| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Epidemiology and Public Health (M.K., J.E.F., M.S., D.G., M.E., R.V., M.G.M., J.H.), University College London, UK; and the Finnish Institute of Occupational Health (M.K., J.V.), Helsinki, Finland.
Address correspondence and reprint requests to Mika Kivimäki, Department of Epidemiology and Public Health, University College London, London WC1E 6BT, UK. E-mail: m.kivimaki{at}ucl.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Prospective occupational cohort study of 4250 men and 1812 women free of CHD and hypertension at study entry (the Whitehall II study). Justice was assessed at phase 1 (1985–1988) and phase 2 (1989–1990); systolic and diastolic BP at phases 1, 3 (1992–1993), and 5 (1997–1999); hypertension at phases 3 and 5; and incident CHD from phase 2 to phase 5 (231 events, mean follow-up 9.6 years).
Results: A higher level of organizational justice was associated with a slightly lower mean level of diastolic BP over time. After adjustment for age, sex, ethnicity, and employment grade, higher organizational justice was associated with lower CHD incidence. This association was not attenuated after further adjustment for measures of BP and hypertension, although these measures were associated with increased CHD risk.
Conclusions: This study suggests that sustained lower levels of BP do not represent a key mechanism through which organizational justice protects against CHD. The importance of this study is that it eliminates a strong candidate mediator of the association between organizational justice and CHD and thus allows future research to concentrate on other mechanisms.
Key Words: blood pressure • coronary heart disease • epidemiology • hypertension • organizational justice • psychosocial factors
Abbreviations: BP = blood pressure; CHD = coronary heart disease; CI = confidence interval; ECG = electrocardiogram; HR = hazard ratio; ICD9 = International Classification of Diseases, revision 9; MONICA = Multinational Monitoring of Trends and Determinants of Cardiovascular Disease.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
There are plausible mechanisms connecting organizational justice to CHD. A study of employees with multiple supervisors showed smaller blood pressure (BP) elevations on days worked under a supervisor perceived as fair compared with days worked under one perceived as unfair (10). This is a potential underlying mechanism, because lower BP levels are related to a lower risk of CHD whereas sustained high BP can cause damage to blood vessels, accelerate atherosclerosis, and increase CHD risk (11,12).
Data from the Whitehall II study of British civil servants, an ongoing large-scale prospective occupational cohort study, enabled us to examine the contribution of lower levels of BP and hypertension to the inverse association between organizational justice and CHD (13). A unique feature of the study is the availability of repeated measures of BP and hypertension after the assessment of organizational justice and the ability to link this information to data on CHD incidence over a long period of time. If lower levels of BP represent a key mechanism through which organizational justice protects against CHD, then adjustment for BP and hypertension should substantially reduce the strength of the association between organizational justice and incident CHD.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
Design
The baseline and odd-numbered phases of the Whitehall II involve a clinical screening and self-administrated questionnaire, whereas the even-numbered phases are questionnaire only. Organizational justice was measured at phases 1 and 2; systolic and diastolic BP at phase 1, phase 3 (1992–1993), and phase 5 (1997–1999), and hypertension at phases 3 and 5 of the Whitehall II study. Follow-up for incident CHD was from the end of phase 2 to phase 5.
Organizational Justice
As in earlier publications from the Whitehall II study, we used a five-item self-reported justice scale (Cronbach = 0.72 at phases 1 and 2; a sample item: "Do you ever get criticized unfairly?" (reverse scored)) (7,8,14). Participants rated their response to each of the items on a 4-point scale. For each participant, we averaged the scores of the five items at each phase. These average scores were scaled from 0 (low level) to 100 (high level) so that the score reflected the percentage of the maximum score. As long-term rather than short-term levels of psychosocial factors are assumed to affect CHD incidence (15), our exposure variable was the mean of phase 1 and phase 2 averaged scores (repeatability = 0.44 for the two measurements). The mean of these two scores was treated as a continuous variable to indicate the long-term justice levels.
Hypertension and Blood Pressure
At each phase, systolic and diastolic BP were measured twice using a random-zero sphygmomanometer (Hawksley, Lancing, Sussex, UK) with the participant sitting after a 5-minute rest. The average of these two measurements was recorded. Systolic BP was the pressure at which the Korotkoff sound was first heard clearly and diastolic BP was the pressure at which the sound disappeared. Participants were asked whether they were being treated with antihypertensive medication. In keeping with standard definitions, subjects with systolic BP >140 mm Hg and diastolic BP >90 mm Hg or on antihypertensive treatment were considered to be hypertensive.
Coronary Heart Disease
Incident CHD comprised CHD death, a first nonfatal myocardial infarction or definite angina. For the assessment of fatal CHD, >99.9% of the participants were successfully flagged at the National Health Service Central Registry, which provided information on the date and cause of death. Fatal CHD was defined according to the International Classification of Diseases, Ninth Revision (ICD 9 codes 410 through 414 as underlying causes of death) (16). Potential new cases of nonfatal myocardial infarction were ascertained by questionnaire items on chest pain (17) and physician's diagnosis of a heart attack. Confirmation of myocardial infarction according to Multinational Monitoring of Trends and Determinants of Cardiovascular Disease (MONICA) criteria (18) was based on electrocardiographic findings, markers of myocardial necrosis, and a history of chest pain in the medical records. The assessment of angina was based on the participant's reports of symptoms, with corroboration in medical records or abnormalities in a resting electrocardiogram (ECG), an exercise ECG, or a coronary angiogram. The outcome was comprised of clinically verified incident coronary events only.
Baseline Covariates
The following baseline demographic characteristics were treated as covariates in the analysis: sex, age, ethnicity (white versus other), and socioeconomic position, as indicated by employment grade (high = administrative, e.g., personnel director; intermediate = professionals and middle administrators, e.g., nurse, graphic designer; low = clerical, e.g., clerk).
Statistical Methods
Correlation coefficients were calculated to examine associations between demographic characteristics (age, sex, ethnicity, and socioeconomic position), measures of BP and hypertension, and organizational justice in a cohort combining men and women. To study the associations of BP, hypertension, and organizational justice with incident CHD, Cox proportional-hazards models were fitted. There was no evidence for any sex differences in the associations of BP/hypertension (all interaction terms p > .14) or organizational justice (interaction p = .74) on CHD. Hazard ratios for incident CHD adjusted for age, sex, ethnicity, and socioeconomic position were calculated per 1-standard deviation (SD) increase in standardized BP, hypertension incidence, and organizational justice measures. Fully adjusted models included organizational justice with all measures of BP (baseline, mean and the slope, i.e., change) and hypertension slope as well as age, sex, ethnicity, and socioeconomic position. These models were rerun by sex to confirm that the findings were similar for men and women.
Estimates of mean BP and the change in BP and hypertension over time were obtained from multilevel models with measurement occasion (i.e., study phase) nested within individuals using the PROC MIXED maximum likelihood procedure of the SAS software, version 9.1 (SAS Institute, Cary, NC). Mean exposure for each individual was estimated from the mean of predicted values. Although participants were free of hypertension at baseline, BP readings at each phase, as well as hypertension status at follow-up, varied. To obtain estimates of change of time, models with random intercept and slope were fitted with systolic BP, diastolic BP, and hypertension as the dependent variables and phase of measurement as a covariate. An unstructured covariance matrix was used to specify a random intercept-slope model with different variances for the intercept and slope and a covariance between them. The use of multilevel models to estimate means and change allowed the inclusion of participants with missing data either at phase 3 or 5.
If the effect of an exposure on an outcome is partly or wholly mediated through another variable, observational data should show an association between exposure and outcome which will be attenuated after adjustment for the mediator variable (19). The greater the attenuation, the larger the mediated effect. Applied to our hypothesis, this would mean that 1) there would be an association between organizational justice (exposure) and BP and/or hypertension (possible mediator); 2) BP and hypertension (possible mediator) would predict incident CHD (outcome); 3) organizational justice (exposure) would predict incident CHD (outcome); and 4) this association would attenuate after adjustment for BP and hypertension (possible mediator). Failure to meet all four criteria would suggest that the effects of BP and hypertension may not explain the inverse association between organizational justice and incident CHD.
Statistical significance was obtained from two-tailed tests, p < .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
Associations Between Organizational Justice, BP, and Hypertension
Table 2 shows strong correlations between the baseline, mean, and change for the two BP measures, moderate correlations between measures of BP and change in hypertension, and weak negative correlations between organizational justice and mean diastolic BP and the change in diastolic BP. After adjustment for age, sex, ethnicity, and socioeconomic position, there was also a weak negative correlation between organizational justice and change in systolic BP. Higher levels of BP and hypertension were observed in older participants and men.
|
Association of BP and Hypertension With Incident CHD
Among the measures of hypertension and BP, the slope of hypertension over time, baseline BP (systolic and diastolic) and mean BP were all strong predictors of CHD (Table 3, basic models). After adjustment for age, sex, ethnicity, and socioeconomic position, a 1-SD increase in these measures was associated with between 1.29 and 1.48 times higher hazard ratio for incident CHD. However, slope for BP was not associated with incident CHD.
|
Contribution of BP to the Association Between Organizational Justice and CHD
The adjusted (age, sex, ethnicity, and socioeconomic position) hazard ratio for CHD per 1-SD increase in organizational justice is 0.87 (95% Confidence Interval (CI) 0.77–0.98) (Table 3). Additional adjustments for measures of BP and hypertension had little or no effect on this association. This was also the case after adjustment for those specific measures of BP that were associated with organizational justice, i.e., mean diastolic BP and the change in systolic and diastolic BP, or after adjustment for all BP and hypertension measures (data not shown). A stratified analysis by sex replicated these findings (in men, the hazard ratio for justice is 0.88, 95% CI 0.76–1.02 before and after additional adjustment; in women, the corresponding hazard ratios are 0.84, 95% CI 0.67–1.04 and 0.83, 95% CI 0.67–1.04, respectively).
Sensitivity Analysis
We retested the mediation hypothesis with a design where there was no overlap between either the exposure (justice) and the mediators (measurements of BP and hypertension), or between the mediators and the outcome (CHD). In this analysis, we compared a model adjusted for age, sex, ethnicity, and socioeconomic position with a model additionally adjusted for systolic and diastolic BP and hypertension at phase 3. The effect sizes were similar with wider confidence intervals, as the follow-up for CHD started after phase 3 and the number of incident cases was substantially reduced. The hazard ratio for organizational justice was 0.85 (95% CI 0.70–1.05) before adjustment for BP and hypertension and 0.86 (95% CI 0.70–1.06) after this adjustment (n = 5635 and 86 events in both models).
We also retested the hypothesis using fatal CHD and nonfatal myocardial infarction as the outcome (n = 117) but retaining the original design. The hazard ratio adjusted for age, sex, ethnicity, and socioeconomic position was 0.85 (95% CI 0.71–1.01). There was no attenuation after further adjustment for all measures of systolic and diastolic BP and hypertension slope (adjusted hazard ratio 0.84, 95% CI 0.71–1.00).
In sum, these findings indicate that BP fails to meet the four criteria for mediator status.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The present findings accord with a prospective cohort study from Finland showing no substantial change in the association between organizational justice and CHD after adjustment for baseline risk factors including systolic BP (9). As the assessment of baseline BP does not cover trends in BP and hypertension after baseline, those findings provide an incomplete test of the mediator hypothesis. We measured BP and hypertension twice during follow-up and constructed estimates of long-term exposure to BP and change in BP and hypertension over time. Using these data provided a mediator test more sensitive than that in the Finnish investigation and exclusion of participants with prevalent hypertension at baseline reduced bias due to reversed causality between organizational justice, BP, and hypertension.
Imprecision in the measurement of BP and hypertension may underestimate their role in the association between organizational justice and CHD. However, the strong associations of the BP and hypertension measures with incident CHD suggest that our assessments tapped the risk factor accurately (20). Therefore, we consider measurement imprecision as an unlikely explanation of our results. Selection bias is possible as 31% of the baseline cohort was excluded due to missing data on organizational justice, BP, or antihypertensive treatment. Loss to follow-up was greater among older civil servants, women, non-whites, and those of lower socioeconomic position. All these characteristics were controlled for in our analyses. Thus, although selective processes may have resulted in a slight underestimation of the association between organizational justice and CHD, they are unlikely to have completely masked the role of BP and hypertension in this relationship.
Several strengths of this study reduce the possibility of a false-negative conclusion. Our findings were based on a large well-characterized cohort; a 10-year follow-up period; repeated measurements of the key factors: justice at work, BP, and hypertension; and a sophisticated statistical analysis (21). Furthermore, all components of the outcome—CHD deaths, first nonfatal myocardial infarctions, and definite angina—were confirmed in clinical records. However, future research with more diverse samples, including a larger variety of ethnic groups, is needed to confirm the generalizability of our findings.
Scientific Value of a Negative Finding
Can a negative finding such as ours be scientifically important? We believe that it can. Research interest in whether the organizational justice-CHD association is causal has increased (7,9). Although large-scale, randomized, controlled trials would be ideal to resolve this, practical problems limit their feasibility. However, observational studies with data on plausible biological mechanisms that could link organizational justice and CHD may be able to provide guidance on whether the association is likely to be causal. We believe that the importance of our negative evidence on hypertension is that it eliminates a strong candidate mediator of the organizational justice-CHD association and thus allows limited research resources to be used on testing more promising mechanisms. Moreover, reporting negative findings reduces publication bias (a tendency of journals to accept preferentially papers reporting an association over papers reporting no association).
Implications for Future Research on Mediating Mechanisms
The association between organizational justice and casual BP observed in the present study was statistically significant but not large enough to be of clinical relevance. In contrast, a previous small-scale study found that employees with multiple supervisors had a 15 mm Hg lower ambulatory systolic BP on days worked under a supervisor perceived as fair than on days worked under an unfavorably perceived supervisor (10). If sustained, this is a difference of considerable clinical significance (12,22). We recommend that future large-scale studies of organizational justice and CHD measure ambulatory rather than casual clinic BP.
Ambulatory but not casual clinic measurements enable detection of transient BP elevations at work. It is possible that BP variability, rather than general BP level, underlies the association between organizational justice and CHD (23–25). In hypertension, the magnitude of BP variability correlates closely with target-organ damage and the incidence of cardiovascular events, independent of absolute BP levels (26). Moreover, high BP variability (transient BP elevations) may act as a trigger for CHD events among vulnerable individuals with preexisting atherosclerosis within a proinflammatory and prothrombotic milieu (27). These and other possible mechanisms, such as those related to the metabolic syndrome, inflammatory responses, and health behaviors, would provide an interesting avenue for further research on factors that may mediate the association between organizational justice and CHD.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The Whitehall II study has been supported by grants from the Medical Research Council; British Heart Foundation; Health and Safety Executive; Department of Health; Grant HL36310 from the National Heart Lung and Blood Institute; Grant AG13196 from the National Institutes of Health; National Institute on Aging; National Institutes of Health; Grant HS06516 from Agency for Health Care Policy Research; and the John D. and Catherine T. MacArthur Foundation Research Networks on Successful Midlife Development and Socio-economic Status and Health. M.K. and J.V. are supported by Projects no. 105195 and 117604 from the Academy of Finland and the Finnish Environment Fund. J.E.F. is supported by Grant number G8802774 from the Medical Research Council. M.J.S. by a grant from the British Heart Foundation, and M.G.M. by a Medical Research Council Research Professorship.
DOI:10.1097/PSY.0b013e31815aaca3
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
