| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
LETTERS TO THE EDITOR |
Department of Psychology, and Behavioral Medicine Research Center; University of Miami; Coral Gables, Florida (Schneiderman)
Department of Psychiatry and Behavioral Sciences, and Behavioral Medicine Research Center; Duke University; Durham, North Carolina (Williams)
"Resolved: Psychosocial interventions can improve clinical outcomes in organic disease." The Great Debate, or so it was hyped, was held on March 9, 2001, at the 59th annual meeting of the American Psychosomatic Society and subsequently published in redacted form in this journal (1–5). Five years later, the journal published an editorial (6) about the debate. The first sentence of the editorial declared that "the Great Debate was a pivotal event in the history of behavioral medicine, or at least it should have been." As agreed on in advance, however, the debate consisted of a rather modest discussion of 24 papers dealing with epidemiologic investigations, studies on intermediate outcomes, and clinical studies that had examined whether psychosocial interventions that are directed at emotional processes can change the course of serious organic disease. Excluded from the debate, at the insistence of Drs. Arnold Relman and Marcia Angell, representing the negative side, were meta-analyses, animal studies, and randomized clinical trials (RCT) directed at improving the course of serious organic diseases by changing behavior (e.g., adherence, diet, exercise). The exclusion of these studies as a condition for the debate provoked a heated exchange of e-mails before the Great Debate but was reluctantly agreed on by the affirmative side. (Otherwise, there would have been no Great Debate!)
After Drs. Redford Williams and Neil Schneiderman, who represented the affirmative side, presented epidemiologic evidence linking psychosocial variables to the development of serious organic disease, Dr. Relman conceded that "the mechanisms by which brain and mind interact with the body may be debatable, but the fact of the connection is established and we do not doubt that." Nevertheless, according to the editorial by Freedland et al. (6), "Dr. Angell’s critique of the epidemiological studies was scathing. She noted that, ‘observational epidemiological studies are fraught with difficulties, and these articles exemplify most of them.’ " Among the difficulties Dr. Angell cited was confounding, a particular problem in psychosocial studies because of their "weak effects, not strong ones" in which there is typically "a relative risk, or odds ratio, of no more than 3 or 4." We are surprised that a psychosocial factor that increases risk by 300% to 400% (i.e., a relative risk, or odds ratio of 3 or 4) would be considered to have a "weak effect."
Not cited in the editorial was Dr. Angell’s argument that "in the study of hostility on coronary artery calcification (7), it would be essential to adjust very, very finely for important factors that would be related both to hostility and to coronary artery calcification .... In fact, to be on the safe side, a good study of the question would enroll only nonsmokers, nondrinkers, and people with the same income, same education and so forth." Most qualified epidemiologists, of course, would be concerned with the external validity of any study with such a design and find it more appropriate to adjust for potentially confounding variables. Moreover, if increased smoking, drinking, and caloric intake, which are associated both cross-sectionally (8) and prospectively (9) with hostility, are indeed mediators of the increased cardiovascular disease (CVD) risk associated with hostility, such a strategy would lead to failure to detect real effects of hostility on risk, albeit effects mediated by risky health behaviors that cluster in hostile persons. We can think of many words to characterize the quality of arguments like this one advanced by the negative team, but "scathing" is not one of them.
The major challenge that the affirmative side faced in the debate is that there had been few large-scale RCTs addressing the issue of whether "psychosocial interventions can improve clinical outcomes in organic disorders." Nevertheless, one or two good RCTs should be enough to show the affirmative case has merit and should not be rejected out of hand. One RCT that was chosen for the debate was the Recurrent Coronary Prevention Project, which randomized 862 post–myocardial infarction (MI) patients to either type A counseling plus cardiac counseling or cardiac counseling alone (10). The average recurrence rate in the type A counseling plus cardiac counseling group was significantly lower (i.e., a 50% decrease) than the recurrence rate in the cardiac counseling group over 4.5 years. Subsequent follow-up study showed that the reduced cardiac recurrence rate was maintained for the 4 years after the treatment ended (11). Dr. Relman did not challenge the findings but felt that "this study is considerably weakened by the fact that the beneficial long-term effects of counseling could all be due to unmeasured, indirect effects (such as subsequent changes in lifestyle, including diet, exercise), or to differences in compliance with medical treatment." This criticism, of course, is substantially weakened because there was a cardiac counseling control group that received vigorous counseling to reduce differences between treatment groups in these lifestyle factors. A detailed comparison of changes in lifestyle and medication adherences would be helpful given 2002 versus 1986 clinical trial standards. But even if such a comparison did reveal improved lifestyle measures in the type A counseling group, it would not weaken the case that the intervention is effective; it would only indicate that one pathway to benefit was via improved lifestyle behaviors. To argue otherwise would be like saying that the well-documented decreased CVD risk associated with antihypertensive medication had nothing to do with the lower blood pressure in the medicated group.
A second RCT discussed by the affirmative side compared the effects of transcendental meditation, progressive muscle relaxation, and a lifestyle education condition on blood pressure in 111 older African Americans with elevated blood pressure (12). Both the meditation and relaxation interventions had significantly greater effects than the health education control condition, and the medication condition was significantly better than relaxation. Neither Dr. Relman nor Dr. Angell criticized this study.
Given the weaknesses noted above in the negative side’s critiques of the evidence presented by the affirmative side, we were surprised to learn, 5 years later, that "after hearing the closing arguments, the audience might have wanted to go down to the beach and drown itself en masse ..." (6, p. 180). We did not feel then, nor do we now, that "we were swept away by a scientific tsunami that day in Monterey" (6, p. 183).
One can only speculate about the motivation behind a six-page self-flagellating editorial that focuses on a 5-year-old debate that was modest in scope and inconclusive in outcome. Perhaps the editorial was prompted by the fact that a major psychosocial RCT, ENRICHD, which was published after the debate, provided null results (13). If so, the concern is probably misguided; the finding of a null or even negative result in an RCT is not unusual and typically does not throw the scientific community into despair. During the past few years, for example, the Women’s Health Initiative, a major RCT that has cost in excess of $700 million, reported that hormone replacement therapy (14) and a low-fat diet (15) do not prevent heart disease in women and that calcium and vitamin D supplements do not reduce risk of fractures (16). To our knowledge, this tsunami of null results has not produced an exodus of people, en masse or otherwise, heading down to the beach to drown themselves because the scientific community had expected the trial to produce positive results. Instead, the results will be used to guide further researched aimed at informing and improving medical practice.
Whether targeting physical or psychosocial risk factors, investigators can learn much from RCTs that report null or negative results, and such trials can provide a foundation for improving subsequent investigations. The ENRICHD RCT (13), for example, which attempted to decrease depression and improve social support in order to reduce mortality and morbidity in post-MI patients, modestly influenced psychosocial variables but overall had no effect on organic outcomes. However, a secondary analysis published in Psychosomatic Medicine found that the ENRICHD intervention may have decreased both morbidity and mortality in white men while being ineffective in women and minority patients (17). These findings suggest a need for replication in white men but, more important, the need for tailoring psychosocial interventions for both women and minority patients. Thus, for example, there may be benefit in conducting group interventions designed for and including only women.
In another secondary analysis, recently reported in the Archives of General Psychiatry (18), the risk of death or recurrent MI was significantly lower in the nonrandom subset of ENRICHD patients who were treated with selective serotonin reuptake inhibitors (SSRIs) compared with patients who did not receive these antidepressants. Though requiring replication, these exploratory findings suggest the desirability of conducting an RCT examining the effects of SSRIs on morbidity and mortality in post-MI patients. Because SSRIs influence variables such as platelet function (19), as well as depression, their influence may or may not be related to depression and needs to be assessed.
The primary purpose of an RCT is usually to determine the effects of an intervention on a particular outcome rather than to explore complex intervening mechanisms. Psychosocial interventions may influence behaviors (e.g., adherence, diet, exercise), psychophysiological processes (e.g., stress hormones), or a combination of both. To the extent that emotional processes (e.g., depression) may influence behavior (e.g., medication adherence, smoking, alcohol consumption) or physiologic processes (e.g., cortisol release), psychosocial interventions may influence one or more pathways to disease. In fact, they cannot affect disease risk without influencing these pathways. It is perhaps worth noting that in our most informative psychosocial primate model of organic disease (e.g., cynomolgus monkeys), the exacerbation of atherosclerosis depends on the interactionamong social position, an unstable social environment, and an atherogenic diet (20).
The attempt to unravel the effects of emotional dysregulation, behavior, psychophysiological processes, and organic disease is probably best accomplished in relatively small experimental studies that are able to examine multiple psychosocial and biological (including genetic) variables. In contrast, large-scale RCTs are designed to determine the extent to which specific interventions can influence morbidity and mortality. The question in the Great Debate was whether "psychosocial interventions can improve clinical outcomes in organic disease." If this is the question one wishes to address, then it makes no difference whether a psychosocially induced reduction in hostility that decreases MI recurrence (10) accomplishes its goal by altering behaviors, stress hormones, or both. As perceptively noted by Dr. Jerry Markovitz, the moderator, at the outset of the Great Debate, "Psychosocial interventions have been shown to improve the quality of life of patients with established disease, and seem to influence biological processes thought to ameliorate disease progression. Small- scale studies are useful for specifying the conditions under which psychosocial factors impact quality of life, biological factors, and disease progression. Only large-scale clinical trials, however, can determine the extent to which these psychosocial interventions may impact morbidity and mortality" (21). In the 5 years since the Great Debate took place, a large number of important mechanistic experiments have been published (22–26), which have set the stage for new RCTs. In addition, existing RCTs such as ENRICHD and the Diabetes Prevention Project (27) have provided provocative results that may guide the research agenda of future experiments and large-scale RCTs. Those interested in applying knowledge about health and behavior to improve health and prevent disease should develop new experimental studies and RCTs to move the field forward. Who knows: maybe it won’t take 25 years, as implied by Freedland et al. (6); maybe it will even happen before we and our peers "have long since retired."
DOI:10.1097/01.psy.0000227691.88628.69
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  L. POZUELO, G. TESAR, J. ZHANG, M. PENN, K. FRANCO, and W. JIANG Depression and heart disease: What do we know, and where are we headed? Cleveland Clinic Journal of Medicine, January 1, 2009; 76(1): 59 - 70. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
