Attention to human behavior has been intertwined with the epidemiologic method in the study of health and disease at least since the days of Hippocrates. For the Hippocratean writers, the study of human behavior was an important component of all medical investigation. "One should consider the mode in which the inhabitants live and what are their pursuits, whether they are fond of drinking and eating to excess and given to indolence or are fond of exercise and labor."17 A much later talent in the history of medicine, John Snow, revealed a sensitivity to human behavioral psychology in the anecdotal evidence he cited to support his explanation of the London cholera epidemics. We recall the arrogant landlord who drank his tenant's water to prove that it was not contaminated, and forthwith contracted cholera and died; and the sick woman who lived far from the Broad Street well but had sent her son to fetch that water because it tasted best to her.46
The rapid development and amazing success of the germ theory in the early 20th century turned the view of medical scientists towards the infectious agent and away from the host. One cannot question the great scientific progress and improvement in health brought about by the efflorescence of laboratory medicine, but in turning from in vivo to in vitro concerns the science of medicine sustained a serious loss in perspective.
A reawakening of interest in human behavior is now taking place within medicine for two reasons. The first is scientists' increasing social consciousness and active concern for the human problems of 20th century society, and the second is the growing ability to study human behavior with precise and quantified methods. Techniques for measuring attitudes, beliefs, feelings, traits, group interaction and interpersonal and social conditions make it possible to include behavioral variables in medical research and practice without sacrificing the rigor of scientific method. This is a fitting accompaniment to the noteworthy advances in laboratory techniques and clinical medicine,
The behavioral sciences (I am referring here primarily to psychology, sociology and anthropology) can make great potential contributions to the scientific study of health and disease in human communities. The ability of a community to function is intimately related to the personal reaction of individuals to their social environment. Psychosomatic medicine can illuminate many layers of this interactive fabric of social environment, personal reaction, health and disease. Additional progress probably depends on the degree to which psychosomatics is able to move towards the socio-psychosomatic approach. In so doing, it would need to move from the consulting room and the hospital into the community. If psychosomatic medicine is to move from studying small samples of people in clinical settings to studying populations within communities, it will need to obtain from the behavioral sciences the basis for reliable measures which can be applied validly and economically in large-scale community studies.
What kinds of contributions can the behavioral sciences make to psychosomatic research? Three levels of contribution immediately come to mind: theories, concepts and techniques. Over the years psychologists and sociologists have developed a variety of theories in an attempt to explain and predict human behavior, and certain of these are useful frames of reference for study of the interrelations between behavior and health.6 The relevance of a particular behavioral theory, of course, depends on the issue under study; different theories would apply if the issue involved psychophysiology or habit formation of healthrelevant group behavior, Although behavioral science theories have provided many significant concepts, others equa/íy important were derived from pragmatic generalizations by astute observers. These concepts provide health scientists with an established symbol system which supplements and extends psychiatric theories and thus aids in describing and interpreting human behavior. The third contribution is the methodologies for gathering, measuring and analyzing information about knowledge, attitudes, feelings, personality structure, overt behavior, social interaction, cultural norms and other important variables. A wide variety of psychological tests and sociological indices are now available, and the methodologies themselves can be used to construct new rating scales, questionnaires or tests for any kind of overt behavior or attitude system which the clinician can describe clearly and identify constantly.
The behavioral sciences can make important contributions in a number of health areas, but I should like to confine my discussion to four frontiers of psychosomatic research: (1) studies of the etiology of disease; (2) studies of psychological and behavioral concomitants of physiological change; (3) studies of the "sick role" and seeking of medical care; and (4) studies of the course of disease: recovery, chronicity and fatal outcome.
STUDIES OF THE ETIOLOGY OF DISEASE
When epidemiologic study is pursued in a programatic fashion, investigators are soon confronted with the necessity to consider psychic factors. These can affect health adversely either through covert physiologic mechanisms or through overt behaviors, and psychosomatic medicine needs to discover the psychodynamics in the latter situation as well as the former.
Epidemiologists studying lung cancer have considered many possible causal variables including heredity and the physical environment (particularly atmospheric pollution), but the strongest risk factor proved to be a behavioral one, cigarette smoking. Studies of the epidemiology of lung cancer repeatedly lead back to the behavior of cigarette smoking. We should not stop at this point, however, but should try to push back the veil of the unknown one step further and study the determinants of smoking.
Two groups of variables have been found associated with this behavior: host factors and environmental factors. A recent study of a large group of employed men from a variety of industries revealed that heavy cigarette smokers differed from others (those who had quit smoking or who had never smoked) in serum lipids, hematocrit, type of occupation, level of income and amount of schooling.27 Studies of the smoking habits among adolescents have revealed the importance of the social environment; young people who smoke are more likely to associate with a peer group where smoking is the norm and to have parents who smoke, in clear contrast to nonsmoking adolescents in the same schools.2
We now turn to a very different disease, paralytic poliomyelitis. Epidemiologic studies of polio have shown that the individuals most at risk of a paralytic polio attack are those who failed to obtain a natural immunity through very early childhood exposure to the polio virus and who have further failed to obtain an artificial immunity through receiving an effective vaccine. (Thus the epidemiology of paralytic polio is reduced to the epidemiology of the absence of appropriate antibodies.) Serological surveys conducted in various parts of the world before the mass distribution of polio vaccines revealed the association between environment and antibody levels. It was first thought that tropical environments favored the dispersion of the polio virus with consequent early acquisition of naturai immunity and reduced risk of paralytic polio later in life. Later studies showed that the key factor was not the geographic latitude of the community but its social class, a sociological concept.39 Backett subsequently pursued this association to more specific levels of cause, and these were found to be behavioral such as the play habits and personal hygiene of children.1
The development and distribution of safe, live-virus vaccines means that in an environment where sanitation is good and infant mortality low, the epidemiology of susceptibility to polio (i.e. lack of antibodies) can be reduced to the epidemiology of nonacceptance of polio vaccine. Such nonacceptance has been studied intensively, and while standard epidemiologic variables of age, race and sex have been associated with the rate of vaccine acceptance, sociological and psychological variables emerged as even more potent predictors. The perception that most of one's friends had taken vaccine or a personal orientation towards planning ahead, trust in medical experts, ability to postpone gratification and a generally "productive" orientation were strongly associated with the acceptance of polio vaccine, and these traits can be studied in communities.29 Would it not also be appropriate to turn to psychiatry, psychology and psychosomatics to clarify the psychodynamics behind the nonacceptance of proven measures for disease prevention?
Psychological and social factors have also been found to relate quite directly to the dynamics of respiratory infections. Holmes' studies argue for the complicity of life change and life crisis in timing the clinical appearance of tuberculosis,20 and Hinkle and colleagues have demonstrated the temporal clustering of life stress and illness over a long time span.16 A series of studies of military recruits showed that particular human environments generate different levels of nonspecific susceptibility to illness. The social environment was a more accurate predictor of differences in sickness rates than were such specific causes as spread of an infectious organism or exposure to a common hazard,47 and incongruity in social status (that is, simultaneous possession of attributes derived from contrasting social levels) was associated with higher risk of respiratory infection.49
THE RISK FACTORS OF CORONARY HEART DISEASE
Coronary heart disease, the pandemic of the modern world, is a disease of complex etiology, but epidemiologists have traditionally been reluctant to consider systematically the role of behavior in both predisposition to and precipitation of coronary disease. CHO has both behaviorai and biological risk factors, and many of its biological factors are demonstrably influenced by psychological ones.
Most investigators agree that the best-established risk factors for CHD are cigarette smoking, elevated serum cholesterol, elevations of other lighter-density serum lipids and elevated blood pressure.8 Excess weight has been implicated by some but not by others. Diets rich in fats have been associated with coronary disease when whole nations are compared, but diet has not usually been found associated with CHD rates9 (nor even with serum cholesterol levels) when free-living persons within a nation are studied.30,37 Let us consider the role of behavior in each of these risk factors.
Cigarette smoking is obviously a learned behavior; the habit is initiated by imitation of desired role models and maintained through both pharmacologic and psychologic reinforcement. The most effective procedures for discontinuing the habit are behavioral: the learning of incompatible responses, substitution of other oral gratifications and deconditioning through eversive reinforcement. Considerable experimental and clinical literature exists on the influence of psychological factors on serum lipid levels, and it is common knowledge that excitement raises the blood pressure. Both these issues are discussed in the next section of this paper.
Three additional CHD risk factors are diet, obesity and exercise. Dietary practices are influenced by an individual's social and cultural background as well as by his current economic state. Psychoanalytic theory holds that early experiences, particularly the relationship with the mother and the act of feeding, make a lifelong imprint on character structure, and even those who do not accept many derivations from psychoanalysis find it hard to deny that childhood eating experiences influence dietary habits and the role of food in the psychic economy.7·15 Obesity is an area of traditional concern to psychosomatic researchers, taking on added social importance because of its association with several of our major killing and debilitating diseases.
The amount of physical exercise performed is related to both CHD incidence and mortality.10 Much of the average person's daily physical exertion occurs in his occupation, but now that even blue-collar occupations are becoming largely sedentary, voluntary (nonoccupational) physical activity emerges as the major exercise opportunity for most people. Sheldon has asserted that the propensity for active sports is in part constitutionally determined.45 It is also likely that habits of physical activity can be learned and, if socially reinforced, will be maintained. Psychologically oriented studies could isolate and specify the personality characteristics of persons most and least likely to engage in physical exercise, and could delineate the social circumstances that best promote such healthful activity.
In addition to these traditional risk factors with their psychosocial dimension, research studies implicate some purely psychological and social variables as additional, independent risk factors for CHD.24 They are observed and measured at the psychosocial level, but their roles in the pathogenesis of coronary disease are no doubt mediated by neurophysiological or biochemical mechanisms.
The hypothesis that stress promotes CHD had led to including indices of anxiety and neuroticism in heart research. High scores on some scales of the Minnesota Multiphasic Personality Inventory (MMPl), especially Hypochondriasis and Hysteria, have been repeatedly associated with coronary disease, particularly when it manifests itself as angina pectoris. Other psychometrically-derived scales of anxiety or stress have registered higher scores among coronary disease patients, but in general these scales have not been used in subsequent studies so the generality of the findings remains untested. It should be noted, however, that in studies where the stress-anxiety issue was approached both by obvious interview questions and by use of a psychometrically validated scale, only the latter showed a significant association with coronary disease.24
One of the more intensively studied psychosocial variables is the coronaryprone behavior pattern, a style of living characterized by competitiveness, striving for achievement, aggressiveness (though this is sometimes stringently repressed), haste, impatience, restlessness, hyperalertness, explosive speech, tense facial muscles and persistent feelings of being under the pressure of time and the challenge of responsibility. This concept has developed historically at least from the time of Osier.11 A broad array of studies all over the world converges in finding that one or more aspects of these symptoms are indeed correlated with the presence of CHD.40 The work of Rosenman, Friedman and their colleagues has shown that the coronary-prone behavior pattern (Type A) is associated both with CHD prevalence and with CHD incidence in a prospective study of a large industrial population,44 and their findings have been confirmed in the prevalence survey among monks reported by Caffrey.s
The cardiologists who developed the concept of the coronary-prone behavior pattern and directed the research of the Harold Brunn Institute and the Western Collaborative Group Study have called upon psychologists and psychometricians for a wide variety of scientific contributions. The interview procedure for determining the coronary-prone behavior pattern was subjected to content analysis, and clusters of items having similar content were studied in terms of their relationship to "silent" myocardial infarctions and blood lipid levels.23 A psychologist conducted reliability studies of the interview procedure to determine the degree of interjudge agreement and the stability of the behavior type over time. Other psychologists learned the procedures for behavior typing and applied them in independent studies of coronary disease.5·32 Bortner has developed new methods for ascertaining behavior type, a battery of performance tests and a short self-administered rating scale.3 A cardiologist and an electrical engineer, using advanced electronics, developed a method for analyzing the patterns in changes of voice amplitude which are associated with the behavior type and with existing CHD.12
An intensive research program for more precise measurement of the coronary-prone behavior pattern was initiated in 1964 by the Harold Brunn Institute and the University of North Carolina's Department of Epidemiology. Thus far it has been shown that higher Type A scores (i.e. higher scores on the A-B scale of the Jenkins Activity Survey) are found in patients having coronary disease as compared to nondiseased individuals in the same population.28 Patients with CHD who subsequently have a recurrent episode of disease, such as another myocardial infarction, have higher scores on the A-B scale than do those patients who continue without further infarctions. Preliminary study of a limited sample of men who developed their initial coronary attack after taking the JAS does not show them to have scored significantly higher than their fellows who remained healthy. This raises the possibility that the test in its present form may not be able to predict the CHD candidate even though it has modest success in distinguishing between the behavior traits of men already having coronary disease and those without the disease.26 Further study and refinement of this test are needed.
STUDIES OF PSYCHOLOGICAL AND BEHAVIORAL CONCOMITANTS OF PHYSIOLOGICAL CHANGES
Alterations in physiological function commonly accompany changes in the individual's life situation or his interpretation of the situation. Such alterations may be adaptive or maladaptive. Sustained deviations of certain biochemical and physiological parameters have been implicated as risk factors for a number of diseases.
The serum lipids, particularly serum cholesterol, have been identified for several decades as risk factors in coronary heart disease and cerebrovascular accident. Researchers of a variety of backgrounds, but sharing a psychosomatic bent, have searched out the behavioral and emotional correlates of deviations in serum lipid levels. One of the earliest of these studies, by Friedman, Rosenman and Carroll,13 observed cyclical changes in the serum cholesterol of a group of accountants who were studied for a year. Many of the accountants showed a sharp rise in serum cholesterol late in March and during most of April, which led to the hypothesis that stress of income tax preparation increased as the April 15th deadline approached and was accompanied by an elevation of serum cholesterol. This hypothesis was strengthened when the researchers discovered that many of the men who did not show this choiesteroi rise during April were corporate accountants whose deadline period coincided with the closing of the fiscal year, December 31. A wide variety of subsequent studies have corroborated this association between work pressure and elevated cholesterol.
The literature suggests that somewhat different emotional parameters may be involved in altering the levels of lighter-density lipids - serum triglycerides and beta and pre-beta lipoprotein fractions. These levels seem to rise when the individuai is either in a state of excitation or is actively suppressing his impulses. It is not known whether these behavioral factors operate through alteration of diet or exercise, or act directly through the central nervous system or endocrine system to affect the mobilization or degradation of these substances in the blood. The general thrust of these findings is that more active, aggressive, outgoing, stimulated people have higher cholestérols as do those who are conscientious, suppress their impulses, show high levels of self-control and propriety, and are self-critical.25
Psychosomatic research has concerned itself for some time with the hypertensive patient and has demonstrated the role of anxiety and hostility in raising blood pressure levels.41 Recently, Hokanson has shown that experimental manipulation of hostility also influences blood pressure. His experiment placed college students in a game situation where they could give or receive electric shocks to the finger. Blood pressures rose in subjects receiving many shocks. If these subjects were then allowed to retaliate by shocking the experimenter, their blood pressures promptly declined to baseline levels; when the opportunity for revenge was withheld, blood pressure remained elevated for a longer time.18,19
Adults have been trained through operant conditioning to alter blood pressure by changing their heart rate, increasing or decreasing it independent of respiration rate or overt body movement.36 Various other laboratory experiments and observations of natural situations have also shown the connection between psychological factors and blood pressure.38 It must be pointed out that most of this experimental work deals with transient alterations in blood pressure level rather than with sustained hypertension, but epidemiologic studies have shown that a single casual reading of elevated blood pressure has predictive value concerning CHD risk.
Serum uric acid has occasionaJJy been suggested as a risk factor in coronary disease, but most large-scale studies have not found it a valuable predictor. It is invariably present in high concentrations, however, in persons with gout. Folklore has long held that gout is the result of extravagant living, particularly over-consumption of rich food and drink; and studies have shown that indeed serum uric acid levels average higher in upper and middle social classes than in lower social classes. High levels have also been noted in geniuses and high-achieving individuals. Psychologists sought to distinguish among these three related concomitants by use of intelligence tests, scales measuring drive for achievement and a social class index for each of these three variables could be measured on a continuous scale and its relation to serum uric acid studied simultaneously with other controlled variables. The work of Brooks and Mueller4 and of Kasl and his co/teagues31 suggests that drive for achievement is the most important correlate of serum uric acid levels among these three factors. The contribution of behavioral scientists was to introduce the methods of measurement on continuous scales of intelligence and drive for achievement.
STUDIES OF "SICK ROLE" AND SEEKING OF MEDICAL CARE
As we place greater emphasis on studying the utilization of health services, psychosomatic researchers will be called upon to delineate the various factors that influence people to seek help from health professionals. Certainly the presence of disease or physiological changes brings a person to the doctor, but not all people with these same conditions will seek medical care. However, many persons present themselves at clinics and doctors' offices without detectable disease and without apparent physiological abnormality. The use of professional time and facilities by these people is substantial.
Medical sociologists have defined as the "sick role" that set of attitudes towards the self which lead a person to claim the privileges afforded to the "sick" and to initiate corrective actions.34 If we are to better understand the utilization of medical services, we must understand how the sick role is learned and used. Identifying the reasons behind unwarranted use of the sick role could make possible reduction of the load on medical care facilities, and learning how this role is denied inappropriately could make it possible to bring into earlier care those people whose denial and delay make subsequent curative medicine more difficult. While sick role is a concept developed by sociologists, understanding its dynamics is a psychiatric problem and effective measurement of its properties is a psychometric one.
Closely related to the sick role is the way in which culture and learning alter the modes of experiencing and expressing physical symptoms. Zola,50 in a study of outpatients at an ear, nose and throat clinic in Boston, found that patients of Italian descent described their symptoms in expansive terms with vivid and dramatic language, implying the involvement of the whole body. In contrast, patients of Irish descent reported localized symptoms in rather laconic terms and did not seem to be suffering unduly. Analysis of the medical charts revealed that the two groups had essentially the same distribution of diagnoses. The interrelation of culture, personality and health is obvious in this example.
STUDIES OF THE COURSE OF DISEASE: RECOVERY, CHRONICITY AND FATAL OUTCOME
Different processes seem to be involved in primary resistance to disease than are involved in recovery after a disease or injury has occurred. A rather broad variety of precursors has been reported as being associated with specific disease risks, but rapid recovery from disease and disability appears to coalesce around two major psychological factors: high ego strength and lack of depression.
Delayed recovery from mononucleosis was found to be associated with low scores on the ego-strength scale of the MMPI and similar scales,14 and delayed recovery from chronic brucellosis was associated with depression and morale loss as measured on the MMPI scales.21 Both of these studies, however, were retrospective, and one might properly ask whether the delayed recovery might encourage patients to report complaints and inadequacies consistent with lowered ego strength and greater depression. To answer this question of which came first, lmboden and his colleagues did a prospective study of recovery from Asian influenza.22 The MMPI scores did not correlate with who became ill with the flu and who remained free of disease, but the depression and morale-loss scores obtained before illness were again significantly lower in those who recovered from the flu in less than median time.
The Bender-Gestalt test was administered to persons with gastric ulcer, and the results evaluated by a technique which gives scores indicative of high ego strength and impulse control. Persons with intractable ulcer scored much more poorly on these variables.42 In a replication study similar findings were observed.
Physicians and nurses working on hospital wards containing numbers of terminal cancer patients have commented that the friendly, cooperative patients seemed to die more quickly while the complaining, resistive, cantankerous patients survived longer. It would be easy to assert that the difficult patients just seem to survive longer and dismiss the whole hypothesis. Klopfer decided to test this issue empirically and found that the clinical personnel's observation was indeed true. Certain indices on the Rorschach test discriminated between patients with fast-growing tumors. He also discusses previous studies, one of which used the MMPI, that show similar findings.35
SURVIVORS AND NONSURVIVORS AMONG AGING
Tests measuring personality attributes and flexibility of thought processes have been found to discriminate between survivors and nonsurvivors in an aging population. Riegel and his colleagues found seven test scores in their lengthy battery that predicted which persons between ages 55 and 64 would die in the ensuing five-year period. His predictions were significantly better than chance though of course far from perfect.43 Rigidity and dogmatism were among the significant variables.
Two prospective studies of open-heart surgery, performed at about the same time in different medical centers, agreed in showing that fatal outcome of surgery is far more frequent in patients who are depressed or hopeless before the operation. Anxiety did not correlate nearly so strongly with risk of fatal outcome. Both Kimball at Yale University and Tufo and Ostfeld in Chicago found that over 80 per cent of patients who showed severe preoperative depression and inability to discuss postoperative plans died during or after surgery. This is in contrast to approximately 20 per cent average mortality for the total patient groups.33,48
These examples show the importance of psychiatric concepts in studying recovery from disease, treating situations of delayed recovery and assessing more adequately the risks of adverse response to surgery. Psychosomatic medicine's contribution to such practical matters of patient care is still in its infancy, and rapid advance is possible.
The field of psychosomatic medicine has traditionally been concerned with the question, "What facets of personality or behavior raise chances of various disease manifestations?" This will continue to be its key concern in the years immediately ahead. What I have presented are some ideas concerning theories, concepts, methodologies and research settings that will play an increased role as psychosomatics renews its attack on the traditional questions.
Theories developed by the behavioral sciences about the interaction of culture, social structures, personality, styles of behavior and common habits will generate new hypotheses for psychosomatic research. The concepts and language of the behavioral sciences, when used selectively and with an understanding of their limitations, can provide basic building blocks for the "behavior" side of the "behavior-disease equation." The methodologies of epidemiology, survey research and psychological measurement can strengthen the research designs, data-gathering systems and data-analysis techniques of psychosomatic research. The opportunities for developing reliable quantitative measures of clinical concepts are particularly challenging. Social and epidemiologic perspectives beckon the psychosomatic researcher to use the community - the natural habitat of health and disease- as his laboratory. Finally, psychosomatics may profitably broaden its areas of concern to include not only the pathogenesis of disease but also alterations in physiological functioning, the psychodynamics of inappropriately -claiming or rejecting the sick role and the processes altering rate of recovery and rehabilitation. J offer these intimations about anticipated developments not as original ideas, but rather as emerging changes in emphasis within a field which has traditionally been characterized by continual evolution and growth.
1. Backett. E. M. Social patterns of antibody to virus. The Lancet Il (1957), 778.
2. Bajda, L. A survey of smoking habits of students of Newton High School- a cooperative project. Amer. J. Pub. Health 54 (1964). 441.
3. Bortner. R. W. and Rosenman, R. H. The measurement of pattern A behavior. J. Chronic Dis. 20 (1967). 525-533.
4. Brooks, GW, and Mueller, E. Serum urate concentrations among university professors. JAMA 195 (1966), 415.
5. Caffrey, B. A multivariate analysts of socio-psychological factors in monks with myocardial infarctions. Amer. J. Pub. Health 60 (1970). 452.
6. Cassel. J. C. Social science theory as a source of hypotheses in epidemiologic research, Amer. J. Pub. Health 54 (1964), 1482-1488.
7. Dwyer, J. T., Feldman, J. J. and Mayer, J. The social psychology of dieting. J, Health and Soc. Behav. Il (1970). 269.
8 Epstein, F. H. The epidemiology of coronary heart disease: a review. J. Chronic Dis. 18 (1965). 735,
9. Epstein, F. H. Coronary heart disease epidemiology. In Trends in Epidemiology, ed, G. T. Stewart. Springfield: C. C. Thomas. 1971.
10. Frank, C. W.. Weinblatt, E.. Shapiro et al. Myocardial infarction in men - role of physical activity and smoking in incidence and mortality. JAMA 198 (1966). 1241.
11. Friedman, M, Pathogenesis of Coronary Artery Disease. New York: McGraw-Hill, 1969.
12. Friedman, M., Brown, A. E. and Rosenman, R. H. Voice analysis test for detection of behavior pattern: responses of normal men and coronary patients, JAMA 208 (1969), 828.
13. Friedman, M., Rosenman, R. H. ano Carroll. V. Changes in serum cholesterol and blood clotting time in men subjected to cyclic variation of occupational stress. Circulation 1 7 (1958), 852.
14. Greenfield, N. S,, Roessler, R, and Crosley, A. P. Ego strength and length of recovery from infectious mononucleosis. J. Nerv. Ment. Dis. 128 (1959), 125.
15. Hamburger, W. W. Psychological aspects of obesity. Bull. N. Y. Acad. Med. 33 (1957). 771.
16. Hinkle. L. et al. Ecologie investigations of the relationship between illness, life experience and social environment. Ann. Int. Med. 49(1957), 1373.
17. Hippocrates. On air. water and places. Cited in Epidemiologic Methods, Preface, B. McMahon et al. Boston: Little, Brown, 1969.
18. Hokanson. J, E. and Edelman, R. Effects of three social responses on vascular processes. J. Abn. Soc. Psychol. 3 (1966), 442.
19. Hokanson, J, E., Willers, K, R. and Koropsak, E. The modification of autonomic responses during aggressive interchange. J. Personality 36 (1968), 386.
20. Holmes, T. H. Multidiscipline studies in tuberculosis. In Personality, Stress and Tuberculosis, ed. P. J. Sparer, New York: Intemat'l Universities Press, 1956.
21. Imboden, J1, Canter, A.. Cluff, L. et al. Brucellosis: III psychologic aspects of delayed convalescence, Arch. Int. Med. 103 (1959), 406.
22. Imboden. J. B., Canter. A. and Cluff, L. E. Convalescence from influenza, Arch. Int. Med. /05(1961),
23. Jenkins, C. D, Components of the coronary-prone behavior pattern: their relation to silent myocardial infarction and blood lipids. J. Chronic Dis. 19 (1966).
24. Jenkins, C. D. Psychologic and social precursors of coronary disease. N. Eng. J. Med. 284 (1971). 244. 307. (In two 25. Jenkins, C. D., Hames. C. G., Zyzanski, S. J. et al, Psychological traits and serum lipids. I. Findings from the California Psyetiological Inventory. Psychosom. Med. 31 (1969). 115.
26. Jenkins, CD.. Zyzanski, S. J., Rosenman, R. H. et al. Association of coronary-prone behavior scores with recurrence of coronary heart disease. J. Chronic Dis. 24 (1971). 601.
27. Jenkins. C. D., Rosenman, R. H, and Zyzanski, S, J. Epidemiologic study of cigarette smokers and nonsmokers. Submitted for publication (1971).
28. Jenkins, C. D., Zyzanski, S. J. and senman. R. H. Progress toward validation of a computer-scored test for the type A coronary-prone behavior pattern. Psychosom. Med. 33(1971), 193.
29. Johnson, A. L, Jenkins. CD.. Patrick, C et al. Epidemiology of Polio Vaccine Acceptance: A Social and Psychological Analysis. Florida State Board of Health, Monograph #3 (1962).
30.Kahn, H. A., Medalie, J. H., Neufeld, N. et al, Serum cholesterol: its distribution and association with dieting and other variables in a survey of 1 0.000 men. Israel J. Med. Sci. 5(1969). 1117.
31. Kasl, S. V.. Brooks, G. W. and Rodgers, W. L. Serum uric acid and cholesterol in achievement behavior and motivation. I. The relationship to ability, grades, test performance and motivation. JAMA 213 (1970). 1158.
32. Keith. R. A.. Lown. B. and Stare. F. J. Coronary heart disease and behavior patterns: an examination of method, Psychosom. Med. 27(1965). 424-434.
33. Kimball. C P. The experience of openheart surgery: psychological responses to surgery. Psychosom. Med. 30 (1968), 552.
34. King, S. H. Chapter 4 in Handbook of Medical Sociology, ed. M. E. Freeman, S. Levine and L. G. Reeder. Englewood Cliffs. New Jersey: Prentice-Hall, 1963. See especially pages 111-116.
35. Klopfer, B. Psychological factors human cancer. Jour. Proj. Tech. 21 (1957), 331.
36. Levene. H. I,, Engel. B. T. and Pearson, J. A. Differential operant conditioning of heart rate. Psychosom. Med. 30 (1968), 837.
37. McDonough, J. R., Hames, C. G., Stulb, S. C et al, Coronary heart disease among Negroes and Whites in Evans County, Georgia. J. Chronic Dis. 18 (1965), 443.
38. McGinn, E. F., Harburg, E., Julius. S. et al. Psychological correlates of blood pressure, Psychol. Bull. 61 (1964). 209.
39. Melnick, J. L.. Walton, M.. Isaacson, P. et al. Environmental studies of endemic enteric virus infections. I. Community seroimmune patterns and polio virus infection rates. Amer. J. Hyg. 65 (1957), 1.
40. Myocardial infarction and other psychosomatic disturbances. Special Issue, Psychother. Psychosom. 16(1968), 189-292.
41. O ken. D. An experimental study of suppressed anger and blood pressure. Arch. Gen. Psychiat. 2 (1960), 441.
42. Pascal. G. R. and Thoroughman, J. C Relationship between Bender-Gestalt test scores and the response of patients with intractable duodenal ulcer to surgery. Psychosom. Med. 20 (1964). 625.
43. Riegel, K. F., Riegel, R. M. and Mayer. G. A study of the dropout rates in longitudinal research on aging and the prediction of death. Jour. Pers. and Soc. Psych. 3(1967), 342.
44. Rosenman, R, H., Friedman, M,, Straus, R. et al. Coronary heart disease in the Western Collaborative Group Study. A follow-up experience of 41/2 years. J. Chronic Dis. 23 (1970), 173.
45. Sheldon, W, H. The Varieties of Temperament: A Psychology of Constitutional Differences. New York: Harper, 1944.
46. Snow, John. On the mode of communication in cholera. In Snow on Cholera: A Reprint of Two Papers. New York: Commonwealth Fund, 1936.
47. Stewart, G. T., Voors. A. W.. Jenkins, C D. et al. Determinants of sickness in Marine recruits. Amer. J. Epidem. 89 (1969), 254.
48. Tufo, HM. and Ostfeld. A, M. A prospective study of open-heart surgery. Psychosom. Med. 30(1968), 552.
49. Voors, A. W-, Stewart. G. T., Gutekunst, C F. et al. Respiratory infection in Marine recruits: influence of personal characteristics. Amer. Rev. Resp. Dis. 98 (1968), 800.
50. Zola, I. K. Culture and symptoms- an analysis of patient's presenting complaints. Amer. Socio!. Rev. 31 (1966), 615.