included in analyses of the LEO group. The within subjects variables were: CVD, diabetes, hypercholesterolemia, hypertension, physical inactivity, and elevated BMI. The effects of duration of time in the profession on CVD and on CVD risk factors, adjusting for age, also were estimated using MANOVA. The within subjects variables were: CVD, stress, diabetes, hypercholesterolemia, hypertension, physical inactivity, and elevated BMI. Age and duration of time in the law enforcement profession were entered as continuous variables.
As with the combined LEO and general population analysis, logistic regression was used to generate the best subset regression model to predict which LEOs were in the group with CVD and the group without CVD. Forward selection logistic regression was used to predict group membership using seven independent variables, followed by stepwise selection and backward elimination logistic regression, to select the best model for predicting CVD group membership.
All models were estimated by Statistical Package for the Social Sciences (SPSS) software. Statistical significance was set at P < 0.05.
In the LEO group, 2818 of 4627 (61%) officers returned surveys. In the subset of subjects used for the comparison of CVD prevalence between the officers and the general population, 1791 of a possible 2722 officers (66%) returned surveys. Respondents included LEOs from state highway patrols (n = 2470), areas of criminal investigation (n = 230), narcotics enforcement (n = 33), and other entities within their Departments of Public Safety (n = 85). LEO respondents ranged in age from 21 to 65 years, with mean of 37.3 (SD = 8.7). The mean number of years of employment in law enforcement was 13.7 (SD = 8.5; range = 1 to 40). The prevalence of CVD did not differ between LEOs and the general population [2.3 (SD = 15) versus 3.1 (SD = 17);P = 0.095]. The prevalence of CVD risk factors is reported in Table 1. The LEOs were younger [37.3 (SD = 8.7) versus 41.7 (SD = 11.7) years;P < 0.001] and had greater prevalence of hypertension, hypercholesterolemia, tobacco use, and increased BMI compared to the general population. Fewer LEOs reported having diabetes or being physically inactive.
As shown in Table 2, several risk factors were associated with CVD in the group of LEOs and citizens from the four states. Being a LEO per se was not associated with an increased risk for CVD (OR = 1.03, 95% CI = 0.59–1.78). Physical inactivity, hypercholesterolemia, hypertension, and age were directly associated with increases in CVD. These findings need to be considered in context. To account for interdependence among CVD risk factors in the multiple logistic regression model, nonmissing data must be present for all subjects. Consequently, a subject who failed to respond to one risk factor question but otherwise completed the entire survey would be excluded from such analyses. Complete data were available for only 3,012 respondents, of which 1648 were LEOs. Tobacco use was excluded from this model because the very low response rate to such questions in the general population (Table 1) would have greatly exacerbated the reduction in sample size. Thus, these findings are based on 69% of the eligible pool of respondents (3012 of 4366 total respondents), with the effects of tobacco use not included.
To mitigate the effect of reduced sample size due to missing data, CVD prevalence also was explored using bivariate cross-tabulated odds ratios. By evaluating each risk factor independently of the others, we were able to include all subjects who provided responses about a given risk factor even if they did not provide data about other risk factors. With this technique, diabetes was also found to be a significant independent contributor to CVD (Table 3).
In evaluating only the law enforcement officers, significant contributors to CVD were duration of time in the profession, hypertension, and perceived stress (Table 4). These observations were made on the 91% of the LEOs from all nine states that provided complete data (2572 of 2818 respondents). In the LEO group (Table 4), the Wald statistics indicated that stress, time as a LEO, and hypertension significantly contributed to CVD. However, the OR for these variables indicated little change in the likelihood of CVD development.
In assessing the relationship among stress, CVD morbidity, and CVD risk factors in the LEO cohort, CVD was directly associated with stress (η2 = 0.003;P = 0.008). Moreover, several CVD risk factors were affected significantly by stress: cholesterol (η2 = 0.007;P = 0.001), hypertension (η2 = 0.006;P = 0.001), and physical inactivity (η2 = 0.014;P = 0.001). Multivariate tests from the model indicated an overall significant effect of stress on the combination of the seven CVD variables.
When testing whether CVD and CVD risk factors were affected by duration of time in the profession of law enforcement, while adjusting for age, it was found that CVD was affected significantly by age only (P = 0.022). However, stress was affected significantly by duration of time on the job (P = 0.004). Cholesterol, hypertension, and diabetes were affected significantly by both age and time in the profession of law enforcement. Elevated body mass index and physical inactivity were affected significantly only by age. Tobacco use was not affected by either age or time in the profession.
This study examined the relationships among CVD morbidity, risk factors, and the perception of stress among male LEOs compared to men in the general population. The findings from this study are important because the unique characteristics of law enforcement may increase an officer’s risk of CVD exclusive of conventional risk factors, and because CVD morbidity may be a more useful endpoint than the much more commonly assessed outcome of mortality. We hypothesized that perceived stress might be a factor in the prevalence of CVD in the LEO cohort. In comparing LEOs to the general population, the main findings of this study were : (1) a similar prevalence of CVD among employed LEOs to that found in the general male population, but (2) differences in the prevalence of major risk factors for CVD. Among LEOs there is a higher prevalence of hypertension, hypercholesterolemia, elevated BMI, and tobacco use, but lower prevalence of diabetes and physical inactivity. In evaluating the LEOs only, we found: (1) an association between greater perceived stress and CVD prevalence; (2) duration of time in the profession, which was not a result of aging alone, contributed to CVD and perceived stress levels; and (3) an association between perceived stress and some CVD risk factors.
The law enforcement profession per se appeared to be unassociated with CVD morbidity in comparing employed LEOs with the general adult male population. Diagnosed CVD is considered to be job-related in many states by the respective Departments of Public Safety. Therefore, most officers who develop CVD are quite likely to retire due to this medical disability. This “healthy worker effort”33 limited the number of officers who completed the survey to those who were working and therefore less likely to have diagnosed CVD. In this context, one could argue that the prevalence of CVD in the LEO cohort actually may have been higher than expected, and the association between stress and CVD in this group actually may be more pronounced than our findings suggest.
The present study supports the conclusion of Sparrow, 1 who, when comparing CVD mortality and morbidity among 220 employed LEOs and 1428 non-LEOs, concluded that the risk of CVD in law enforcement was not significantly different from that found in the general population (OR = 1.4, 95% CI = 0.7–2.4). However, this finding of diminished risk in the LEO cohort is contrary to the findings of Franke and colleagues, 2 who compared 232 retired LEOs to 817 similarly aged males. Franke and colleagues found that LEO employment was associated with increased CVD morbidity (OR = 2.34; 95% CI = 1.53–3.58), and that the relationship persisted after considering conventional risk factors. In Franke’s study, 89 of the LEOs (32%) believed that employment in law enforcement increased their risk for CVD; the most common risk factor identified by the LEOs as contributing to CVD risk was stress (81%). An important difference between these studies was the age of the officers. Franke’s 2 study evaluated retirees [65.5 (SD = 7.1) years] while the cohorts examined in the present study and by Sparrow were active, younger officers [40.7 (SD = 8.1) years]. Franke’s findings may be due to the older cohort having been employed for a longer time and thereby experiencing prolonged exposure to the deleterious effects of law enforcement, as well as CVD manifesting itself more readily in an older population.
Several well-recognized risk factors were associated with CVD prevalence in our combined analysis of LEOs and the general population. Other researchers also have suggested that increased presence of CVD risk factors may account for increased prevalence of CVD in the law enforcement profession. 12–22 In other words, conventional CVD risk factors contribute to CVD in LEOs just as readily as in the general population. The present study suggests that the former may be at greater risk from the effects of hypertension, hypercholesterolemia, tobacco use, and obesity due to their greater prevalence among LEOs. However, the factors associated with an increased prevalence of CVD among the LEOs were duration of employment and perceived stress, while hypertension was inversely associated with CVD (Table 4). In other words, none of the conventional risk factors for CVD contributed to this model, and hypertension did so inversely. It is difficult to reconcile the disparity among these findings.
Some risk factors increase in severity among LEOs in a time-dependent manner. 14,17,21 The severity of CVD risk factors contributes importantly to the rapidity with which CVD manifests itself; we could not assess the severity, but only the presence, of these risk factors. Likewise, we did not determine whether the officers currently were being treated for any of the CVD risk factors, such as hypertension. Consequently, we hypothesize that, for these relatively young officers, stress and duration of employment were more profound contributors to CVD development than were conventional risk factors because the former may be less likely to be treated. On the other hand, hypertension typically is treated medically once it is identified. Therefore, we hypothesize that hypertension may be inversely associated with CVD among these officers simply because it was being treated.
In the LEO cohort, perceived stress clearly was associated with CVD. Stress can affect CVD prevalence both directly, and indirectly by exacerbating risk factors that contribute to the development of CVD. 28 The indirect influence interpretation is supported here, since greater levels of perceived stress were associated with several risk factors (hypercholesterolemia, hypertension, and physical inactivity) in the LEOs. This finding is consistent with the observation that, in other law enforcement groups, increased stress may result in excessive ingestion of food, including foods with a higher fat content, increased use of tobacco products, and decreased participation in physical activity. 11
Our use of the PSS does not enable us to identify the sources of stress among the LEOs. Thus, it is unclear whether the stress originates from their job, law enforcement, or elsewhere. Nevertheless, LEOs encounter sources of stress unique to their profession, such as high-speed chases and exposure to violence, and ensuing consequences including risk of death. However, these sources often are missing when officers self-report their stressors. 26,32 The more potent stressors may be more mundane. The hierarchical command structure of most public safety departments and the “bureaucracy” of law enforcement, such as excessive paperwork, lack of input into decision-making, and poor communication practices, may play important roles in inducing stress. 26 In his review of the relationship between work stress and CVD, Tennant 28 noted that jobs with either high demands and low decision-making latitude, or requiring high effort with low intrinsic rewards, are associated with increased incidence of heart disease. These scenarios typically are seen in the paramilitary command structure of most law enforcement organizations.
Our data also indicate that not all officers are greatly stressed. The range of LEO scores on the PSS was 2 to 52, with mean of 19.2 (SD = 7.3). This mean score was very similar to that of a randomly sampled cohort of 949 American males [18.8 (SD = 6.9)], 34 but lower than that found in a group of 234 Florida highway patrol officers [22.2 (SD = 8.0)]. 27 This mean score also was lower than those in a male community smoking cessation group [24.0 (SD = 7.8), n = 27] and two groups of male college students [22.4 (SD = 6.8;n = 127; and 21.7 (SD = 8.4;n = 60)]. 30 Thus, it appears that the officers assessed in the present study had relatively normal levels of perceived stress, although stress levels varied considerably. Nevertheless, officers with high levels of perceived stress appeared to be at increased risk of CVD. Parenthetically, because previous research using the PSS with highway patrol officers had a smaller sample size than that seen in the present study, it is unclear whether those results constitute normative values for PSS scores or if the results obtained from this LEO cohort are more likely to establish a standard.
Several limitations to the present study need to be recognized. All the variables assessed in the present study were self-reported. Feedback provided by the subjects’ health professionals helped form the basis for many of the self-reported responses to the questionnaire. Because different medical professionals may embrace different thresholds for high-risk criterion values, variation in these responses is inevitable. A cross-sectional study, such as the present one, does not allow for the determination of causality. Thus, our finding of an association between perceived stress and CVD among LEOs does not necessarily imply that stress per se will lead to CVD. The present study also did not determine the sources of officer stress. Caution therefore must be used in ascribing LEO stress to their profession. The LEOs studied here were a unique group—White males who were mostly troopers working in the Midwest. As such, officers in different working environments, such as municipal police or officers working in inner cities, may have markedly different stress levels than the LEOs assessed here. Similarly, the relationships among CVD, CVD risk factors, and stress may differ between that seen here and what would be obtained in groups of female or minority officers. Consequently, generalizing our findings to other groups of LEOs should be done with caution. Finally, a low response rate to some questions on the BRFSS, and consequently a large proportion of missing observations for some variables, may have led to underestimates of the prevalence of some morbidity measures in the general population, such as tobacco use in our cohort. In addition, only residents with telephones can participate in the BRFSS. The Third National Health and Nutrition Examination Survey found that people without telephones are more likely to be obese and to have diabetes. 35 Therefore, some BRFSS comparison results may be underreported.
In conclusion, this group of LEOs had higher prevalence of hypertension, hypercholesterolemia, excess body weight, and tobacco use than that seen in the general population cohort. Among LEOs, CVD was directly associated with perceived stress and duration of employment but inversely associated with hypertension. Higher levels of perceived stress were associated with increased prevalence of hypercholesterolemia, hypertension, and physical inactivity, as well as increased CVD morbidity, among LEOs. However, as a group, these officers had relatively low levels of perceived stress. The small number of self-reported CVD cases in this study and the association between time as an officer and CVD suggest that future research should include surveying retirees from the law enforcement profession.
The authors thank Sigma Theta Tau International, and the Research Institute for Studies in Education and the Department of Health and Human Performance at Iowa State University, for their financial support of this study.
1. Sparrow D, Thomas HE Jr, Weiss ST. Coronary heart disease in police officers participating in the normative aging study. Am J Epidemiol. 1983; 118: 508–513.
2. Franke WD, Collins S, Hinz PN. Cardiovascular disease morbidity in an Iowa law enforcement cohort, compared with the general population. J Occup Environ Med. 1998; 40: 441–444.
3. Calvert GM, Merling JW, Burnett CA. Ischemic heart disease mortality and occupation among 16–60 year old males. J Occup Environ Med. 1999; 41: 960–966.
4. Dubrow R, Burnett CA, Gute DM, Brockert JE. Ischemic heart disease and acute myocardial infarction mortality among police officers. J Occup Med. 1988; 30: 650–654.
5. Guralnick L. Mortality by occupation and industry among men 20–64 years of age: United States 1950. Vital Statistics-Special Reports 1962; 53, Bethesda, MD, 1962.
6. Milham Jr S. Occupational mortality in Washington State. 1950–1979, DHHS (NIOSH) Publication No. 83–116. Washington, DC, 1983.
7. Sardinas A, Wang-Miller J, Hansen H. Ischemic heart disease mortality of firemen and policemen. J Public Health. 1986; 76: 1140–1151.
8. Demers PA, Heyer NJ, Rosenstock L. Mortality among firefighters from three northwestern United States cities. Br J Ind Med. 1992; 49: 644–660.
9. Feuer E, Rosenman K. Mortality in police and firefighters in New Jersey. Am J Ind Med. 1986; 9: 517–527.
10. Forastiere F, Perucci CA, DiPietro A et al. Mortality among urban policemen in Rome. Am J Ind Med. 1994; 26: 785–798.
11. Vena JE, Violanti MJ, Marshall J, Fiedler RC. Mortality of a municipal worker cohort of police officers. Am J Ind Med. 1986; 10: 383–397.
12. Ekelund LG, Haskell WL, Johnson JL, Whaley FS, Criqui MH, Sheps DS. Physical fitness as a predictor of cardiovascular fitness in asymptomatic American men. The Lipid Research Clinics Mortality Follow-up Study N Engl J Med. 1988; 319: 1379–1384.
13. Ely DL, Mostardi RA. The effects of recent life events stress, life assets and temperament patterns on cardiovascular risk factors for Akron city police officers. J Human Stress. 1986; Summer; 12: 77–91.
14. Franke WD, Anderson DF. Relationship between physical activity and risk factors for cardiovascular disease among law enforcement officers. J Occup Med. 1994; 36: 1127–1132.
15. Franke WD, Cox, DF, Schulte DP, Anderson DF. Coronary heart disease risk factors in employees of Iowa’s department of public safety compared to a cohort of general population. Am J Ind Med. 1997; 31: 733–737.
16. Peters RK, Cady LD, Bischoff DP, Bernstein L, Pike MC. Physical fitness and subsequent myocardial infarction in healthy workers. JAMA. 1983; 249: 3052–3056.
17. Pollock ML, Gettman LR, Ullman Meyer B. Analysis of physical fitness and coronary heart disease risk of Dallas area police officers. J Occup Med. 1978; 20: 392–398.
18. Pyörälä M, Miettinen H, Laakso M, Pyörälä K. Hyperinsulinemia predicts coronary heart disease risk in healthy middle-aged men: the 22-year follow-up results of the Helsinki policemen study. Circulation. 1998; 8: 398–404.
19. Richmond RL, Wodak A, Kehoe L, Heather N. How healthy are police? A survey of life-style factors. Addiction. 1998; 93: 1729–1737.
20. Stamford BA, Weltman A, Moffatt RJ, Fulco C. Status of police officers with regard to selected cardio-pulmonary and body compositional fitness variables. Med Sci Sports Exerc. 1978; 10: 294–297.
21. Williams MA, Petrallis MA, Baechle TR, Ryschon KL, Campain JJ, Sketch MH. Frequency of physical activity exercise capacity and atherosclerotic heart disease risk factors in male police officers. J Occup Med. 1987; 29: 596–600.
22. Piercecchi M, Leonetti G, Pelissier AL. Evaluation of biological stress markers in police officers. Med and Law. 1999; 18: 125–144.
23. Wood SD, Kreitner R, Friedman GA, Edwards M, Sova MA. Cost effective wellness screening: a case study of 4534 law enforcement officers. J Police Sci. 1998; 10: 273–278.
24. Kros W. Society’s victim-the policeman. Springfield, IL: Bannerstone House, 1976.
25. Russek LG, King SH, Russek SJ, Russek HI. The Harvard mastery of stress study 35 year follow-up: prognostic significance of patterns of psychological arousal and adaptation. Psychosom Med. 1990; 52: 271–285.
26. Finn P, Tomz JE. Developing a law enforcement stress program for officers and their families. Washington, DC: National Institute of Justice, 1997.
27. Hills H, Norvell N. An examination of hardiness and neuroticism as potential moderators of stress outcomes. Behavioral Med. 1991; 17: 31–38.
28. Tennant C. Work stress and coronary heart disease. J Cardiovasc Risk. 2000; 7: 273–276.
29. Knutsson A, Akerstedt T, Jonsson BG, Orth-Gomer K. Increased risk of ischemic heart disease in shift workers. Lancet. 1986; 12: 89–92.
30. Cohen S, Kamarck T, Mermelstein RA. Global measure of perceived stress. J Health Soc Behav. 1983; 24: 385–396.
31. Monroe SM, Kelley JM. Measurement of stress appraisal. In: S Cohen, RC Kessler, L Underwood Gordon (eds.). Measuring Stress: A Guide for Health and Social Scientists. New York: Oxford University Press, 1997.
32. Malloy TE, Bays GL. The police stress hypothesis: A critical evaluation. Crim. Justice Behav. 1984; 11: 197–224.
33. Howe GR, Chiarelli AM, Lindsay JP. Components and modifiers of the healthy worker effect: evidence for three occupational cohorts and implications for industrial compensation. Am J Epidemiol. 1998; 128: 1364–1375.
34. Cohen S, Williamson GM. Perceived stress in a probability sample of the United States. In: S Spacapan, S Oskamp (eds.). The Social Psychology of Health. Beverly Hills, CA: Sage Publications, 1988.
35. Harris MI, Flegal KM, Cowie CC, et al. Prevalence of diabetes impaired fasting glucose and unimpaired glucose tolerance in U. S. adults: the third national health and nutrition examination survey. Diabetes Care. 1998; 21: 518–524.