Personality as a Host Risk Factor: Enthusiasm, Evidence, and Their Interaction

Abstract

In Dr. Eysenck's interesting article, he essentially addresses two questions: (a) Do personality or dispositional tendencies place individuals at heightened risk for disease, especially cancer and cardiovascular disease? (b) If such identified dispositional tendencies are changed, does one either reduce risk of disease incidence or alter the course of established disease once it is diagnosed? Certainly from reading Dr. Eysenck's article, his answers regarding both of these questions are clear. He concludes that personality factors, particularly the ones that he has identified in his and his colleagues' work, do contribute to disease risk and course, and the modification of these factors, in the appropriate direction, both reduces incidence and retards the progress of pathological biological processes. But before accepting his conclusions, it would be well to reflect on both the nature of his argument and the empirical base on which he rests his case. Actually, the first issue addressed by Dr. Eysenck-do personality factors matter in physical disease?-has been considered by both laypersons and scientists for centuries. Certainly the whole field of psychosomatic medicine (and more recently, behavioral medicine) is founded on the assumption that personality, or more broadly, psychosocial factors, contribute to physical disease. In terms of psychosocial factors playing a role in physical morbidity and mortality, probably the best evidence lies in the area of social support and social integration. In a recent review of this literature, House, Landis, and Umberson (1988) made a persuasive case for the health relevance of this variable across several populations. Additionally, specific personality types or habitual response patterns have been put forward as increasing risk for identified diseases. For example, Type A behavior (now much modified from its original form) has been exhaustively studied as a risk factor for cardiovascular disease (Matthews, 1988; Matthews & Haynes, 1986; Rosenman et al., 1975). And a passive response style, referred to as Type C (Temoshok, 1983, 1985), stoicism or helplessness (Greer, Morris, & Pettingale, 1979; Greer & Watson, 1985), adjustment (Rogentine et al., 1979; Temoshok & Fox, 1984), or object-dependent tendency (Grossarth-Maticek & Eysenck, 1990), has been identified as a risk factor for malignancy. Over the years there have also been rather careful reviews of the empirical evidence linking personality styles with disease risk. For example, in the cancer area, Cox and MacKay (1982) carried out a careful examination of the state of the evidence linking psychosocial factors broadly with cancer endpoints. They concluded that lack of emotional expression and passivity appear to be empirically associated with increased disease risk across studies. More recently, H. S. Friedman and Booth-Kewley (1987) and H. S. Friedman (1990) reviewed the evidence linking personality with cardiovascular disease and other conditions. In their 1987 review, these authors suggested that perhaps a generic diseaseprone personality does exist, characterized by unhappiness and depression-proneness, along with chronic distress. Thus, the weight of the evidence seems to suggest that certain psychosocial factors (e.g., unhappiness or apathy or isolation) probably affect health risk. But the precise nature of the psychosocial factors and the mechanisms linking such factors to biologically pathological states remain matters of dispute and speculation. It should not be surprising, then, that the answer to the second question-does the modification of such psychosocial factors alter disease risk or progression?-is not in. Intervention studies in the cardiovascular area (M. Friedman et al., 1984; Jenni & Wollersheim, 1979; Levenkorn, Cohen, Mueller, & Fisher, 1983; Powell, M. Friedman, Thoresen, Gill, & Ulmer, 1984), attempting to modify Type A behavior, with potential implications for subsequent disease risk, have shown mixed findings. In the cancer area, the most recent promising study was reported by Spiegel, Bloom, Kraemer, and Gottheil (1989), who demonstrated an effect on mortality for advanced-breast-cancer patients randomized to a psychosocial group-treatment protocol. Patients who received the social-support-group intervention lived statistically significantly longer than those randomized to the control group. The overall conclusion drawn by the scientific community is that Spiegel et al.'s work did in fact demonstrate an effect on survival from one form of malignancy as a function of psychosocial intervention. What actually made the difference, and what mechanisms were involved in the process, remain unknown. At any rate, although less is known regarding this second area with which Eysenck is concerned, this is a major research direction for the field of behavioral medicine today. Within this general historical framework, I now turn to Eysenck's article, and critically address both shortfalls, as well as positive contributions made by his effort. On this critical base, I turn again to emerging research directions for future consideration.