Psychosomatic Medicine
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Published online before print February 6, 2008, 10.1097/PSY.0b013e3181642ef5
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sacker, A.
Right arrow Articles by Bartley, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sacker, A.
Right arrow Articles by Bartley, M.
Related Collections
Right arrow Social Class
Right arrow Other Cardiovascular Medicine
Psychosomatic Medicine 70:133-140 (2008)
© 2008 American Psychosomatic Society

ORIGINAL ARTICLES

Impact of Coronary Heart Disease on Health Functioning in an Aging Population: Are There Differences According to Socioeconomic Position?

Amanda Sacker, PhD, Jenny Head, MSc and Mel Bartley, PhD

From the Department of Epidemiology and Public Health, University College London, London, United Kingdom.

Address correspondence and reprint requests to Amanda Sacker, Institute for Social and Economic Research (ISER) University of Essex Wivenhoe Park, Colchester, CO4 3SQ, UK. E-mail: asacker{at}essex.ac.uk


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
Objective: To examine differences in the effect of coronary heart disease (CHD) on health functioning according to socioeconomic position. Research on social inequality in health has tended to concentrate on differences in disease prevalence and mortality rather than on the impact of disease on functioning. As outcomes improve for those with CHD, it is important to know how their health is affected in the long term.

Methods: The analysis uses questionnaire data from phase 3 (1991–1993) to phase 7 (2003–2005) of the Whitehall II Study of civil servants (n = 8292). Differences between those in higher and lower employment grades in the relationship between CHD and physical and mental health functioning were measured according to the Short Form 36 General Health Survey (SF-36). A growth curve model of change in SF-36 physical and mental health from five repeated-measures over the 12-year period was then estimated.

Results: The differences in SF-36 health between those with and without preexisting CHD depended on employment grade. For those with CHD, physical health was initially poorer in lower grades than in higher ones; this difference persisted throughout. The mental health of respondents with CHD in the lowest grades deteriorated over time whereas for members of the higher grades, the prevailing trend was for improving mental health.

Conclusions: CHD has a more detrimental effect on physical and mental health functioning among those in more disadvantaged socioeconomic positions.

Key Words: coronary heart disease • aging • cohort studies • cost of illness • longitudinal studies • socioeconomic factors

Abbreviations: CHD = coronary heart disease; GCM = growth curve model; ECG = electrocardiogram; ML = maximum likelihood; SF-36 = Short Form 36 General Health Survey; PCS = physical health component score; MCS = mental health component score.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
There has been a welcome improvement in outcome for persons with coronary heart disease (CHD) over the 1990s (1–5). However, as survival times increase, it is important to know how living with CHD affects quality of life in the long term. Many longitudinal studies on the impact of CHD on functioning exist but these have usually been short-term studies, typically with a 3- to 6-month follow-up. Longer-term studies have taken place (6–10) but, to our knowledge, follow-up has never exceeded 5 years, neither have analyses considered measurement error bias.

Existing studies indicate that in the short term after diagnosis and treatment, physical and mental health functioning improves for most persons but by no means all patients with CHD. Factors associated with nonimprovement include depression and anxiety, and manual social class (11–13). Longer-term follow-up studies also present a complex picture. All showed poorer health functioning at follow-up for participants with CHD than control groups. In some studies, health-related quality of life at 2- and 4-year follow-up visits were found to be lower only if angina was still present (7,9). In another study, mental functioning remained poorer for those with CHD than for their peers; although the gap in physical functioning narrowed between survivors with myocardial infarction and their peers, this finding was partly attributed to the peer group becoming more disabled over time (10). This natural decline in functioning as individuals age may account for the finding by Brown et al. that at a median 4 years postmyocardial infarction, poorer-than-normative health functioning was confined to patients <65 years (8).

It is also known that socioeconomic inequalities in physical and mental health functioning increase as people age (14); physical functioning declines faster and mental functioning shows less improvement. Socioeconomic inequalities in age-related changes in health functioning may be due to the differences in disease onset or the impact of disease on functioning. Research on health inequalities has tended to concentrate on the former, that is, on factors thought to be implicated in socioeconomic variation in etiology. Less attention has been paid to the ways in which prevalent disease may interact with socioeconomic conditions of life and the possible results of this interaction for disease progression. Those studies that addressed this question have found social differences in the impact of disease on changes in health functioning (15,16).

It has been shown that health behaviors are important for physical health functioning, (17,18) and that they account for some of the socioeconomic gradient in the onset of illness and functional decline (19,20). Another Whitehall II study found that working beyond retirement at age 60 years was associated with declines in mental health, especially among lower grades (21). Improvements in mental health after retirement were only found for higher grades. Given these findings, it is important to control for health behaviors and employment status for a fuller understanding of the joint impact of CHD and socioeconomic position on health functioning over time.

Evidence is inconclusive on socioeconomic differences in the impact of CHD on physical and mental functioning, mainly due to the small number of cases and short follow-up in prior work. This paper adds to the existing literature by following up individuals over a 12-year interval. It estimates a growth curve model (GCM) of change in physical and mental health functioning in a cohort of civil servants (Whitehall II Study). Study participants completed the Short Form 36 General Health Survey (SF-36) questionnaire roughly every 3 years. It has therefore been possible to investigate change among participants with and without CHD at the beginning of follow-up. Health functioning is expected to be poorer, with physical health declining more slowly and mental health improving more slowly in persons with CHD. However, the focus here is to explore the joint influence of CHD and socioeconomic position on health functioning over time and to assess if this is independent of health behaviors and employment status.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
Study Participants
The Whitehall II cohort of civil servants was established between 1985 and 1988 (phase 1) to investigate the degree and causes of the social gradient in morbidity and mortality (22). All male and female nonindustrial civil servants aged 35 to 55 years working in the London offices of 20 Civil Service departments were invited to participate. Altogether 6895 male and 3413 female civil servants participated in phase 1, giving an overall response rate of 73%. Participants were followed up in 1989 (phase 2), 1991–1993 (phase 3), 1995–1996 (phase 4), 1997–1999 (phase 5), 2000–2002 (phase 6) and 2003–2005 (phase 7). This paper uses questionnaire data from phase 3 to phase 7 when participation rates were 79% and 67%, respectively (n = 8292).

Measures

Health Functioning
Health was assessed at all five phases by the SF-36, a 36- item instrument covering physical, psychological, and social health that is comprised of eight subscales: physical health, role limitations due to physical problems, social health, bodily pain, general mental health, role limitations due to emotional problems, vitality, and general health perceptions (23). These eight scales of the SF-36 can be summarized into physical and mental component scores using factor analysis (24,25). We used the physical health component (PCS) and mental health component scores (MCS) in the analysis. These are scaled using general US population norms to have mean ± standard deviation (SD) values of 50 ± 10. In a comparison of UK specific population norms for SF-36 PCS and MCS and scoring using US norms, Jenkinson concluded that sensitivity to change was almost identical with both methods and suggested US scoring be adopted throughout the world (26). Low scores imply poor functioning. Angina and myocardial infarction in a general US population were associated with a 2.75 to 3.75 reduction in PCS and a 0.18 to 2.36 reduction in MCS (25). A 1 standard error of measurement (2.8 PCS and 3.2 MCS points) has been recommended for a clinically meaningful change in health functioning, as has an effect size of 0.5 standard deviates (25,27,28).

Coronary Heart Disease
Potential cases of CHD were determined by questionnaire items on physician- diagnosed heart disease, or Rose angina symptoms of chest pain (29). Details of physician diagnoses and investigation results were sought from clinical records for all potential cases of myocardial infarction. Twelve-lead resting electrocardiograms (ECGs) were performed at phases 3 and 5 (Siemens Mingorec, Siemens Medical Solutions, Erlingen, Germany) and assigned Minnesota codes (30). Based on all available data (from questionnaires, study ECGs, hospital acute ECGs, coronary angiograms, and cardiac enzymes), prevalent CHD at phase 3 was identified.

Employment Grade
Employment grade was measured at phase 3 from participants’ Civil Service grade title. The Civil Service identifies 12 nonindustrial grade levels on the basis of salary. For analysis purposes, these grade levels were grouped into six categories ranging from top administrative grades to clerical/office support grades. For descriptive purposes, the grades were grouped into Administrative (grades 1–2), Executive (grades 3–5), and Clerical/Support (grade 6).

Employment Status
At each phase, participants reported whether they were still employed (in the Civil Service or elsewhere) or not.

Health Behaviors
Smoking status was classified into five groups: nonsmoker, ex-smoker, 1 to 10 cigarettes per day, 11 to 20 cigarettes per day, and ≥21 cigarettes per day. Physical activity was measured as frequency of vigorous, moderate, and mild exercise per week. Alcohol intake was measured by asking participants how many glasses of wine, pints of beer, and measures of spirits they had consumed in the preceding 7 days. Intake was converted into grams of alcohol and then subdivided into three categories using UK recommendations for sensible drinking (none, ≤ gender specific recommendations, > recommendations).

Statistical Analysis
The GCMs are specified as latent variable models that account for baseline health status at phase 3 and change over time from phase 3 to phase 7, using individually varying times of observation. The models are fitted using the software package Mplus v4 (31). The GCM is applied to all available SF-36 data and not just for individuals who responded at all five phases. The Mplus robust maximum likelihood (ML) algorithm is conditioned on complete data for the independent variables and adjusts estimates for nonresponse in the dependent variables if the data are missing at random. Fuller descriptions of GCM within a latent variable framework are available (32–34).

Initial analyses indicated that a quadratic GCM for health was overfitted, producing nonadmissible solutions that were no better a representation of individual’s health changes with time than a linear GCM. Hereafter we represent the relationship of health over time as a linear function. The baseline and slope latent variables are regressed on the time-invariant variables gender, age, and grade, all at phase 3: grade is entered into the models as a continuous variable; age and age-squared terms are centered at age 50 years; gender is coded 0 for males and 1 for females. Gender did not interact with any other independent variables in the models; neither did gender-specific models show any differences in the pattern of results to those presented.

Separate two group (those with and without CHD at baseline) models for mental and physical health are estimated. The equations for these models are given by



Formula 1



Formula 2



Formula 3

where ygjt denotes the SF-36 score for group g (g = 1,2), individual j, and time point t (t = 1,2,...,5),

{eta}gIj is the intercept growth factor, {eta}gSj is the slope growth factor, atj is the time score for individual j at time t, and xgj are the time-invariant covariates, age, gender, and grade. Then, the (V({zeta}g)) represents individual differences in the slope and intercept growth factors, and (V({varepsilon}gjt)) captures unique, time-specific individual variation in health functioning.

The impact of grade on the growth function is estimated separately for each group (those with and without CHD), thus allowing for an interaction between grade and CHD status. The statistical significance of the interaction is tested using a Wald test of equality of the regression estimates of the baseline and slope latent variables on grade across the two groups. Two sets of possible confounding variables are then included in the models: employment status and health behaviors. Accordingly, three additional GCMs are fitted to estimate the effect of grade, and its interaction with CHD, after controlling for employment status (employed, nonemployed) and health behaviors (drinking, smoking, and exercise) separately and in a fully adjusted model. The control variables are entered as time-varying variables which directly influence the observed SF-36 health scores. Equation 1 is replaced by the following for the models with time-varying covariates xgjt:



Formula 4


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 NOTES
 REFERENCES
 
Table 1 describes the demographic and physical and mental health status of respondents at phase 3 by CHD status. It also gives these data for those still in the study at phase 7. Whitehall II respondents with CHD were 2 years older on average than respondents without CHD. The high civil service grades were dominated by male and the low grades by female respondents, contributing to the excess of female respondents with CHD. A clear grade gradient can be seen in the prevalence of CHD. Respondents with CHD at phase 3 also had poorer mean PCS than respondents without, but there was a less clear difference in MCS at phase 3. Overall, PCS declined over time, whereas MCS tended to improve. Over 90% of respondents were working at phase 3 but this dropped to around 50% by phase 7. Those with CHD were less likely to be employed, drink excessive alcohol, or exercise, and more likely to smoke. There were some differences in health behaviors at phase 7 compared with phase 3 but these may reflect differential drop-out from the study. The GCMs described below take account of missing data, assuming it is random once age, gender, grade, and health in previous phases are known.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Demographic and Health Characteristics by CHD Status at Phase 3

 

Table 2 shows the differences in baseline physical health functioning and change over time in health functioning between those with and without CHD according to employment grade. The intercept values refer to mean PCS values for male members of the highest Civil Service grade (grade 1) who were aged 50 years at phase 3 of the study. The baseline intercept is the mean SF-36 PCS summary score at phase 3 for this reference group, and the mean rate of change in scores between phases 3 and 7 is given as the slope intercept. Differences in initial health scores associated with a unit reduction in grade are given in terms of the regression of baseline on grade; differences in the change in health scores between grades is given in terms of the regression of slope on grade.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Model Estimates and 95% Confidence Intervals in Parentheses for the Linear Growth Curve Model of Short Form 36 General Health Survey Physical Health Functioning From Phase 3 to Phase 7

 

After age and gender differences between the grades were taken into account, mean PCS score at phase 3 for the reference group (highest grade men aged 50 years) without CHD was estimated to be 55.94. The scores for lower grades were significantly poorer, with mean PCS score for the lowest-grade respondents estimated to be 54.29 (55.940 – 0.329 * 5). In high-grade men without CHD, PCS scores declined by an average of –0.24 points per year (slope intercept –0.24). The regression of the slope on grade showed that the rate of change in physical health was significantly faster in those in the lower grades than in those in the highest grades without CHD (slope regressed on grade –0.022). Thus, for the reference group, mean PCS was 55.94 at phase 3 and declined to 53.06 by phase 7, whereas a 50-year-old man in the lowest grade had a PCS of 54.29, which declined to 50.09.

The growth curve estimates for PCS among those with CHD are shown in the second column of Table 2. Mean baseline PCS was poorer for those with CHD than without, but the rate of decline was slower. There was weak evidence of a decline in PCS over time for those already reporting CHD: the mean rate of change for a 50-year-old highest-grade man was –0.18 (95% Confidence Interval (CI) = –0.38–0.02). Grade differences in baseline scores were considerably greater in those with CHD than in those without (Wald statistic 9.74 for 1 df, p = .002). For each 1-point difference in grade, baseline PCS was almost 1 point lower for those with CHD (–0.963), three times the difference by grade for those without CHD (–0.329). Physical function scores then declined over time at the same rate in all civil servants with CHD (slope regressed on grade –0.006). For 50-year-old highest-grade men with CHD, mean PCS was 52.75 at phase 3, which declined to 50.63 by phase 7, whereas their lowest-grade contemporaries had mean PCS of 47.94 at phase 3, which declined to 45.46. These grade differences in decline were similar to that seen for those without CHD (Wald statistic 0.55 for 1 df, p = .46).

Figure 1 shows the growth curves for 50-year-old highest-grade and lowest-grade men. At baseline, men in the highest grades without CHD had the best physical health and those in the lowest grades with CHD had the poorest health. PCS was better for a lowest-grade man without CHD than a highest-grade man with CHD. By phase 7, the difference between the best-off and the worst-off is not much altered. However, over time, PCS in lowest-grade men without CHD declined to just below that in highest-grade men with disease. That is, the impact on decline in PCS of being in the lowest grade for men without CHD was slightly greater than the impact of CHD for men in the highest grade.


Figure 11
View larger version (21K):
[in this window]
[in a new window]

  		Figure 1. Short Form 36 General Health Survey (SF-36) physical health component score (PCS) latent growth curve for Civil Service highest and lowest grades, with and without coronary heart disease at phase 3. CHD = coronary heart disease.

 

Table 3 shows the social variation in SF-36 MCS scores and change in scores in those with and without CHD. Baseline MCS at phase 3 in the reference group was better for those with no CHD (52.80) than for those with CHD (49.36). Mental health tended to improve in the reference group of highest-grade civil servants: the slope intercept was positive in those with and without CHD. There seemed to be more improvement in individuals with CHD, although this estimate lacks precision given the wide CI. There was a significant interaction between grade and CHD status on the growth function for MCS (Wald test 7.22 for 2 df, p = .03). Being a member of any other than the highest grade had both a negative effect on baseline MCS and on changes in MCS. Grade had a greater negative effect on baseline MCS among those with CHD (–0.45) and a greater impact on the rate of change in MCS (–0.06) than in those without CHD (–0.15 and –0.03, respectively), although taken alone these differences failed to reach significance.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Model Estimates and 95% Confidence Intervals in Parentheses for the Linear Growth Curve Model of Short Form 36 General Health Survey Mental Health Functioning From Phase 3 to Phase 7

 

Grade differences in the impact of CHD on MCS can be seen more clearly in Figure 2. At baseline, there were only small differences between highest- and lowest-grade respondents without CHD. However, the negative impact of CHD on MCS was compounded by low grade. By phase 7, all showed improvements in MCS with the exception of the lowest-grade civil servants with CHD. The improvement is particularly noticeable for the highest grades with CHD. Similar to the finding for physical health, the initially better health of lowest-grade men without CHD compared with highest-grade men with CHD could not be maintained over time. By the end of follow-up, MCS was the same for a lowest-grade man without CHD as a highest-grade man with CHD.


Figure 21
View larger version (20K):
[in this window]
[in a new window]

  		Figure 2. Short Form 36 General Health Survey (SF-36) mental health component score (MCS) latent growth curve for Civil Service highest and lowest grades, with and without coronary heart disease at phase 3. CHD = coronary heart disease.

 

Table 4 shows the grade effects on PCS in the smaller sample with covariate data before and after adjustment for risk factors. Employment status and health behaviors were entered into the model as time-varying variables. Grade differences in baseline PCS were attenuated by drinking, smoking, and exercise yet they were still significantly different from each other and from zero. By contrast, grade differences in baseline PCS were intensified after adjusting for employment status, as were grade differences in the impact of CHD on baseline PCS. Despite the pattern of results being similar before and after adjustment for the risk factors, the Wald test indicated that health behaviors, but not employment status, explained the interaction between grade and CHD status in their impact on PCS (Wald statistic 4.86 for 2 df, p = .09 for the former, and Wald statistic 10.83 for 2 df, p = .005 for the latter).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Estimates and 95% Confidence Intervals in Parentheses for Effect of Grade on Short Form 36 General Health Survey Physical and Mental Health Functioning From Phase 3 to Phase 7 Before and After Adjusting for Health Behaviors and Employment Status

 

MCS differences by grade were also explained by differences in behavioral risk factors (Wald statistic of equality of regression estimates 5.15 for 2 df, p = .08), but not by employment status (Wald statistic 6.79 for 2 df, p = .03). About 50% of the grade difference in baseline MCS for those with CHD and roughly 30% of the difference for those without CHD was explained by smoking, drinking, and exercise. But regardless of health behaviors, civil servants in lower grades still showed less improvement in MCS over time, especially if they had CHD; the difference in the regression of grade on the slope between those with and without CHD in the model adjusted in health behaviors was 0.042 (90% CI = 0.00–0.082). By contrast, adjusting for employment status magnified grade differences in the impact of CHD on baseline MCS because the grade difference among civil servants with CHD intensified at the same time it was attenuated for those with no CHD at baseline. Employment status explained some of the grade differences in the impact of CHD on MCS over time; but even in the fully adjusted model, members of the lower grades showed less improvement in MCS.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
NOTES
 REFERENCES
 
This study demonstrates that baseline levels of physical and mental health and the rate of change in health over time associated with CHD are affected by socioeconomic circumstances. Among London-based civil servants, baseline physical health was, as expected, strongly related to disease status and grade, but CHD had a greater negative impact among lower-grade civil servants. There were also differences between grades in the rate of decline in physical health for those with and without CHD. In study participants with CHD, there was no difference between grades in the rate of decline in physical health. In those persons without CHD, health deteriorated faster among lower grades than higher grades, until after 12 years of observation, the health functioning score for those without disease in the lowest grade was no better than that for those with disease in the highest grade.

Mental health also differed by employment grade. High-grade civil servants showed improvements in their mental health over time. In parallel with its impact on baseline physical health, CHD had a negative effect on baseline mental health. However, among high-grade respondents, the impact of CHD on mental health wore off over time although by the end of follow-up, their mental health was still not as good as that seen among high grades without CHD. Lower-grade civil servants with CHD were not so fortunate. Their declining mental health trajectories went against the prevailing trend of improving mental health.

The poorer baseline PCS and MCS for those with CHD are in line with reported effect sizes in the literature. Our findings are consistent with reports that when heart disease predates retirement age, health functioning is poorer than for community norms (8). Similar to our estimated baseline differences, a large international study found mean PCS and MCS for respondents with ischemic heart disease to be, respectively, 3.3 points and 2.1 points lower than for respondents with no reported chronic condition (35). However, the literature indicates that account needs to be taken of socioeconomic position, because it is known to influence health profiles (36). The current findings accordingly show the contrasting health functioning of respondents with and without CHD over 12 years of follow-up among those in different socioeconomic positions. A previous report also indicated that changes over time in physical functioning are partly due to the declining health of general population control respondents (10). Adding to that study, we found that improvement in mental health functioning took place although this was conditional on social circumstances; MCS scores improved in those in the highest Civil Service grades but not in lower grades.

The great majority of research on inequalities in health has focused, at least implicitly, on etiological processes. Papers have concentrated on social inequalities in prevalence or incidence, despite the call by Zimmer and House for a distinction to be made between the onset of health problems and their progression (37). A smaller body of research has found that socioeconomic disadvantage plays an important role beyond the effects of disease itself. Yelin et al. reported that social and work factors had a far larger effect on the degree of disability among rheumatoid arthritis sufferers than disease factors (38,39). Krokstad et al. found that social and educational factors were a strong influence on receipt of disability benefit in a large longitudinal study in Norway (40). Lantz et al. reported that income was related to health functioning and that the relationship could be explained by exposure to stressful life events (41). Our findings are therefore in line with the small number of similar studies carried out over the past 25 years. Unfortunately, we were not able to assess disease severity in the Whitehall II sample, but socioeconomic conditions have been found to correlate with the severity of CHD (42,43). From this evidence, the impact of CHD on health functioning seems to depend on both psychosocial and physiological correlates of socioeconomic position. These results are consistent with a social model of disability, which proposes that the impact of illness depends not just on biological differences but also on the environment in which people have to manage their illness (44). Both the early work by Yelin et al. and some of the later work by House and Lantz indicated the importance of the psychosocial as well as material environment. Our work adds to these observations by showing that environmental conditions associated with lower grade interact with physical illness to negatively affect mental health.

Martikainen and colleagues reported that health behaviors were the main determinant of socioeconomic differences in change in men's physical health (45). However, Lantz and colleagues reported that socioeconomic differentials in health status change were not explained by health behaviors (46). As in the study by Lantz et al. (46), health behaviors did not explain grade differences in changes in physical health for the CHD and non-CHD groups, even when we included time-varying measures of smoking, drinking, and physical activity. Similarly, although health behaviors explained some of the socioeconomic differential in baseline MCS, they did not explain the socioeconomic differences in MCS change among those without heart disease at baseline or the socioeconomic differences in the impact of CHD on MCS change.

In the current analysis, there was evidence of suppression effects by employment status (47). That is, when employment status was held constant, a lower grade had an even more negative effect on baseline health. This can be explained by the fact that Civil Service grade and reasons for changes in employment status are related to each other at older ages: members of higher grades are more likely to retire early for reasons other than health (for example, because their accumulated pension rights and wealth allow them to do so). Accordingly, when we compare members of different grades within employment status categories, the association between grade and health functioning becomes even clearer.

We found no evidence for any gender differences in effects. As all study participants were employed at the beginning of the study, there is a "healthy worker effect" and few cases of CHD. It is likely that the magnitude of the differences between those in the most-advantaged and least-advantaged occupations would be even greater in a general population sample. The estimates presented here are also likely to underrepresent the true impact of CHD by grade because members of the non-CHD group included those civil servants who developed CHD between phase 3 and phase 7.

A potential limitation of the analyses is selective drop out from the survey. Those from lower grades were more likely to drop out by phase 7, which could result in an underestimation of the differences in the impact of CHD on functioning according to socioeconomic position. However, the latent GCM applied a robust full information ML algorithm to all available data, correcting for bias when data are missing at random (48). The GCM also estimates the extent of measurement error in the SF-36 scores and estimates change net of this error. An advantage of the work is that the SF-36 is a good measure for investigating change in health because it can distinguish between physical and mental dimensions and is responsive to differences and changes in objective assessments of health (15,19).

Although CHD is considered a chronic illness, it seems that its effects on physical health are well managed on average, as longitudinal changes in physical health functioning were unaffected by CHD status. Neither did socioeconomic position interact with CHD to exacerbate declines in physical health. In our analysis, grade-related effects on the impact of CHD on physical health were already evident at phase 3, and over two thirds of these prevalent CHD cases had been identified at earlier phases. The results do not therefore represent social inequalities in the rate of change in health for incident cases at baseline. Lower grade is known to be associated with poorer physical health independently of CHD in this sample, and a grade-related impact of CHD may also happen early in its course. By contrast, grade-related effects on the impact of CHD on mental health are seen over the 12-year follow-up. The argument over whether depression and possibly anxiety are risk factors for heart disease or whether they are a consequence of disease is still unresolved (49). Given the high sensitivity and specificity of the MCS for detecting depression, (50) our data offer greater support for the latter theory than the former.

This long-term follow-up of respondents with and without CHD has shown that cross-sectional health functioning is poorer for those with CHD than without, and for those in more social disadvantage than less, with socioeconomic position and CHD acting synergistically to negatively affect functioning. Mental health functioning improved over time for the socially advantaged, but declined among the disadvantaged with heart disease. This study has shown that the short-term improvements in health functioning reported in previous studies do not seem to return patients with CHD to the same level of functioning as their more healthy peers. Social disadvantage acted to undermine any recovery of mental health functioning and may have important prognostic implications.

We thank all participating Civil Service departments and their welfare, personnel, and establishment officers; the Occupational Health and Safety Agency; the Council of Civil Service Unions; all participating civil servants in the Whitehall II study; and all members of the Whitehall II study team.


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 NOTES
REFERENCES
 
Author’s note: Amanda Sacker is now at the Institute for Social and Economic Research (ISER) University of Essex Wivenhoe Park, Colchester, UK.

Received for publication March 16, 2007; revision received October 12, 2007.

The Whitehall II study has been supported by grants from the Medical Research Council; British Heart Foundation; Health and Safety Executive; Department of Health; National Heart Lung and Blood Institute (HL36310), US, National Institutes of Health: National Institute on Ageing (AG13196), US, NIH; Agency for Health Care Policy Research (HS06516); and the John D. and Catherine T. MacArthur Foundation Research Networks on Successful Midlife Development and Socioeconomic Status and Health. A.S. was supported by an MRC Co-operative group grant (G0100222) during the preparation of this paper. M.B. is supported by the MRC Programme Grant G8802744: The Whitehall II Study. J.H. is supported by the National Institute on Ageing (AG13196), US, NIH.

DOI:10.1097/PSY.0b013e3181642ef5


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 

  1. Rautio A, Lundberg V, Messner T, Nasic S, Stegmayr B, Eliasson M. Favourable trends in the incidence and outcome of myocardial infarction in nondiabetic, but not in diabetic, subjects: findings from the MONICA myocardial infarction registry in northern Sweden in 1989–2000. J Intern Med 2005;258:369–77.[CrossRef][Medline]
  2. Arveiler D, Wagner A, Ducimetière P, Montaye M, Ruidavets JB, Bingham A, Ferrières J, Amouyel P, Haas B. Trends in coronary heart disease in France during the second half of the 1990s. Eur J Cardiovasc Prev Rehabil 2005;12:209–15.[CrossRef][Medline]
  3. Fox CS, Evans JC, Larson MG, Kannel WB, Levy D. Temporal trends in coronary heart disease mortality and sudden cardiac death from 1950 to 1999—the Framingham heart study. Circulation 2004;110:522–7.[Abstract/Free Full Text]
  4. Quan HD, Cujec B, Ma YJ, Johnson D. Acute myocardial infarction in Alberta: Temporal changes in outcomes, 1994 to 1999. Can J Cardiol 2004;20:213–9.[Medline]
  5. Yu ITS, Li W, Wong TW. Effects of age, period and cohort on acute myocardial infarction mortality in Hong Kong. Int J Cardiol 2004;97:63–8.[CrossRef][Medline]
  6. Bengtsson I, Hagman M, Wahrborg P, Wedel H. Lasting impact on health-related quality of life after a first myocardial infarction. Int J Cardiol 2004;97:509–16.[CrossRef][Medline]
  7. Brorsson B, Bernstein SJ, Brook RH, Werko L. Quality of life of patients with chronic stable angina before and four years after coronary revascularisation compared with a normal population. Heart 2002;87:140–5.[Abstract/Free Full Text]
  8. Brown N, Melville M, Gray D, Young T, Munro J, Skene AM, Hampton JR. Quality of life four years after acute myocardial infarction: short form 36 scores compared with a normal population. Heart 1999;81:352–8.[Abstract/Free Full Text]
  9. Kiessling A, Henriksson P. Time trends of chest pain symptoms and health related quality of life in coronary artery disease. Health Qual Life Outcomes 2007;5:13.[CrossRef][Medline]
  10. Plevier CM, Mooy JM, de Mheen P, Stouthard MEA, Visser MC, Grobbee DE, Gunning-Schepers LJ. Persistent impaired emotional functioning in survivors of a myocardial infarction? Qual Life Res 2001;10:123–32.[CrossRef][Medline]
  11. Hofer S, Doering S, Rumpold G, Oldridge N, Benzer W. Determinants of health-related quality of life in patients with coronary artery disease. Eur J Cardiovasc Prev Rehabil 2006;13:398–406.[CrossRef][Medline]
  12. Failde II, Soto MM. Changes in health related quality of life 3 months after an acute coronary syndrome. BMC Public Health 2006;6:18.[CrossRef][Medline]
  13. Le Grande MR, Elliott PC, Murphy BM, Worcester MU, Higgins RO, Ernest CS, Goble AJ. Health related quality of life trajectories and predictors following coronary artery bypass surgery. Health Qual Life Outcomes 2006;4:49.[CrossRef][Medline]
  14. Chandola T, Ferrie J, Sacker A, Marmot M. Social inequalities in self reported health in early old age: follow-up of prospective cohort study. BMJ 2007;334:990.[Abstract/Free Full Text]
  15. Hemingway H, Stafford M, Stansfeld S, Shipley M, Marmot M. Is the SF-36 a valid measure of change in population health? Results from the Whitehall II Study. BMJ 1997;315:1273–9.[Abstract/Free Full Text]
  16. Koster A, Bosma H, Kempen GIJM, van Lenthe FJ, van Eijk JTM, Mackenbach JP. Socioeconomic inequalities in mobility decline in chronic disease groups (asthma/COPD, heart disease, diabetes mellitus, low back pain): only a minor role for disease severity and comorbidity. J Epidemiol Community Health 2004;58:862–9.[Abstract/Free Full Text]
  17. Stewart A, Hays R, Wells K, Rogers W, Spritzer K, Greenfield S. Long-term functioning and well-being outcomes associated with physical activity and exercise in patients with chronic conditions in the medical outcomes study. J Clin Epidemiol 1994;47:719–30.[CrossRef][Medline]
  18. Stafford M, Hemingway H, Stansfeld S, Brunner E, Marmot M. Behavioural and biological correlates of physical functioning in middle aged office workers: the UK Whitehall II study. J Epidemiol Community Health 1998;52:353–8.[Abstract]
  19. Kumari M, Seeman T, Marmot M. Biological predictors of change in functioning in the Whitehall II study. Ann Epidemiol 2004;14:250–7.[CrossRef][Medline]
  20. Stansfeld SA, Head J, Fuhrer R, Wardle J, Cattell V. Social inequalities in depressive symptoms and physical functioning in the Whitehall II study: exploring a common cause explanation. J Epidemiol Community Health 2003;57:361–7.[Abstract/Free Full Text]
  21. Mein G, Martikainen P, Hemingway H, Stansfeld S, Marmot MG. Is retirement good or bad for mental and physical health functioning? Whitehall II longitudinal study of civil servants. J Epidemiol Community Health 2003;57:46–9.[Abstract/Free Full Text]
  22. Marmot M, Smith G, Stansfeld S, Patel C, North F, Head J, White I, Brunner E, Feeney A. Health inequalities among British civil servants: the Whitehall II study. Lancet 1991;338:58–9.[Medline]
  23. Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection. Med Care 1992;30:473–83.[Medline]
  24. Ware J, Kosinski M, Bayliss M, McHorney C, Rogers W, Raczek A. Comparison of methods for the scoring and statistical analysis of the SF-36 health profile and summary measures: summary of results from the medical outcomes study. Med Care 1995;33:AS264–AS279.[Medline]
  25. Ware J, Kosinski M, Keller S. SF-36 Physical And Mental Component Summary Measures: A User's Manual. Boston, MA: New England Medical Center; 1994.
  26. Jenkinson C. Comparison of UK and US scoring methods for weighting and scoring the SF-36 summary measures. J Public Health Med 1999;21:372–6.[Abstract/Free Full Text]
  27. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol, 2008;61(2):102–109.[CrossRef][Medline]
  28. Wyrwich KW, Tierney WM, Wolinsky FD. Further evidence supporting an SEM-based criterion for identifying meaningful intra-individual changes in health-related quality of life. J Clin Epidemiol 1999;52:861–73.[CrossRef][Medline]
  29. Rose G, Hamilton P, Keen H, Reid D, McCartney P, Jarrett R. Myocardial ischaemia, risk factors and death from coronary heart-disease. Lancet 1977;1:105–9.[CrossRef][Medline]
  30. Prineas R, Crown R, Blackburn H. The Minnesota Code Manual of Electrocardiographic Findings: Standards and Procedures for Measurement and Classification. Bristol, UK: John Wright; 1982.
  31. Muthén LK, Muthén BO. Mplus User's Guide. Fourth edition. Los Angeles, CA: Muthén & Muthén; 2006.
  32. Kaplan D. Structural Equation Modeling: Foundations and Extensions. London: Sage Publications; 2000.
  33. Duncan TE, Duncan SC. Modeling the processes of development via latent variable growth curve methodology. Structural Equation Modeling 1995;2:189–213.
  34. Little T, Schnabel K, Baumert J. Modeling Longitudinal and Multilevel Data. Mahwah, NJ: Lawrence Erlbaum Associates; 2000.
  35. Alonso J, Ferrer M, Gandek B, Ware JE, Aaronson NK, Mosconi P, Rasmussen NK, Bullinger M, Fukuhara S, Kaasa S, Leplege A. Health-related quality of life associated with chronic conditions in eight countries: results from the international quality of life assessment (IQOLA) Project. Qual Life Res 2004;13:283–98.[CrossRef][Medline]
  36. Mazeika PK. Quality of life four years after myocardial infarction: short form 36 scores compared with a normal population. Heart 2000;83:104–5.[Medline]
  37. Zimmer Z, House JS. Education, income, and functional limitation transitions among American adults: contrasting onset and progression. Int J Epidemiol 2003;32:1089–97.[Abstract/Free Full Text]
  38. Yelin E, Meenan R, Nevitt M, Epstein W. Work disability in rheumatoid arthritis: effects of disease, social, and work factors. Ann Intern Med 1980;93:551–6.[Abstract/Free Full Text]
  39. Yelin E, Nevitt M, Epstein W. Toward an epidemiology of work disability. Milbank Mem Fund Q 1980;58:386–415.[CrossRef]
  40. Krokstad S, Johnsen R, Westin S. Social determinants of disability pension: a 10-year follow-up of 62,000 people in a Norwegian county population. Soc Sci Med 1999;48:445–69.[CrossRef][Medline]
  41. Lantz PM, House JS, Mero RP, Williams DR. Stress, life events, and socioeconomic disparities in health: results from the Americans' changing lives study. J Health Soc Behav 2005;46:274–88.[CrossRef][Medline]
  42. Tyroler HA. Coronary heart disease epidemiology in the 21st century. Epidemiol Rev 2000;22:7–13.[Free Full Text]
  43. Liljefor I, Rahe RH. An identical twin study of psychosocial factors in coronary heart disease in Sweden. Psychosom Med 1970;32:523–42.[Abstract/Free Full Text]
  44. Crow L. Including all of our lives: renewing the social model of disability. In: Barnes C, Mercer G, editors. Exploring the Divide. Leeds, UK: Disability Press; 1996.
  45. Martikainen P, Stansfeld S, Hemingway H, Marmot M. Determinants of socioeconomic differences in change in physical and mental functioning. Soc Sci Med 1999;49:499–507.[CrossRef][Medline]
  46. Lantz PM, Lynch JW, House JS, Lepkowski JM, Mero RP, Musick MA, Williams DR. Socioeconomic disparities in health change in a longitudinal study of US adults: the role of health-risk behaviors. Soc Sci Med 2001;53:29–40.[CrossRef][Medline]
  47. MacKinnon DP, Krull JL, Lockwood CM. Equivalence of the mediation, confounding, and suppression effect. Prev Sci 2000;1:173–81.[Medline]
  48. Little TD, Rubin DB. Statistical Analysis with Missing Data. New York, NY: John Wiley and Sons; 1987.
  49. Nicholson A, Kuper H, Hemingway H. Depression as an aetiologic and prognostic factor in coronary heart disease: a meta-analysis of 6362 events among 146 538 participants in 54 observational studies. Eur Heart J 2006;27:2763–74.[Abstract/Free Full Text]
  50. Elliott TE, Renier CM, Palcher JA. Chronic pain, depression, and quality of life: Correlations and predictive value of the SF-36. Pain Medicine 2003;4:331–9.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 J Gerontol A Biol Sci Med SciHome page
E. Brunner, M. Shipley, V. Spencer, M. Kivimaki, T. Chandola, D. Gimeno, A. Singh-Manoux, J. Guralnik, and M. Marmot
Social Inequality in Walking Speed in Early Old Age in the Whitehall II Study
J Gerontol A Biol Sci Med Sci, October 1, 2009; 64A(10): 1082 - 1089.
[Abstract] [Full Text] [PDF]

Home page
 Psychosom. Med.Home page
A. Sacker, J. Head, D. Gimeno, and M. Bartley
Social Inequality in Physical and Mental Health Comorbidity Dynamics
Psychosom Med, September 1, 2009; 71(7): 763 - 770.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sacker, A.
Right arrow Articles by Bartley, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sacker, A.
Right arrow Articles by Bartley, M.
Related Collections
Right arrow Social Class
Right arrow Other Cardiovascular Medicine

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS