Comparing the risk associated with psychosocial work conditions and health behaviours on incident hypertension over a nine-year period in Ontario, Canada.
Smith, Peter M. ; Mustard, Cameron A. ; Lu, Hong 等
The prevention and management of hypertension is a longstanding
public health concern in developed countries. In Ontario, Canada's
largest province, rates of hypertension increased from 153.1 to 244.8
per 1000 Ontarians between 1995 and 2005; a relative increase of 60%.
(1) This increased hypertension prevalence is thought to be attributed
to increasing rates of obesity and sedentary lifestyles, as well as
improved survival among the affected population. (1,2) From a public
health perspective, it is important to understand the relative impact of
various modifiable risk factors on hypertension incidence. While many
public health efforts have focused on changes to health behaviours,
relatively little research has focused on the relative contribution of
the psychosocial work environment to hypertension risk (3) --this
despite evidence that aspects of the psychosocial work environment are
associated with elevated hypertension risk, (4-6) and that aspects of
the work environment can be modified. (7,8)
In particular, there have been relatively few Canadian studies
examining the relationship between the psychosocial work environment and
hypertension. (9,10) The most recent Canadian study we are aware of
examined the relationship between cumulative exposure to job strain
(where job control is low and psychological demands are high) and
measured blood pressure over a 7.5-year period among employees of public
organizations in Quebec City. (11) In this study, exposure to job strain
at time two, and exposure at both time one and time two, were associated
with elevated systolic blood pressure readings among men, but not women,
compared to respondents with no exposure to job strain at either time
point. We are not aware of any Canadian studies that have examined the
temporal relationship between the psychosocial work environment and
subsequent risk of incident hypertension, among a population free of
hypertension when work stress was assessed.
The objectives of this paper are to address this research gap by
examining the relationships between the psychosocial work environment
and subsequent hypertension over a 9-year period in Ontario, Canada; and
to compare the risks associated with the psycho social work environment
to those obtained from other health behaviours (smoking, alcohol
consumption, physical activity, and fruit and vegetable consumption).
METHODS
This study used secondary data from Ontario respondents to the
2000-01 Canadian Community Health Survey (CCHS) linked to the Ontario
Health Insurance Plan (OHIP) database covering physician services as
well as the Canadian Institute for Health Information Discharge Abstract
Database (CIHI-DAD) for hospital admissions, at the individual level.
Follow-up information from the OHIP and CIHI-DAD databases was available
up to March 31, 2010. For the purpose of this analysis, we focused on
labour market participants aged 35 to 60 who: had not been previously
diagnosed with hypertension (using both administrative health records
and self-reported information from the CCHS); were not self-employed;
and were working more than 10 hours per week, more than 20 weeks in the
previous 12 months (N = 7,171).
Main outcome
Incidence of Hypertension
Incidence of hypertension was classified if respondents had one
hospital admission with a hypertension diagnosis, or two physician
service claims with a hypertension diagnosis within a two-year period
(ICD9 codes 401, 402, 403, 404 or 405; ICD10 codes I10, I12, I13, or
I15). (1,12) Previous work at the Institute for Clinical Evaluative
Sciences has demonstrated that this classification of hypertension using
medical records has a sensitivity of 73% and specificity of 95%, when
compared to physician chart-recorded hypertension diagnoses. (12)
Main independent variables
Psychosocial Work Environment
Psychosocial working conditions included job control, psychological
demands and social support, assessed by an abbreviated measure of the
Job Content Questionnaire (JCQ). (13) Job control and psychological
demand measures were also combined to form a measure of job strain
(where psychological demands are high and job control is low).
Health Behaviours
We had information on four health behaviours that may be associated
with an elevated risk of hypertension. These were: leisuretime physical
activity (inactive; moderately active; active); smoking status (never;
former; occasional; 1-10 cigarettes per day; >10 cigarettes per day);
alcohol consumption (non-drinker; non-binge drinker; binge drinker less
than once per month; binge drinker once a month or more); and daily
fruit and vegetable consumption ([greater than or equal to] 5 servings a
day; <5 servings a day).
Potential confounders
Models were adjusted for a range of variables that may confound the
relationship among working conditions, health behaviours and
hypertension. These included: the respondent's shift schedule
(regular; evening or night; rotating; other); the physical activity of
their current occupation (usually sits; stands or walks about quite a
lot, but does not carry or lift things; usually lifts or carries light
loads, or has to climb stairs or hills often, or does heavy work or
carries very heavy loads); the number of hours usually worked per week;
if they worked less than 40 weeks in the previous 12 months (yes/no); if
they were working multiple jobs (yes/no); ethnicity; immigration status
and length of time in Canada; age (grouped); marital status; body mass
index (grouped); living location (urban or rural); highest level of
education completed; the presence of heart disease or diabetes at
baseline (yes/no); and if the respondent had a long-term mental or
physical health problem that limited the amount or the kind of activity
they could do at work (yes/no).
Statistical methods
Our original sample of labour market participants aged 35-60 years,
free of hypertension at baseline, totalled 7,171 respondents. Of this
sample, 560 (7.8%) were missing information on ethnicity, length of time
in Canada, sex, education level, information on working conditions or
other covariates of interest, leaving a final sample of 6,611
respondents, i.e., 92.2% of the original sample. Older respondents,
those who were female, and those with less than secondary education were
more likely to be missing responses on work variables. Females were also
more likely to be missing information on other covariates.
Cox-proportional hazard regression models examined the relationship
between psychosocial work conditions and health behaviours and the
probability of hypertension diagnosis over the nine-ear follow-up
period. To account for the complex sample design of the CCHS, confidence
intervals have been adjusted using a bootstrap technique. (14) In
addition, all analyses were weighted to account for the probability of
selection into the original sample and non-response. All analyses were
conducted using SAS 9.2. (15) We also estimated the population
attributable fraction of hypertension that would be reduced if specific
modifiable risk factors were eliminated. This was done using the general
formula: PAF = pd x [([HR.sub.adj] -1)/[HR.sub.adj]; where PAF =
population attributable fraction; pd = proportion of the population with
hypertension in each category; and [HR.sub.adj] is the hazard ratio from
the fully adjusted model. (16) Due to previously noted differences in
the impact of psychosocial work conditions on hypertension among men and
women, all models were stratified by gender. (11,17)
RESULTS
Table 1 presents descriptive information for hypertension incidence
across our main independent variables. Over our study period, we had
53,573 person-years of follow-up (median follow-up 8.85 years); 19.4% of
our study population developed hypertension, with a higher incidence
among men (20.9%) than women (17.9%). Focusing on our main independent
variables, higher incidence of hypertension was apparent across
respondents with lower levels of job control, but these differences were
far greater among men than women.
Table 2 presents the results of our regression models. Model one
presents the hazard ratios for measures of the psychosocial work
environment and health behaviours after adjustment for all confounders.
Among men, a statistically significant elevated risk of hypertension was
observed for respondents with the lowest levels of job control compared
to those with the highest. Only irregular binge drinking (less than once
per month) was associated with elevated hypertension risk among women,
with no statistically significant relationships between health
behaviours and hypertension observed among men. We also ran models with
a measure of job strain (high psychological demands and low job control
using the median split method) in place of psychological demands and job
control. No relationship was observed between high job strain and
incident hypertension among men or women (results available on request).
Table 3 presents the category-specific population attributable
fraction (PAF) for psychosocial work conditions and health behaviours
from our fully adjusted models. Among men, the PAF associated with low
job control was 11.8%. Among women, a comparable PAF was associated with
occasional binge drinking (11.5%). The highest PAF associated with a
modifiable variable among both men and women was obesity, which had a
PAF of 26.0% among men and 18.2% among women (results not shown but
available on request).
DISCUSSION
The primary objective of this paper was to examine the
relationships between the psychosocial work environment and hyper
tension among employed respondents in Ontario over a 9-year period, and
to compare the risks associated with the psychosocial work environment
to those obtained from health behaviours (smoking, alcohol consumption,
physical activity, and fruit and vegetable consumption). We found that
low job control was associated with an increased risk of hypertension
among men, but not among women. We also found that occasional binge
drinking was associated with an increased risk of hypertension among
women. These findings suggest that job control is also an important
modifiable risk factor for hypertension in Ontario among male labour
market participants.
The results of this study support previous research in Canada
demonstrating that the psychosocial work environment is an important
determinant of hypertension risk among men, but not among women. (11)
However, unlike this previous study, we found only a relationship
between job control and hypertension, and not job strain (the
combination of low job control and high psychological demands). Further
comparisons of our study with this previous study are hampered by the
differing classification of outcomes between these two studies, with
measured blood pressure used in the previous study, while we relied upon
administrative records. However, our finding that low job control is
more important than psychological demands and social support in
determining cardiovascular risk is congruent with results from studies
on British Civil Servants. (18) Low job control was the only
work-related variable in our fully adjusted model associated with an
increased risk of hypertension (albeit only among men). Unlike the
previous Canadian studies examining the impact of sitting at work and
cardiovascular risk, (19,20) we did not find a relationship between
self-reported physical demands of work and incident hypertension.
The results of this study, however, should be interpreted within
the context of the following limitations. Our study relied upon
administrative health care records to determine incident hypertension
over our follow-up period. As such, there may be a proportion of our
sample that are hypertensive, but remain undiagnosed. We did examine the
last contact with the health care system across respondents in our
sample, and found that 99% had contact with the health care system
during our follow-up period, with 85% having their last contact in our
last three years of follow-up. There may be more barriers to health care
utilization across lower socio-economic participants in our sample.
Given that low job control is associated with lower socio-economic
position, (21) the hazards associated with low job control in our sample
may be biased to the null. While this is a concern in any study using
health care data, Ontario has universal health coverage, with studies
finding socio-economic differences in health care in Canada being
predominantly associated with specialist care, but not general
practitioner care. (22,23) We did not have information on the length of
time respondents were exposed to each working condition, or subsequent
changes in health behaviours, which may lead to misclassification,
biasing the results reported here to the null.24 The exclusion of
participants with hypertension from our baseline sample may have
produced a population who are less susceptible to certain hypertension
determinants (e.g., health behaviours) relative to others (e.g., low job
control). This may be one reason for the lack of effect of health
behaviours on hypertension risk in this sample. However, the exclusion
of hypertensive respondents was required in order to ensure the correct
temporality between our main independent variables and the study
outcome. Our study also has a number of strengths, including a large
representative data source containing information on a wide variety of
information on working conditions, health behaviours and other
sociodemographic variables, allowing for one of the first longitudinal
examinations of the relationships between the psychosocial work
environment and hypertension in Canada.
Our findings underpin the importance of psychosocial working
conditions--in particular, control at work--in understanding the social
patterning of diseases such as hypertension. We did not observe a strong
relationship between health behaviours and hypertension in our sample,
with the exception of non-regular binge drinking among women. However,
while primary prevention programs in Canada and elsewhere often target
health behaviours as determining hypertension risk, very little
attention is given to the impact that the working environment may have
on the development of hypertension. In this study, the PAFs for job
control were second only to obesity among men. As such, the inclusion of
aspects of work--in particular, job control among men--should be
considered along with health behaviours as part of a comprehensive
primary prevention strategy for hypertension in Canada.
Received: July 10, 2012 Accepted: November 22, 2012
Conflict of Interest: None to declare.
REFERENCES
(1.) Tu K, Chen ZL, Lipscombe LL. Prevalence and incidence of
hypertension from 1995 to 2005: A population-based study. CMAJ
2008;178:1429-35.
(2.) Katzmarzyk PT, Mason C. Prevalence of class I, II and III
obesity in Canada. CMAJ 2006;174:156-57.
(3.) Smith P, Frank J, Mustard C. The monitoring and surveillance
of the psychosocial work environment in Canada: A forgotten determinant
of health. Can J Public Health 2008;99:475-77.
(4.) Steptoe A, Cropley M, Joekes K. Job strain, blood pressure and
response to uncontrollable stress. J Hypertension 1999;17:193-200.
(5.) Steptoe A, Siegrist J, Kirschbaum C, Marmot M. Effort-reward
imbalance, over-commitment, and measures of cortisol and blood pressure
over the working day. Psychosomatic Med 2004;66:323-29.
(6.) Hemingway H, Marmot MG. Psychosocial factors in the aetiology
and prognosis of coronary heart disease: Systematic review of
prospective cohort studies. BMJ 1999;318:1460-67.
(7.) Bourbonnais R, Brisson C, Vinet A, Vezina M, Abdous B, Gaudet
M. Effectiveness of a participative intervention on psychosocial work
factors to prevent mental health problems in a hospital setting. Occup
Environ Med 2006;63:335-42.
(8.) Gilbert-Ouimet M, Brisson C, Vezina M, Trudel L, Bourbonnais
R, Masse B, et al. An intervention study on psychosocial work factors
and mental health and musculoskeletal outcomes. Healthcare Papers
2011;11(Sp):47-66.
(9.) Brisson C, Laflamme N, Moisan J, Milot A, Masse B, Vezina M.
Effect of family responsibilities and job strain on ambulatory blood
pressure among white-collar women. Psychosom Med 1999;61:205-13.
(10.) Laflamme N, Brisson C, Moisan J, Milot A, Masse B, Vezina M.
Job strain and ambulatory blood pressure among female white-collar
workers. Scand J Work Environ Health 1998;24:334-43.
(11.) Guimont C, Brisson C, Dagenais GR, Milot A, Vezina M, Masse
B, 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.
(12.) Tu K, Campbell NRC, Chen ZL, Cauch-Dudek J, McAlister FA.
Accuracy of administrative databases in identifying patients with
hypertension. Open Med 2007;1:E18-E26.
(13.) Karasek R, Theorell T. Healthy Work: Stress Productivity and
the Reconstruction of Working Life. New York, NY: Basic Books Inc.,
1990.
(14.) Yeo D, Mantel H, Liu TP. Bootstrap variance estimation for
the National Population Health Survey. American Statistical Association
Conference, Baltimore, MD, 1999;778-83.
(15.) The SAS Institute. The SAS System for Windows, Release 9.2.
2010.
(16.) Hennekens CH, Buring JE. Measures of disease frequency and
association. In: Mayrent SL (Ed.), Epidemiology in Medicine. Toronto,
ON: Little, Brown and Company, 1987;54-98.
(17.) Belkic KL, Landsbergis PA, Schnall PL, Baker D. Is job strain
a major source of cardiovascular disease risk? Scand J Work Environ
Health 2004;30:85-128.
(18.) Bosma H, Marmot MG, Hemingway H, Nicholson AC, Brunner E,
Stansfeld SA. Low job control and risk of coronary heart disease in
Whitehall II (prospective cohort) study. BMJ1997;314:558-65.
(19.) Probert AW, Tremblay MS, Gorber SC. Desk potatoes: The
importance of occupational physical activity on health. Can J Public
Health 2008;99:311-18.
(20.) Katzmarzyk PT, Church TS, Craig CL, Bouchard C. Sitting time
and mortality from all causes, cardiovascular disease, and cancer. Med
Sci Sports Exerc 2009;41:998-1005.
(21.) Smith PM, Frank JW, Mustard CA, Bondy S. Examining the
relationships between job control and health status: A path analysis
approach. J Epidemiol Community Health 2008;62:54-61.
(22.) Roos NP, Mustard CA. Variation in health and health care use
by socioeconomic status in Winnipeg, Canada: Does the system work well?
Yes and no. Milbank Q 1997;75:89-111.
(23.) Glazier RH, Agha MM, Moineddin R, Sibley LM. Universal health
insurance and equity in primary care and specialist office visits: A
population-based study. Ann Fam Med 2009;7:396-405.
(24.) Stringhini S, Sabia S, Shipley M, Brunner E, Nabi H, Kivimaki
M, Singh-Manoux A. Association of socioeconomic position with health
behaviors and mortality. JAMA 2010;303:1159-66.
Peter M. Smith, PhD, MPH, [1-3] Cameron A. Mustard, ScD, [1,2] Hong
Lu, PhD, [4] Richard H. Glazier, MD, MPH2, [4-6]
Author Affiliations
[1.] Institute for Work & Health, Toronto, ON
[2.] Dalla Lana School of Public Health, University of Toronto,
Toronto, ON
[3.] School of Public Health and Preventive Medicine, Monash
University, Victoria, Australia
[4.] Institute for Clinical Evaluative Sciences, Toronto, ON
[5.] Centre for Research on Inner City Health, St. Michael's
Hospital, Toronto, ON
[6.] Department of Family and Community Medicine, St.
Michael's Hospital and University of Toronto, Toronto, ON
Correspondence: Peter Smith, Associate Professor, School of Public
Health and Preventive Medicine, Monash University, Victoria, Australia,
Tel: +613.9903.0283, E-mail: peter.smith@monash.edu
Sources of Support: This work was supported by a grant from the
Canadian Institutes of Health Research (#201246). Peter Smith was
supported by a New Investigator Award from the Canadian Institutes of
Health Research while undertaking this work, and is currently supported
by a Discovery Early Career Research Award from the Australian Research
Council. Approval for the secondary data analyses was obtained through
the University of Toronto, Health Sciences I Ethics committee. This
study was supported by the Institute for Clinical Evaluative Sciences
(ICES), which is funded by an annual grant from the Ontario Ministry of
Health and Long-Term Care (MOHLTC). The opinions, results and
conclusions reported in this paper are those of the authors and are
independent from the funding sources. No endorsement by ICES or the
Ontario MOHLTC is intended or should be inferred.
Table 1. Frequencies of Main Independent Variables and
Hypertension Incidence During a 9-year Follow-up,
Stratified by Gender
Males
Hyper-
tension
N Incidence
3217 20.9%
Working Conditions
Nature of Work
Job control
1stquartile (high) 1089 18.3%
2nd quartile 884 20.8%
3rd quartile 604 19.2%
4th quartile (low) 639 27.0%
Psychosocial demands
1st quartile (high) 484 22.2%
2nd quartile 1125 19.5%
3rd quartile 877 23.2%
4th quartile (low) 730 19.4%
Social support
1st quartile (high) 502 21.1%
2nd quartile 1051 19.1%
3rd quartile 1103 21.4%
4th quartile (low) 560 23.0%
Health Behaviours
Smoking
Never 932 20.4%
Former 1305 22.9%
Occasional 150 23.9%
1-10 per day 165 12.7%
>10 per day 664 19.0%
Physical activity
Active 656 19.9%
Moderately active 789 20.7%
Inactive 1773 21.3%
Alcohol consumption
Non-drinker 420 20.4%
Non-binge drinker 1321 20.1%
Binge drinker less than once 706 21.8%
per month
Binge drinker once a month 770 21.7%
or more
Fruit and vegetable consumption
[greater than or equal to]5 964 21.1%
servings a day
<5 servings a day 2253 20.8%
Females
N Hyper-
tension
Incidence
3394 17.9%
Working Conditions
Nature of Work
Job control
1stquartile (high) 846 16.0%
2nd quartile 905 17.3%
3rd quartile 757 18.8%
4th quartile (low) 886 19.4%
Psychosocial demands
1st quartile (high) 720 14.7%
2nd quartile 1182 18.8%
3rd quartile 888 19.3%
4th quartile (low) 604 17.9%
Social support
1st quartile (high) 516 18.8%
2nd quartile 1193 17.4%
3rd quartile 1062 18.3%
4th quartile (low) 622 17.2%
Health Behaviours
Smoking
Never 1241 19.6%
Former 1308 16.7%
Occasional 137 20.7%
1-10 per day 256 18.4%
>10 per day 452 15.4%
Physical activity
Active 548 15.2%
Moderately active 796 14.9%
Inactive 2050 19.8%
Alcohol consumption
Non-drinker 508 19.9%
Non-binge drinker 2082 17.3%
Binge drinker less than once 560 21.6%
per month
Binge drinker once a month 244 10.2%
or more
Fruit and vegetable consumption
[greater than or equal to]5 1437 20.3%
servings a day
<5 servings a day 1957 16.1%
Employees working more than 10 hours per week, more
than 20 weeks in the previous 12 months (N=6611).
Table 2. Adjusted * Hazard Ratios and 95% Confidence Intervals
for Psychosocial Work Conditions and Health Behaviours on Risk of
Hypertension During a 9-year Follow-up Stratified by Gender
Males (N = 3,217)
HR 95% CI
Psychosocial Work Environment
Job control
1st quartile (high) ref
2nd quartile 1.28 (0.92-1.80)
3rd quartile 1.25# 0.90-1.75)#
4th quartile (low) 1.85 (1.26-2.71)
Psychosocial demands
1st quartile (low) ref
2nd quartile 1.23 (0.85-1.77)
3rd quartile 1.00 (0.72-1.39)
4th quartile (high) 1.30 (0.94-1.79)
Social support
1st quartile (high) ref
2nd quartile 1.01 (0.68-1.49)
3rd quartile 0.82 (0.58-1.16)
4th quartile (low) 0.91 (0.67-1.25)
Health Behaviours
Smoking
Never ref
Former 0.96 (0.71-1.31)
Occasional 1.31 (0.73-2.37)
1-10 per day 0.61 (0.33-1.15)
>10 per day 0.86 (0.60-1.23)
Physical activity
Active ref
Moderately active 1.02 (0.72-1.45)
Inactive 1.10 (0.79-1.52)
Alcohol consumption
Non-drinker ref
Non-binge drinker 1.04 (0.69-1.56)
Binge drinker less than once per month 1.23 (0.78-1.94)
Binge drinker once a month or more 1.21 (0.79-1.85)
Fruit and vegetable consumption
[greater than or equal to]5 servings a day ref
<5 servings a day 1.03 (0.81-1.30)
Females (N = 3,394)
HR 95% CI
Psychosocial Work Environment
Job control
1st quartile (high) ref
2nd quartile 0.97 (0.68-1.39)
3rd quartile 1.01 (0.71-1.44)
4th quartile (low) 0.96 (0.64-1.44)
Psychosocial demands
1st quartile (low) ref
2nd quartile 0.95 (0.63-1.44)
3rd quartile 1.24 (0.88-1.74)
4th quartile (high) 1.14 (0.79-1.64)
Social support
1st quartile (high) ref
2nd quartile 1.00 (0.67-1.48)
3rd quartile 1.00 (0.71-1.40)
4th quartile (low) 1.10 (0.78-1.54)
Health Behaviours
Smoking
Never ref
Former 0.83 (0.61-1.13)
Occasional 0.99 (0.52-1.88)
1-10 per day 0.83 (0.52-1.30)
>10 per day 0.81 (0.56-1.19)
Physical activity
Active ref
Moderately active 0.97 (0.67-1.41)
Inactive 1.07 (0.75-1.51)
Alcohol consumption
Non-drinker ref
Non-binge drinker 1.24 (0.82-1.87)
Binge drinker less than once per month 1.92# (1.21-3.04)#
Binge drinker once a month or more 1.01 (0.57-1.77)
Fruit and vegetable consumption
[greater than or equal to]5 servings a day ref
<5 servings a day 0.81 (0.63-1.03)
Estimates associated with statistically significant increased risk
of hypertension (p<0.05) are #.
* Adjusted for age, immigration
status, ethnicity, marital status, urban or rural living location,
body mass index, education, heart disease at baseline, diabetes at
baseline, activity limitations at work due to health problems,
shift schedule, occupational physical activity, work hours, weeks
worked in the previous 12 months and multiple jobs.
Table 3. Population Attributable Fractions for Modifiable
Variables Included in Our Fully Adjusted Model,
Stratified by Gender
Males Females
Psychosocial Work Environment
Job control
1st quartile (high) ref ref
2nd quartile 6.0% -0.8%
3rd quartile 3.4% 0.2%
4th quartile (low) 11.8% -1.2%
Psychosocial demands
1st quartile (low) ref ref
2nd quartile 6.1% -1.9%
3rd quartile 0.0% 5.5%
4th quartile (high) 4.9% 2.2%
Social support
1st quartile (high) ref ref
2nd quartile 0.3% 0.0%
3rd quartile -7.7% 0.0%
4th quartile (low) -1.9% 1.6%
Health Behaviours
Smoking
Never ref ref
Former -1.9% -7.4%
Occasional 1.3% 0.0%
1-10 per day -2.0% -1.6%
>10 per day -3.1% -2.7%
Physical activity
Active ref ref
Moderately active 0.5% -0.6%
Inactive 5.1% 4.4%
Alcohol consumption
Non-drinker ref ref
Non-binge drinker 1.5% 11.5%
Binge drinker less than once per month 4.3% 9.5%
Binge drinker once a month or more 4.3% 0.0%
Fruit and vegetable consumption
[greater than or equal to]5 servings a day ref ref
<5 servings a day 2.0% -12.2%
