Effort-Reward Imbalance at Work and the Prevalence of Unsuccessfully Treated Hypertension Among White-Collar Workers

Abstract

We examined the association between effort-reward imbalance (ERI) exposure at work and unsuccessfully treated hypertension among white-collar workers from a large cohort in Quebec City, Canada. The study used a repeated cross-sectional design involving 3 waves of data collection (2000–2009). The study sample was composed of 474 workers treated for hypertension, accounting for 739 observations. At each observation, ERI was measured using validated scales, and ambulatory blood pressure (BP) was measured every 15 minutes during the working day. Unsuccessfully treated hypertension was defined as daytime ambulatory BP of at least 135/85 mm Hg and was further divided into masked and sustained hypertension. Adjusted prevalence ratios and 95% confidence intervals were estimated. Participants in the highest tertile of ERI exposure had a higher prevalence of unsuccessfully treated hypertension (prevalence ratio = 1.45, 95% confidence interval: 1.16, 1.81) after adjustment for gender, age, education, family history of cardiovascular diseases, body mass index, diabetes, smoking, sedentary behaviors, and alcohol intake. The present study supports the effect of adverse psychosocial work factors from the ERI model on BP control in treated workers. Reducing these frequent exposures at work might lead to substantial benefits on BP control at the population level.

High blood pressure (BP) is a major risk factor for cardiovascular disease (CVD) (1). It accounts for an estimated 54% of all strokes and 47% of all ischemic heart disease events globally (2). Hypertension may affect more than 90% of individuals during their lifetime (3). Despite improvements in BP control over the last decades, unsuccessfully treated hypertension remains a major public health problem. According to national surveys in Canada and in the United States, 1–2 adults out of 5 receiving a pharmacologic treatment for hypertension have BP values higher than the recommended targets (4, 5).

Ambulatory blood pressure (ABP) measurement, obtained by measuring BP during normal daily activities using a wearable device, has improved our ability to assess the effect of antihypertensive treatment (6). ABP monitoring is also a valuable tool to identify barriers to adequate BP control in treated patients. Despite the fact that they are not screened in the doctor's office, work stressors could be important contributors (7). The Siegrist effort-reward imbalance (ERI) model is a well-defined and internationally recognized model used to assess work stressors (8). This model proposes that efforts should be rewarded in various ways: income, respect and esteem, and occupational status control. Workers are in a state of detrimental imbalance, and thus more susceptible to health problems, when high efforts are accompanied by low reward. Effort-reward imbalance at work has been linked to high BP (9–16) and CVD incidence (17) in prospective studies.

There is scarce evidence on the association between work stressors and unsuccessfully treated hypertension. In what was, to our knowledge, the sole previous study on this topic, job strain (a combination of high demands and low control) was not associated with poor BP control among aging workers treated for hypertension (18). This previous study was conducted among older workers (mean age of >60 years) for whom the influence of work stressors on cardiovascular health might differ (19). In addition, BP was measured using a manually operated device, which is not equivalent to ABP measurements, recognized as the gold standard method for the diagnosis of hypertension (20). In fact, hypertension diagnosed from ABP measurements can be classified either as sustained or masked according to the measurements obtained from manual readings. Sustained hypertension results from an agreement between both methods. Masked hypertension, an increasingly recognized clinical entity, is defined by normal manual BP (lower than 140/90 mm Hg) combined with elevated ABP (at least 135/85 mm Hg) (21) and is associated with an increased risk of CVD morbidity (22–24). To our knowledge, no previous study has examined the effect of work stressors on masked uncontrolled hypertension.

The aim of the study was 2-fold: 1) to examine the association between ERI exposure at work and the prevalence of unsuccessfully treated hypertension and 2) to examine whether the effect of ERI exposure differs for masked and sustained uncontrolled hypertension.

METHODS

Population and study design

The study relies on a repeated cross-sectional design. Data collection was performed at 3 time points: at baseline (2000–2004), 3 years later (2004–2006), and 5 years later (2006–2009). The study population was composed of white-collar workers from 3 public insurance institutions. Their main activities were planning and providing insurance services to the general population. Workers completed a self-administered questionnaire on work characteristics and BP risk factors in a dedicated room at the workplace. All employees of the organizations were invited to participate at each measurement period. The study sample was restricted to observations from workers who reported the use of antihypertensive medication (n = 895). Across all 3 measurement waves, we excluded 142 observations due to insufficient or unavailable BP measurements. We also excluded 5 observations for missing information on ERI exposure and 5 observations for missing information on covariates. Finally, 4 observations were excluded for insufficient working time (less than 21 hours), to prevent misclassification caused by limited exposure to adverse psychosocial work factors. The final study sample consisted of 474 workers and 739 observations. These workers participated in 1 (n = 266), 2 (n = 151), or all 3 waves of data collection (n = 57).

Effort-reward imbalance at work

Reward was measured using the 11 items recommended by Siegrist (25). These items were divided into 3 scales assessing esteem (5 items), promotions and salary (4 items), and job security (2 items). The factorial validity and internal consistency of both the original English and the French versions of this instrument have been demonstrated (25, 26). Effort was measured with 2 original items of the Siegrist questionnaire (“over the past few years, my job has become more and more demanding,” and “I am regularly forced to work overtime”) and with 2 proxies (“my tasks are often interrupted before they can be completed, requiring attention at a later time,” and “I have enough time to do my work”) (Cronbach's α = 0.69). The ERI ratio was constructed using effort as the numerator and reward as the denominator. Tertile-based variables were created for each dimension taken separately as well as for ERI exposure.

Ambulatory blood pressure

At each wave, ABP was assessed using the Spacelabs 90207 oscillometric device (Spacelabs Produits Médicaux Ltée, Saint-Laurent, Quebec, Canada), validated by the independent investigators’ protocol and recommended by the Association for the Advancement of Medical Instrumentation and British Hypertension Society (27, 28). The device was placed on the nondominant arm if BP difference measured on both arms was less than 10 mm Hg. Otherwise, it was placed on the arm showing the higher BP level. ABP was measured every 15 minutes during daytime working hours. ABP was defined as the mean of all readings taken during the working day. In the present study, participants were required to be measured at least 20 times with each use of the ABP monitor, which is in accordance with more stringent criteria recently recommended by expert committees (29). Participants were white-collar workers whose tasks were performed mainly in a sitting position. To minimize motion artifact, clear instructions were given to participants in case the monitor performed a reading while they were in a standing position (stop walking with their arm resting at their side). Unsuccessfully treated hypertension was defined as daytime systolic or diastolic ABP greater than or equal to 135 mm Hg or 85 mm Hg, respectively (30). A sensitivity analysis was conducted in which the threshold was modified to 125/85 mm Hg for participants reporting a doctor's diagnosis of diabetes (12% of study sample) (20). This analysis yielded similar results (not shown).

Unsuccessfully treated hypertension was further divided into categories of white-coat, masked, and sustained hypertension according to manually operated blood pressure (MOBP) and ABP measurements. MOBP was defined as the average of the first 3 readings taken after the participant had been sitting for 5 minutes, in the presence of trained personnel. Workers were classified into 4 hypertension subtypes, using the European Society of Hypertension guidelines (21): 1) controlled hypertension, in which MOBP was <140/90 mm Hg and ABP was <135/85 mm Hg; 2) white-coat hypertension, in which MOBP was ≥140/90 mm Hg and ABP was <135/85 mm Hg; 3) masked hypertension, in which MOBP was <140/90 mm Hg and ABP was ≥135/85 mm Hg; and 4) sustained hypertension, in which MOBP was ≥140/90 mm Hg and ABP was ≥135/85 mm Hg.

Covariates

Several factors were considered as potential confounders including cigarette smoking status, and alcohol intake. Smoking status was defined as the daily consumption of at least 1 cigarette per day. Alcohol intake was measured using the following 3 categories, related to weekly intake frequency during the previous 12 months: <1 drink per week, 1–5 drinks per week, and ≥6 drinks per week. Other factors also considered as potential confounders were age, education (less than college completion, college completed, university completed), diabetes (self-reported), and family history of cardiovascular disease. The definition of the latter group was based on a report by the participant of a cardiovascular event—such as angina, myocardial infarction, coronary revascularization, or stroke—suffered by their father, mother, brother, or sister before the age of 60 years. The risk factors listed above were evaluated using validated protocols (31, 32).

Analyses

The χ² test was used to compare participant characteristics. Generalized estimating equations were used to examine the association between exposure to psychosocial work factors and hypertension. We included all employed individuals at each measurement event in the analyses. The repeated cross-sectional design used in the present study is suited particularly well to analyses of dynamic populations, such as work organizations, because it accommodates new participants’ in-migration. Generalized estimating equations provide estimates of marginal means and take into account within-subject correlation.

Differences in mean ABP according to ERI exposure were modeled using analysis of covariance. Prevalence ratios of unsuccessfully treated hypertension and 95% confidence intervals were modeled using robust Poisson regression (33). It should be noted that lifestyle risk factors might act as mediating variables (i.e., intervene in the causal pathway linking work stress to BP) (34). We have therefore examined 2 sets of adjusted models: 1) adjusted for sociodemographic and cardiovascular risk factors (age, education, diabetes, and family history of CVD) and 2) additionally adjusted for lifestyle-related risk factors (smoking, body mass index, physical activity, and alcohol intake). Because they yielded similar estimates, only the latter are presented. All analyses were performed with SAS, version 9.4 (SAS Institute, Inc., Cary, North Carolina) (35). This study was approved by the ethical review board of the Centre Hospitalier Universitaire de Québec.

RESULTS

Table 1 presents the description of the study population. There was the same proportion of men and women. Mean age was 50.8 (standard deviation, 6.5) years. The prevalence of uncontrolled hypertension was 44.3%. Table 2 shows mean BP and hypertension subtypes, defined using the cross-classification of MOBP and ABP. Mean BP was 131.0/80.8 mm Hg for MOBP and 129.7/81.8 mm Hg for ABP. The prevalences of masked and sustained uncontrolled hypertension were comparable (21.2% vs. 23.0%), while white-coat uncontrolled hypertension prevalence was significantly lower (4.8%).

Table 1.

Description of the Study Population, Quebec City, Canada, 2000–2009

Characteristic	No. of Observations (n = 739)	%
Sex
Men	364	49.3
Women	375	50.7
Mean age, yearsa	50.8 (6.5)
Mean body mass indexa,b	28.9 (5.4)
Sedentary behaviors	205	27.7
Education
Less than college completion	230	31.1
College completed	207	28.0
University completed	302	40.9
Family history of CVD	340	46.0
Alcohol consumption, drinks/week
<1	239	32.3
1–5	265	35.9
≥6	235	31.8
Smoking status	99	13.4
Unsuccessfully treated hypertensionc	327	44.3

Abbreviation: CVD, cardiovascular disease.

^a Data expressed as mean values (standard deviation).

^b BMI was calculated as weight (kg)/height (m)².

^c Prevalence of unsuccessfully treated hypertension according to ambulatory blood pressure measurements (defined as ≥135/85 mm Hg).

Table 2.

Mean Blood Pressure and Hypertension Subtypes, Quebec City, Canada, 2000–2009

Blood Pressure and Hypertension Subtypes	Mean (SD)	No. of Observations	%
Mean blood pressure
Manually operated blood   pressure, mm Hg
Systolic	131.0 (13.2)
Diastolic	80.8 (9.4)
Ambulatory blood pressure,   mm Hg
Systolic	129.7 (11.6)
Diastolic	81.8 (8.5)
Hypertension subtypes
Controlled hypertension		377	51.0
White-coat uncontrolled  hypertension		35	4.8
Masked uncontrolled  hypertension		157	21.2
Sustained uncontrolled  hypertension		170	23.0

Abbreviation: SD, standard deviation.

As shown in Figure 1, participants exposed to ERI had higher systolic (by 2.9 mm Hg, 95% confidence interval (CI): 0.8, 4.9) and diastolic (by 2.9 mm Hg, 95% CI: 1.4, 4.4) BP mean values compared with participants unexposed to ERI. Table 3 presents the association between ERI at work and unsuccessfully treated hypertension. The prevalence was higher in participants in the highest tertile of ERI exposure (prevalence ratio (PR) = 1.45, 95% CI: 1.16, 1.81). The associations between ERI exposure and unsuccessfully treated hypertension were robust to adjustment for demographic and lifestyle-related risk factors. When taken separately, the effort and reward components showed significant (for effort, PR = 1.37, 95% CI: 1.10, 1.71) or borderline significant (for reward, PR = 1.21, 95% CI: 0.97, 1.51) associations. Table 4 presents the association between ERI exposure and uncontrolled hypertension subtypes. Being in the highest tertile of ERI exposure was associated with masked (PR = 1.49, 95% CI: 1.06, 2.10) and sustained (PR = 1.67, 95% CI: 1.18, 2.37) uncontrolled hypertension.

Figure 1.

Differences in mean ambulatory blood pressure (BP) in mm Hg according to effort-reward imbalance exposure among workers treated for hypertension (n = 739 observations), Quebec City, Canada, 2000–2009. A) Systolic BP: first tertile (unexposed workers), mean = 128.5 mm Hg; second tertile, mean difference (MD) = 0.8 mm Hg; third tertile, MD = 2.9 mm Hg (P < 0.05). B) Diastolic BP: first tertile (unexposed workers) mean = 80.5 mm Hg; second tertile, MD = 1.2 mm Hg; third tertile, MD = 2.9 mm Hg (P < 0.05). The models adjusted for sex, age, education, family history of cardiovascular diseases, body mass index, diabetes, smoking, sedentary behaviors, and alcohol intake.

Table 3.

Prevalence of Unsuccessfully Treated Hypertension^a Diagnosed From Ambulatory Blood Pressure According to Effort-Reward Imbalance at Work, Quebec City, Canada, 2000–2009

Work Stressors	No. of Observations (n = 739)	Adjusted Prevalence, %b	PRb	95% CI
ERI exposure
1st tertile	236	33.2	1.00	Referent
2nd tertile	258	44.3	1.33	1.09, 1.63
3rd tertile	245	48.1	1.45	1.16, 1.81
High Effort
1st tertile	331	36.1	1.00	Referent
2nd tertile	253	45.3	1.25	1.04, 1.52
3rd tertile	155	49.5	1.37	1.10, 1.71
Low reward
1st tertile	192	36.5	1.00	Referent
2nd tertile	251	43.9	1.20	0.96, 1.51
3rd tertile	296	44.1	1.21	0.97, 1.51

Abbreviations: CI, confidence interval; ERI, effort-reward imbalance; PR, prevalence ratio.

^a Uncontrolled hypertension was defined as ambulatory blood pressure of ≥135/85 mm Hg.

^b The models adjusted for sex, age, education, family history of cardiovascular diseases, diabetes, body mass index, smoking, sedentary behaviors, and alcohol intake.

Table 4.

Hypertension Subtypes According to Effort-Reward Imbalance at Work, Quebec City, Canada, 2000–2009

ERI Exposure Tertile	No. of Observations (n = 739)	White-Coat Uncontrolled Hypertension		Masked Uncontrolled Hypertension		Sustained Uncontrolled Hypertension
		PRa	95% CI	PRa	95% CI	PRa	95% CI
1st tertile	236	1.00	Referent	1.00	Referent	1.00	Referent
2nd tertile	258	1.04	0.50, 2.19	1.37	0.98, 1.93	1.50	1.09, 2.07
3rd tertile	245	0.70	0.28, 1.75	1.49	1.06, 2.10	1.67	1.18, 2.37

Abbreviations: CI, confidence interval; ERI, effort-reward imbalance; PR, prevalence ratio.

^a The models adjusted for sex, age, education, family history of cardiovascular diseases, diabetes, body mass index, smoking, sedentary behaviors, and alcohol intake.

DISCUSSION

In the present study, workers exposed to ERI at work were more likely to have unsuccessfully treated hypertension when compared with unexposed workers. This association was robust to adjustment for demographic, lifestyle-related, and cardiovascular risk factors and was observed for both masked and sustained hypertension. Our results suggest that psychosocial factors from the work environment could contribute to the poor BP control observed in national surveys, which is of high public health significance. The proportions of working men and women exposed to these adverse factors at work have been shown to be about 20%–25% in previous prospective studies conducted in industrialized countries (36). Our study suggests that reducing these exposures could result in substantial benefits in BP control.

Previous evidence for an effect of adverse psychosocial work factors on hypertension control is scarce. In what is, to our knowledge, the sole previous study on this topic, job strain, as defined by high psychological demands and low job control, was not associated with unsuccessful treatment in a nationally representative sample of older adult in the United States (18). In this previous study, BP was measured on a single occasion with a manually operated device. As supported by our findings, the use of such measurement methodology does not fully capture the effect of work stress exposure on uncontrolled hypertension. In a supplementary analysis, we examined the association between ERI exposure and hypertension, diagnosed exclusively from MOBP. This analysis showed that both the prevalence of unsuccessful treatment and its association with ERI exposure are underestimated using this manual method (Web Table 1, available at https://academic.oup.com/aje).

To our knowledge, our study is the first to address the association between work stressors and masked uncontrolled hypertension. In our population, the prevalence of masked uncontrolled hypertension was around 20%, which is within the range of previous reports (37, 38). Our findings showed a significant deleterious association of ERI exposure with masked uncontrolled hypertension, which was only slightly lower than that observed for sustained uncontrolled hypertension. Our results therefore support the need for future national surveys to include ABP measurement in order to give a more valid estimation of the magnitude of the public health burden of unsuccessfully treated hypertension and to examine its related risk factors.

Previous studies have reported a deleterious effect of ERI exposure on high BP (9–16). Direct physiological mechanisms, involving activation of the sympathetic and parasympathetic nervous system, have been suggested to explain the effect of psychosocial stressors on CVD risk (39–41). Indirect pathways, through behavioral and lifestyle modifications, have also been put forward (42–44). Nonadherence to prescription therapy has been identified as a leading contributor to failure to control BP and cholesterol levels (45). Psychosocial risk factors at work were found to be associated with nonadherence to health recommendation among older working adults with CVD (46). It is therefore reasonable to presume that the effect of adverse psychosocial stressors at work on unsuccessfully treated hypertension could be explained partly by a negative influence on adherence behaviors to treatment recommendations. More studies are needed to examine this hypothesis.

Our study has some limitations. First, information on the type and dose of antihypertensive treatment and on adherence was not available. Poor adherence might be an important pathway explaining the observed effect of ERI exposure on BP control. Examining this pathway was beyond the scope of the present study. Second, participants were systematically informed about their hypertensive status between measurements and referred to their physician when their ABP values were compatible with the diagnosis of hypertension. The effect of ERI exposure might have been underestimated through more intensive treatment over the study period among screened participants. In a post hoc analysis, we have documented this potential underestimation. The stability of the effect of ERI exposure was examined by entering a term for interaction between ERI and time and by stratifying our results by measurement event. Although the P value for the interaction term suggests a constant effect over time (P = 0.72), stratified analyses (Web Table 2) showed a slight attenuation at follow-up, which could be explained partly by our screening procedure. Third, our modeling strategy did not enable us to investigate duration of ERI exposure and its effect on uncontrolled hypertension. However, this repeated cross-sectional approach was optimal to investigate the proximal effect of work stressors exposure. Finally, the study population was composed of white-collar workers, and so results may not generalize to other populations. However, this choice had the advantage of minimizing physical work constraints that may influence BP at work. The effect of adverse psychosocial work factors on unsuccessfully treated hypertension in other populations, such as in individuals of lower socioeconomic status or living in developing countries, should be examined in future studies.

Our study also has important strengths. It is one of the few that has examined the association of work stressors with hypertension control in treated individuals with hypertension and, to our knowledge, the first to examine on the association with masked and sustained uncontrolled hypertension. This was done using sound methods. The sample of treated hypertensives was drawn from a large population of white-collar workers. Psychosocial work stressors from the ERI model were assessed with validated scales. Finally, potential confounders were taken into account, supporting the robustness of the observed association between work stress exposure and unsuccessfully treated hypertension.

The prevalence of unsuccessfully treated hypertension was high in this white-collar workers population. Our results suggest that work stressors from the ERI model have a deleterious effect on BP control and could undermine the success of pharmacological antihypertensive treatment. Reducing these adverse and frequent exposures at work should be considered in efforts to improve BP control.