| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
EDITORIAL COMMENT |
The Zena and Michael A. Wiener Cardiovascular Institute, Mount Sinai Medical Center, New York, NY 10029-6574
There is a tide in the affairs of men, and one that has been flowing strongly in the field of behavioral medicine is the steadily increasing body of evidence that the psychosocial aspects of people’s environment can have a substantial effect on their physical health. Several types of chronic stress, such as bereavement, job strain, and low socioeconomic status are emerging as risk factors for chronic illness, in particular cardiovascular disease. A common factor that seems to make a situation stressful is when it is perceived as being a threat to the individual or a situation that is beyond one’s control. This concept makes intuitive sense, because if a potentially adverse situation is viewed as one over which one can exert control, it ceases to be threatening. Perceived control is an integral part of the job strain model, which has been shown to be able to predict the development of coronary heart disease and hypertension in a number of studies (1). Control has also proven to be of great importance in the Whitehall study of British civil servants, where there is a marked and continuous gradient of disease prevalence according to the occupational grade. In that study, conventional risk factors such as smoking and blood cholesterol only explain a small proportion of the cardiovascular disease-SES gradient, and control, which decreases in the lower grades, has emerged as a major contributor (2). Like many things British, the civil service is very hierarchical, and whether these results can be extrapolated to other settings is uncertain.
Most of the studies that have demonstrated these associations between psychosocial factors and chronic physical disease have been observational, and the data have grown stronger as the field has moved from cross-sectional to prospective studies. A good example of this is the association between depression and coronary heart disease: although the two coexist more commonly than would be expected by chance, there has been dispute as to whether depression is a cause or a consequence of heart disease. Several recent prospective studies have shown that depression can predict the onset and recurrence of heart disease, thus strengthening the argument for a causal role (3–6).
Organizations such as the American Heart Association have been reluctant to declare depression as a major risk factor for heart disease, perhaps because they are waiting to see the results of a major intervention trial [Enhancing Recovery in Coronary Heart Disease (ENRICHD)], which could be available in the next year or so (7).
To be sure that a putative risk factor actually is involved in the causal chain of disease development, it is necessary to show that eliminating or reducing the risk factor will lower the risk of disease. Thus it has been very difficult to relate salt intake to high blood pressure by epidemiological studies alone, and the strength of the evidence rests on interventional studies, which have shown that lowering salt intake lowers blood pressure (8). Another relevant example is the role of estrogens in the development of coronary heart disease. Numerous observational studies showed that women who take estrogens are at lower risk of heart disease; therefore, it was a major shock when the first intervention study [the Heart and Estrogen/Progestin Replacement Study (HERS)] not only found no evidence of benefit, but also evidence that there may be harm when estrogens are given prophylactically (9).
We are entering the age of evidence-based medicine, a system that has developed a series of criteria for gauging the strength of evidence supporting associations between risk factors and disease, and the effectiveness of interventions (10). Category 1, the strongest, requires evidence from randomized controlled clinical trials. Category 2 requires evidence from a quasiexperimental or controlled study, whereas Category 3 requires evidence from nonexperimental descriptive studies. For most of the evidence linking psychosocial factors with chronic disease, we are at the Category 3 stage. If behavioral medicine is to enter the mainstream of medical practice, we must begin doing Category 1 and 2 studies.
The article by Theorell et al. (11) in this issue describes the results of an attempt by a Swedish insurance company to reduce workers’ stress at a time when the workers’ job security was being threatened. There was one intervention group and three control groups. The intervention focused on the managers and could be classified as a cognitive behavioral intervention, with an emphasis on improving control for the employees. The main finding was that both managers and employees in the intervention group reported more decision authority or control at work. Serum cortisol, measured at the beginning of the workday, decreased in the employees of the intervention group, but not in the managers.
An earlier study by the same group attempted a comprehensive intervention in Swedish civil servants (12). This consisted of a 2-day educational course focusing on stress management and relaxation training, as well as attempts to increase control at work and to improve social support. This resulted in some improvement in perceived control 8 months later, but only a marginal change in social support. The changes in cardiovascular risk factors were somewhat disappointing: there was no effect on blood pressure or cholesterol, although there was a 6% change in the ratio of apolipoproteins B to A1.
The new study adds to a small but growing literature on the effects of interventions of work-related stress, which was recently reviewed by van der Klink et al (13). They identified 48 studies in which work site interventions were carried out, that could be divided into four categories: cognitive behavioral approaches, relaxation techniques, multimodal interventions, and organization-focused interventions. The outcome variables included work-related items such as decision latitude and job demands, somatic symptoms, physiological and psychological measures, and absenteeism. Only 17 of the 48 studies yielded significant effect sizes (expressed as Cohen’s d-, a measure of the standardized difference between the means of the intervention and control groups), the most effective being cognitive-behavioral interventions which focused on the individual (d = 0.68), and the least effective being the organizational intervention (d = 0.08).
Where do we go from here? The next stage in the evolution of behavioral medicine is the provision of evidence showing that interventions designed to relieve the putative causal factors such as job stress or negative affects can mitigate the course of chronic disease. There are two stages: first, we must be able to show that the intervention actually improves the independent variable. In the case of anger and depression, there are well-developed strategies that we know are effective, whereas with job strain we are not so sure. The second stage is to show that improving the causal factor will affect the outcome variable. A key issue is how far downstream we need to go to be convinced that the intervention is truly effective. The salt debate illustrates this issue well (8, 14). One of the arguments used against the importance of salt has been that although salt may affect blood pressure in some people, a low salt diet may be associated with an increased morbidity (15). Actually, the evidence for this is very weak, but it raises the issue as to whether we need to be able to prove that lowering salt intake reduces the risk of cardiovascular morbidity. In reality, no one has ever carried out a study to show that reducing salt intake improves prognosis, and probably never will.
The final question is the strength of the evidence that will be required for a behavioral or lifestyle risk factor to be recognized by the general medical community as a major risk factor. Although randomized controlled trials showing that changing the risk factor improves prognosis are the ideal to be strived for, they are not always feasible or desirable. Smoking is a good example here. There is no disagreement that quitting smoking lowers the risk of cardiovascular disease and lung cancer, but this is based on observational studies not controlled trials. This means that smoking will never rate better than a category 3 as an evidenced-based risk factor, but it is clearly neither scientifically desirable nor ethical to perform the randomized trials that would be needed to move it to category 1. Controlled intervention trials of behavioral risk factors are certainly desirable, but likely to be very expensive, particularly if the outcome is to be morbidity rather than an intermediate variable such as cortisol. Given our limited resources, it will require very careful prioritization to decide which types of behavioral intervention trials are likely to yield the best return-on-investment, but we are now at the stage where more such trials should be done. The study by Theorell et al. (11) is a model for what can be achieved.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  A Singh-Manoux, J Macleod, and G Davey Smith Psychosocial factors and public health J. Epidemiol. Community Health, August 1, 2003; 57(8): 553 - 556. [Full Text] [PDF]
|
|
|
|
|
|
J Macleod and G Davey Smith Psychosocial factors and public health: a suitable case for treatment? J. Epidemiol. Community Health, August 1, 2003; 57(8): 565 - 570. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Cesana, R. Sega, M. Ferrario, P. Chiodini, G. Corrao, and G. Mancia Job Strain and Blood Pressure in Employed Men and Women: A Pooled Analysis of Four Northern Italian Population Samples Psychosom Med, July 1, 2003; 65(4): 558 - 563. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
