| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Public Health and Primary Care (P.K.M., R.N.L., P.G.S., N.W.J.W., A.A.W., K.-T.K.), University of Cambridge, Strangeways Research Laboratory, Wort's Causeway, Cambridge, UK; MRC Centre for Nutrition and Cancer (S.A.B.), Cambridge, UK; MRC Epidemiology Unit (N.J.W.), Elsie Widdowson Laboratories, Cambridge, UK; School of Medicine (P.K.M., R.D.S., I.M.H.), Health Policy and Practice, University of East Anglia, Norwich, UK.
Address correspondence and reprint requests to P. K. Myint, Clinical Gerontology Unit, Box-251, Level 2, F&G Block, Addenbrooke's Hospital, Cambridge, CB2 2QQ, UK. E-mail: Pkyawmyint{at}aol.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Participants included 17,777 men and women aged 40 to 79 years at baseline who lived in Norfolk, UK, and had no known cardiovascular disease or cancer, and completed the anglicized Short Form 36-item questionnaire (UK SF-36) during 1996 to 2000 in the European Prospective Investigation into Cancer-Norfolk prospective population study. We examined the relationship between mental functional health derived from the mental component summary scores of the SF-36 and mortality from all causes, cardiovascular disease, cancer, and other causes during an average 6.5-year follow-up.
Results: There were 1065 deaths during a total of 115,550 person-years of follow-up. Impaired mental health-related quality of life was associated with increased risk of all-cause mortality in men and women. A decrease of 1 SD (10 points) in SF-36 mental component summary scores was associated with a 14% increase in all-cause mortality (hazards ratio = 1.14; 95% Confidence Interval: 1.07, 1.21) after controlling for age, gender, body mass index, systolic blood pressure, cholesterol, alcohol consumption, diabetes, smoking, social class, and physical functional health.
Conclusion: Poor self-reported mental functional health is related to increased risk of all-cause mortality in men and women. Interpretation of this association requires further investigation.
Key Words: self-reported mental functional health • HLEQ • UK SF-36 • all-cause mortality
Abbreviations: UK SF-36 = Anglicized version of short form 36-item questionnaire; HRQL = health-related quality of life; MCS = mental component summary; EPIC-Norfolk = European Prospective Investigation into Cancer-Norfolk; BMI = body mass index; HLEQ = Health and Life Experiences Questionnaire; HR = hazards ratio; ICD-9 = Ninth Revision of the International Classification of Diseases; ICD-10 = Tenth Revision of the International Classification of Diseases; PCS = Physical Component Summary.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
More recently, there has been increasing interest in the assessment of health- related quality of life (HRQL). One of the most commonly used validated scoring systems in assessing HRQL is the Short Form-36 (SF-36) (5). Unlike single item self-rated health, the SF-36 provides a profile of multidimensional aspects of functional well-being including mental health. Previous studies using the SF-36 mental component summary (MCS) reported inconsistent relationships between low MCS score and mortality. Moreover, these studies were conducted only on specific patient groups that included those in receipt of coronary artery bypass grafts for cardiac ischemia (6), which showed an unexpected finding of lower MCS being associated with decreased risk of short-term mortality, and those who received dialysis (7), which showed a lower level of MCS predicting mortality.
The relationship between subjective mental health measured by the SF-36 and mortality due to all causes and common chronic diseases such as cardiovascular and cancer death remains unclear in a general population without prevalent illness. In this study, we examined the prospective relationship between self-reported mental health measured by the MCS score of the SF-36, and subsequent mortality due to all causes, cardiovascular disease, cancer, and other causes in men and women free of known diseases at baseline who were living in the general community.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Measurements
At the baseline assessment in 1993 to 1997, measurements were taken by trained staff according to standardized protocols (9). Body mass index (BMI) was calculated as weight (kg)/height (m)2. Blood pressure was measured (Accutorr Sphygmomanometer, Datascope, Huntingdon, UK) after each participant had been seated resting for 3 minutes. Two blood pressure measurements were taken and the mean values were used in analysis. Nonfasting blood samples were taken. Serum levels of total cholesterol, high-density lipoprotein cholesterol, and triglycerides were measured on fresh samples (RA 1000, Bayer Diagnostics, Basingstoke, UK), and low-density lipoprotein cholesterol levels were calculated with the Friedewald formula (10).
Responses to the questions "Have you ever smoked as much as one cigarette a day for as long as a year?" and "Do you smoke cigarettes now?" permitted us to classify the participants into smoking status as current smoker, former smoker, or those who had never smoked.
At baseline, social class was classified according to the UK Registrar General's occupation-based classification scheme (11). Social class I consists of professionals; class II includes managerial and technical occupations; class III is subdivided into nonmanual and manual skilled workers; class IV consists of partly skilled workers; and class V comprises unskilled manual workers (12). In this study, we used social class obtained at the baseline survey from 1993 to 1997. Included within a baseline health questionnaire, the participants were asked: "Has the doctor ever told you that you have any of the following?"—followed by a list of specific conditions including heart attack, stroke, cancer, and diabetes to obtain self-reported prevalent illnesses. Alcohol consumption was derived from food frequency questionnaire data collected at the baseline survey.
Predictor Variables
From 1996 to 2000, the surviving participants, then aged 41 to 80 years, were asked to complete by mail a detailed Health and Life Experiences Questionnaire (HLEQ) (13), which included the anglicized version of Short Form 36 (UK SF-36) (13–16). SF-36 functional health is assessed by eight multi-item independent health dimensions, namely, physical functioning, social functioning, role limitation due to physical problems, role limitation due to emotional problems, mental health, energy/vitality, pain, and general health perception.
The SF-36 subscales were scored on a scale from 0 (worst health) to 100 (best health). The MCS scores were calculated by using the algorithms specified by the original developers (16,17). The scores for all eight health dimensions were aggregated and transformed to z scores and multiplied by their respective factor score coefficients, and standardized as t scores (50 ± 10, mean ± standard deviation (SD)).
Outcome Measures
All individuals were flagged for death certification at the UK Office of National Statistics (ONS), with vital status ascertained for the whole cohort. ONS reports deaths in the cohort via a regular record linkage system. Causes of death were classified as death due to all causes, underlying cardiovascular disease, cancer, and other causes (noncardiovascular, noncancer deaths). Cardiovascular death was defined as Ninth Revision of the International Classification of Diseases (ICD-9), codes 401 to 448, and Tenth Revision of the International Classification of Diseases (ICD-10), codes I10-I79; and cancer deaths as ICD-9, codes 140 to 208, and ICD-10, codes C00-C97. For this study purpose, follow-up for each participant began at the date of completion of the HLEQ. We present the mortality results up to the end of July 2004, approximately 6.5 years average follow-up from the time of completion of the SF-36 questionnaire.
Statistical Analyses
We used the Cox proportional hazards model (18) to determine the independent contributions of self-reported mental functional health represented by MCS (as a continuous variable and with a 10-point decrease in MCS score equivalent to 1 SD) on all-cause, cardiovascular, cancer, and other causes (noncancer, noncardiovascular) mortality with adjustment first for age (continuous) and gender, second for age, gender, and physical component summary (PCS) score of SF-36. Hazards ratios (HRs) for all-cause mortality are calculated with further adjustment for smoking status (current smoker, ex-smoker, never smoked), BMI (continuous), diabetes (yes or no), systolic blood pressure (continuous), cholesterol (continuous), social class (class I-V), and average daily alcohol intake. Missing values for smoking, BMI, prevalence of diabetes, systolic blood pressure, cholesterol, social class, and average daily alcohol intake were imputed using multiple imputation by chained equations (19,20). All variables in the analysis models were included in the imputation model. Results for all-cause, cardiovascular, cancer, and other causes of deaths are presented as HRs (95% confidence intervals (CI)) adjusted for age and gender and PCS and other covariates included in the final model (model C). Gender differences were tested through inclusion of the gender-MCS interaction term. Analyses were repeated after removing those who died within 2 years of completion of the HLEQ to exclude participants who might be most likely to have serious preclinical disease at the time of survey.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Comparison between participants who completed the SF-36 and those who did not showed no material difference in terms of age, gender, BMI, systolic blood pressure, and cholesterol level. The prevalence of current smoking and proportion of people in manual social class were higher in the nonresponders.
Of these, 1758 participants reported having had a heart attack, stroke, or cancer at baseline survey, and were excluded from the study, leaving 17,777 men and women in the current analyses. There were a total of 1065 deaths during 115,527 person-years of follow-up. Not all participants who completed the HLEQ attended the baseline health examination and vice versa. The number of missing cases ranged from 119 for smoking status to 3159 for blood cholesterol.
Men on average had higher SF-36 MCS scores (denoting better mental functional health) than women: 53.0 ± 8.9 for men, and 51.5 ± 9.7 for women, respectively; p < .0001. Table 1 shows gender distribution of sample characteristics for the whole sample.
|
Table 2 shows HRs for mortality for all causes by 10-point decrease in MCS scores. Multivariate models adjusting first for age at the time of HLEQ completion and gender (model A) and second for age, gender, and PCS score (model B) showed increased risk of death from all causes with every 10-point (1-SD) decrease in MCS scores. Age, gender, PCS and other covariates, including BMI, systolic blood pressure, blood cholesterol, diabetes, cigarette smoking, alcohol consumption, and social class were included in the final model (model C). A decrease of 1 SD in MCS was associated with approximately 15% increased rate of all-cause mortality independently of all other risk factors adjusted. The gender interaction term was not significant (p = .17). Repeating analyses after the exclusion of early deaths within the first 2 years of follow-up showed similar results (data not shown).
|
Higher mortality rates for cardiovascular (HR = 1.17; 95% CI: 1.03, 1.32) and other causes of death (HR = 1.31; 95% CI: 1.12, 1.52) were observed with every 1-SD (10- point) decrease in MCS scores. The corresponding HR for cancer death was 1.03 (0.92, 1.16). These results are independent of all factors adjusted, using model C as described above. The gender differences were not significant for cause-specific mortality. Similar associations were observed after the exclusion of deaths occurring within 2 years of follow-up.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
We previously reported the association between self-reported poor physical health, assessed by the SF-36, and increased risk of death in both male and female participants in the EPIC-Norfolk cohort (22). In this study, we primarily focused on the MCS of the SF-36, and mortality risk controlling for PCS and other potential confounders. Our capacity to differentiate physical from mental health may have helped elucidate the role of mental health in mortality.
Much attention has been focused on the association between mental disorders and mortality rather than mental health across the population range and mortality. The strengths of our study include a prospective design, a study population drawn from the community, the ability to adjust for known biological, social, and lifestyle risk factors that could influence both mental health and mortality. We were also able to adjust the physical health, which has been shown to be associated with mortality (22).
The relationship between self-reported mental health assessed by MCS of SF-36 and mortality was independent of age, BMI, systolic blood pressure, cholesterol level, alcohol consumption, prevalence of diabetes, smoking status, and the participants' physical health (PCS scores of SF-36). We also adjusted for occupational social class as this may influence self-reported mental well-being (15) and is also related to mortality risk. It is therefore unlikely that the study findings were confounded by individual level socioeconomic status. However, we could not exclude the possibility of residual confounding.
Interpretation of the association between self-reported mental health and subsequent mortality remains unclear. Idler and colleagues (23) stated that the relationship between self-rated health and mortality may be due to a) the respondents' superior knowledge concerning their own present health status and/or past health risk, b) behavior that modifies health risk during the follow-up period, or c) some combination of these. Growing evidence also suggests that psychological factors are among the determinants of the onset and course of many chronic disorders (24). Poor mental health may contribute to increased mortality risk, indirectly through behaviors such as poor self-care in the context of physical illness, increased smoking and alcohol consumption, and poor diet (25), or directly through influencing physiologic functions such as stress hormones.
In people with physical illness, poor mental health is associated with increased mortality (26,27). We excluded participants with prevalent illnesses at baseline to minimize this possibility. However, it is still possible that the effect of prevalent subclinical disease explains some of the observed associations. There seems to be increased risk of death from cardiovascular and other causes of deaths in those who reported poorer mental health. Most of the deaths from other causes in the current study were due to respiratory- and gastrointestinal-related deaths. It is plausible that people with prevalent respiratory and gastrointestinal diseases may also have poor mental health secondary to physical limitations.
It seems that the relationship between mental health and mortality is not as strong as the relationship between physical health and mortality (every 10-point decrease in PCS score is associated with HR of 1.41 (95% CI: 1.32, 1.51) for all-cause mortality (unpublished data)). Further investigation is required to determine the association between MCS and mortality. Mental health subscale of SF-36, which contributes and correlates most to the MCS, is shown to be a good marker of depressive illness (28). Therefore, it is plausible that the MCS of the SF-36 is sensitive to the subclinical minor depressive symptoms independently of physical health. It is important to better understand the mental health determinants, measured as MCS of SF-36, to search for the mechanisms that explain mind/body links in future research.
Limitations
We required subjects to provide detailed information and participate in a long-term follow-up study; as a result, we had only a population response rate of approximately 40% at baseline, possibly limiting the generalizability of our results. However, the characteristics of this population are comparable to national samples, except for a slightly lower prevalence of smokers (8), and the observed summary scores for functional health outcome in this study population are comparable with findings in other UK studies (21).
Due to limited resources, we could not validate the cause of death but in the UK, cause of death ascertained by ONS is a reliable source and is routinely used in epidemiological studies. For the same reason, we could not confirm self-reported chronic illness by medical record review or physical examination. However, self-reported illness is usually used in large population-based epidemiological studies as a standard procedure. We did not adjust for other potential confounders such as medication use, history of psychiatric illness/diagnosis, or a diagnostic assessment of current psychiatric illness. It might have influenced the data in either way. It may have attenuated the results or the results may be underestimates of the actual association. However, we included several biological and lifestyle factors in multivariate adjustment. Another limitation of the study is our relatively brief assessment of mental health MCS scores of SF-36.
Reverse causality was addressed by a) excluding those respondents who reported heart attack, stroke, and cancer at baseline; and b) repeating analyses after the exclusion of deaths within 2 years of follow-up. Although the prospective nature of this study limits reverse causality, undiagnosed disease that may be present at the time of examination remains a potential limitation.
Implications
The current study suggests that assessment of self-reported mental health using MCS of the SF-36 may be a useful indicator for identifying populations at higher risk of death from all cause, cardiovascular, and other causes of deaths. A better understanding of the determinants of mental functional health may help to identify areas for intervention.
We thank the participants and general practitioners who took part in the study. We also thank the staff of EPIC-Norfolk and our funders.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
K.-T.K., S.A.B., and N.J.W. are principal investigators in the EPIC-Norfolk population study. P.G.S. is the principal investigator of EPIC- Norfolk HLEQ program. R.N.L. is responsible for data management, computing and data linkages. P.K.M. and N.W.J.W. conducted the analysis. All co-authors contributed in writing of this paper. KTK is the guarantor.
Received for publication August 18, 2005; revision received January 28, 2007.
DOI:10.1097/psy.0b013e318068fcd4
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. J. VINKERS and R. C. van der MAST Does Depression Specifically Increase Cardiovascular Mortality? Am J Psychiatry, September 1, 2008; 165(9): 1204 - 1204. [Full Text] [PDF]
|
|
|
|
|
|
P. G. SURTEES Dr. Surtees Replies Am J Psychiatry, September 1, 2008; 165(9): 1204 - 1204. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
