- Email: [email protected]
Stress and psychosomatic illness
Jou,,,u, o,Psychosomum Printed in Great Bntain.
Research. Vol. 26, No.
I, pp.63-67.
1982.
0022.3999/82/010063-09$03.00/O c 1982. Pergamon Press Ltd.
STRESS AND PSYCHOSOMATIC
ILLNESS
CHASE PATTERSONKIMBALL* (Received16 April 1981) Abstract-Increasing attention has been addressed
to the relationship of stress and illness during the past several decades. Despite extensive and intensive research in this area leading to statistically significant correlations, the specificity of the relationship remains vague. Some of the difficulty evolves from the definition of stress and illness. What is stress? How is it defined? Can it be defined independent of an identifiable reaction or change in behavior? How is illness defined? How does stress as a phenomenon relate to its perception and cognition and on what levels? To what extent do the physiological-worded concepts of Selye and Levi relate to the social and psychological ones of Holmes, Rahe and others? What do we mean by psychosomatic illness? How specific is it in terms of discrete entities. If we are in agreement that there is something there-something relating to change in behavior, conceptualized in social, psychological and biological terms with an event occurring external or internal to the individual, what greater precision can we strive for in determining and identifying specificity and the particulate? Is it possible to define a new model of relationships that will assist in identifying, determining and intervening in stress-related illness situations?
THE PROBLEM DESPITE the number of new papers written over the last few years, the inquiry into the relationship between stress and the development of illness remains elusive. This in part results from an initial failure to define these terms adequately. As a result, the studies addressing this subject are at once too broad and too narrow. The broadness is directly related to the failure in defining stress. The narrowness relates to the vagueness of the term illness and its limitation to physical complaints that fit a biological disease classification. The latter omits a wider view of illness, as any change of behavior that results in compromised function. By thus limiting, less attention has been addressed to change in other behavior which may be an intermediary factor in a subsequent illness or specific pathophysiological process. The continued exclusion of altered social, psychological, and physiological (short of pathophysiologic) behavior from most studies has ignored potential data that would sharpen our focus in the specific area of the relationship between better defined terms-stress and illness.
TOWARD
DEFINITION
A renewed attempt to define stress is in order. Among the more recent stress and illness advocates, Hans Selye [l] identified stress as something that correlated with and was measurable in terms of a physiologic response, specifically in the adrenal cortical-pituitary axis. He interpreted this as a general response to an identifiable provoking external factor. The response, he saw as an initial defense mechanism and named it, the General Adaptational Syndrome. This was seen as a normal response when sufficient to handle the external factor and maintain a relatively intact and functioning organism. The physiological alteration returned to baseline once the This paper was delivered at the 5th Congress of the International College of Psychosomatic Medicine held in Jerusalem, Israel, in September 1979. *Division of Biological Sciences, University of Chicago, 950 E. 59th Street, Chicago, Illinois 60637, U.S.A. 63
64
CHASE PATTERSONKIMBALL
factor was controlled. Subsequently, Selye became intrigued with a continued response correlating with provocation by an external agent even after that agent was no longer present. The processes once set in motion did not stop. This resulted in altered and usually an irreversible altered physiological state. These processes, Selye called “diseases of adaptation”. Several concepts are of immediate interest in this formulation. First, is that while Selye was interested in external factors as provoking responses in Claude Bernard’s milieu interna [2], his major emphasis came to be on the continued response of the organism to an external factor, even though that external factor was no longer present. In other words, it was still operating in some way. The idea was not unfamiliar to immunological concepts developed as far back as the late 19th Century, i.e. that an external factor had in one way or another combined with or was similar enough to an internal factor that the latter in its own right became a stimulus keeping the first response going, even when the external factor was no longer present. Second, Selye suggested that for some reason there was a deficiency in Walter Cannon’s [3] feedback mechanism, resulting in a failure to modulate or turn off the provoked response. Third, these concepts arc similar to the reasoning and the vocabulary used in the developmental psychology of Sigmund Freud [4]. Freud suggested that an unresolved trauma occurring at a vulnerable time in the life of an organism was recorded as a conflict within the mind, subsequently repressed or less available to consciousness, but arousable under a new social situation reminiscent of the initial one. (Considerations one and three might be subsumed under the common denominator, memory. In some way the organism remembers and reinvokes a response learned earlier. The reason for a particular response or its failure may be, and has been, explained in terms of a given organism’s vulnerability based on either genetic or early epigenetic factors or both.) Thus from the outset, the definition of stress was contingent upon a (measurable) response of (and in) the organism. Since the response was measured in physiological or biological variables, the definition of stress has had a distinctly biological orientation. [This seems to be so whether or not the concept of internal stressors (psychological, immunological) is accepted.] Franz Alexander [5] similarly defined stress as relating to an event external to the organism but correlating with an identifiable measurable response in a vulnerable organ system. The organ system of response was conditioned in early development either genetically or epigenetically, but in either case, depended upon the symbolic nature of the external stimuli. (In other words, the external factor stimulated an unconscious unresolved memory event which was manifested in the physiologic response. Presumably, the vulnerable organ system was that which correlated physiologically with the developmental phase in which the memory event triggered by the symbolic external event was initially recorded.) The Cornell group of Harold Wolff and Stuart Wolf [6], working at the same time from a more phenomenological orientation, correlated identifiable and measurable physiologic events with social factors outside the organism. They were able to reproduce the physiologic event in simulated situations when the amnestic social factor was reintroduced. Subsequently, Thomas Holmes’ and Richard Rahe’s [7] work considerably altered the definition of stress. Rather than measuring an actual response to an event, they asked individuals to define the effect of an
Stress and psychosomatic
illness
65
event in terms of the amount of (re)adjustment it would take to maintain the status quo or to return to a state of equilibrium. This led to a listing of situations or events, ranked in order of severity based on a consensus in a social group. These were called “life crises”. The term “life crisis” was viewed as the experience induced by the event. The latter was viewed as potentially stressful. This definition does not (at least immediately) imply an identifiable and measurable biological response. [In fact, the choices as potential stressors offered to subjects engaged in the research were almost always social ones. What eventually was visualized by the researchers as a response was the amount of (re)adjustment necessary to handle the event. Adjustment, i.e. response, was rarely identified in measurable biological terms. Rather, response seems to be more identifiable in psychological and social terms, i.e. the amount of psychological and social alteration (accommodation), that would be necessary to handle the hypothetical stressful experience.] Thus, the identifiable and measurable biological response of the early researchers defining stress became the anticipated, hypothetical social or psychological response of the latter. It was an “as if” or fantasized response, what one thought a particular crisis would demand of them. (What was measurable and measured was essentially an anticipated psychological or social response. These were developed on the basis of a social consensus, i.e. ranked in order of the amount of readjustment a particular hypothetical event would require on the part of a defined population.) Thus, the definition of stress came to be defined in terms of the identification and amount of social (re)adjustment (measure) a hypothetical event would require. The scales from Holmes’ and his colleagues’ research subsequently were used to correlate the measure of identifiable and measured accumulated life crisis units with illness. It is imperative to recognize that the accumulated life crisis units were measured in terms of the amount of (re)adjustment a hypothetical event (i.e. before it happened) would require. The measure used for the hypothetical event was automatically used for the actual event, without either verification or assessment of whether there was any correlation between the measure of adjustment for the actual event and that hypothetically anticipated for the fantasized event. This is the glaring and grievous dichotomy in the research. It is an unwarranted assumption that there would be a de facto strong correlation between the anticipated and the actual. An attempt to verify the relationship between the anticipated and the actual needs to be undertaken. Were this done, it is possible that the present low level of correlation between life stress and illness for all individuals would result in a high correlation when anticipated and actual measures were close together and an extremely low correlation when the measures were greatly dichotomous. Ursin et al. [8], on the basis of their study of paratrooper trainees, have recently suggested that the concept of stress requires refinement. They suggest that activation per se is not sufficient to define stress. Rather it is a maladaptive activation, one that is less than (hypo-) or more than (hyper-) sufficient to handle the provoking external stimulus. While they adhere primarily to identifying activation (and consequently what is seen as stress) in measurable biological processes, they also examine the modifying variables affecting the relation of the provoking external stimulus and the response. They emphasize that what comes to be identified as stressful is idiographic, as do Gardner and Taylor [9]. This includes a host of variables including the perception and cognition of the provoking stimulus-the defences against
66
CHASEPATTERSONKIMBALI
these; the affects generated and the defences against these; memory and learning of what is stressful; memory and learning of what is useful in handling (coping with) and intactness of the individual; motivation; role stress; native intelligence identification; the relationship between achievement motivation and actual performance (feedback).
SUMMARY
OF STRESS
In summary, then, even as we begin to address the relationship between stress and illness, there are some intrinsic difficulties deriving from the definition, identification and conceptualization of stress and what is stressful. The studies considered herein for the most part address stress in terms of its response (Selye) or what would be likely to necessitate a response (Holmes, Rahe). Ursin et al. attempts to refine this a little by suggesting that what is stressful has to do with organism’s incapability of responding at an optimal level to that situation. In either case, we are left with the Berkeleyian dilemma [lo], of whether we can call something stressful if there is not an identifiable and measurable response in the perceiver. (Response depends upon a perceiver. Further, we might inquire whether an event is stressful if it is only anticipated as being stressful but is not actually experienced as stressful, in terms of a non-response. We are left with the question whether or not a failure to respond identifies a stress. The answer will be in terms of the consequences to the organism.) More seriously, the identification of what is stressful and what is stress in terms of response practically begs the question of what the relationship is between the two of them. It is simply reductionistic. This is especially so in terms of what has been identified, and what we will here consider as illness. A stress is something that causes a response. The response is identified as a change in the behavior of the organism as identified and measured on some level, most often biologically. The degree of stress or how stressful something is, is measured in the adequacy of the response to handle (cope with) that stress. If the response is optimal, i.e. if it is optimally sufficient in handling the stress, the response is soon over and the system(s) of response returns to baseline. When it is less or more than optimal, the return to baseline in the system(s) of response will be aberrant. These considerations add other connotations to the idea of stress and what is stressful; e.g. direction, system(s) of, and duration of response. Various workers refine the definition of stress in terms of the less than, more than, prolonged and/or otherwise inadequacy of the response to handle what is stressful. These refinements add other problems in considering a proper definition of stress inasmuch as they focus increasingly on the response. The response (or lack of response) becomes increasingly dependent for its definition on the individual organism or what internal mechanisms it has which affects its means of coping, the idiographic ones. This property of of stress resides within, rather than outside, the individual. They may be both quantitatively and qualitatively assumed under the general designation of coping processes, those internal variables that determine and tune responses. Perhaps, the study of stress and what is stressful is best attended to in the particulate by focusing on these determinants and modifiers of response intrinsic to the organism. The languages subserving the identification and description of these modifiers include
Stress and psychosomatic illness
67
social, biological and psychological ones. This is the second part of this paper and reflects the major part of our preoccupation in psychosomatic studies. However, processes are not the focus of our concern in this discussion. Having discussed stress and having identified that to a very large extent, its definition is determined by the identification of response, we must address response and its relationship to the second part of our title illness . . . all the time keeping in mind that since by definition stress and response merge, therefore assuredly stress and illness (as one kind of response) will also. ILLNESS
In opening a discussion of illness, we will arbitrarily define illness as a “poor” (perhaps an untoward) response, one that results in an identifiable and measurable change from that which is considered optimal in an organism’s total functioning. (The clause, “from what is considered optimal”, we agree, is fraught with difficulty because it involves a certain amount of subjectivity on both the part of the organism as observer as well as the outside observer. The word “total” is used merely to emphasize that functioning is inevitably defined at different levels in different languages. To address only one is to ignore major areas of research.) For our purpose, we shall address illness response in terms of the social, psychological and the biological. This is a crucial dimension of our discussion. Few stress researchers have undertaken detailed and careful investigation in each of these areas. Without addressing each, it is difficult to gain adequate or complete understanding of the other and/or the effect of change in one on change in the other. [It is important to focus on processes rather than fixed states in our initial discussion of illness. For instance, an alteration in social behavior viewed as a response to a stress may be more damatic than an alteration in psychological processes or in biological responses. It may forever be co-related with these and depend on these for its manifestations inasmuch as it is psychologically and biologically mediated. On the other hand, the social response may modify the degree of identifiable and measurable response observed in the psychological and/or biological. A similar statement could be made for the psychological and biological in turn. Simultaneously, however, the effect of a social change will in and of itself affect psychological and biological behavior secondarily. Thus the behavior we observe on any one of these levels (e.g. biological) will include the response (R,) as primary to what is stressful and response (RJ as secondary to the responses observed in the other two responding systems (e.g. psychological and social). The severity of illness in any one of these responsive systems will be adjudged by the identification and measurement of the deviation of its chronicity from a baseline state. This should allow for the introduction of a behavior not initially observed in the baseline state, as well as for an alteration of behavior observed in the baseline state.] Succinctly, without considering the other two, we shall not be able to understand their role in modifying the observed behavior in the third. The extent of illness is defined by the amount of, direction of, and chronicity of the change from baseline as determined for each of the responsive systems. This definition depends not only on the organism reacting, but the observer of the reacting organism. At one time or another, the response may be more in one system than another. When in one system, it may obscure or even preclude major change in one or both
68
CHASE
PATTERSON KIMBALL
of the others. (Depending on other circumstances, idiographic to the organisms, e.g. past experience, response is selectively and primarily channelled through one system or another. Thus, we can suggest that response under or above a defined level of optimal change is defined as illness in a given system.) Hence, we have social illness, psychological illness, and biological illness. These illnesses relate to the extent to which measurable indices deviate from the norm. The chronicity of the illness is defined in terms of the fixedness of the behavior, i.e. its non-reversibility. It is at the point where response results in permanent changes, based on altered structure or fixed patterns of response that we begin to talk about diseases and disease states. (Again, an illness or disease seen primarily in terms of the extent and degree of the alteration in social behavior may also effect alterations in psychological and biological processes. When one or the other of these becomes greater in terms of change from baseline than the other, it becomes defined in terms of the illness or disease of the second.)
PSYCHOSOMATIC
ILLNESS
The idea of social and psychological illnesses as distinct from biological ones is especially difficult. First, we wish to distinguish these from the illness behavior which is associated with biological disease. There are patterns of social behavior which can be discerned as illness in and of themselves. (One example is drug usage and abuse, once we exclude the biological addictive properties, as Patrick Hughes [l l] has suggested, in his application of an epidemiologic model describing its spread as similar to that of infectious disease. A lesser, but more common example, is loneliness, a major factor among the elderly visiting medical clinics.) Psychological problems as distinct from psychiatric disease may be seen as examples of behavior unrelated to biological pathology. Such forms of behavior may be observed as relating to stress. As such, they may in some way protect against altered biological function and structure. The failure of stress researchers to focus on changes in behavior other than biological ones neglects a rich area of investigation for understanding coping and adaptive processes. The Holmes and Rahe studies of identifying what is stressful in social changes provide a measure for this. However, their studies largely have ignored the next step, i.e. changes in social and psychological behavior, in their emphasis on identifying altered biological functions. In some studies, they have identified possible processes as end-points, while they may be means of coping. Engel and Schmale have proposed one such psychosocial intermediary pathway for reaction to a social situation identified as stressful. This is the GivingUp-Given-Up state [ 121. This is viewed as putting the stressed individual at risk for a subsequent pathophysiological process, and is composed of identifiable and measurable psychological and social variables. It is probable that on a subtle level altered physiological function would be identifiable. This intermediary step moves us more quickly into the idea of psychosomatic illness, whose definition we need to more fully examine. These difficult abstractions lead us directly into the most difficult area, that of psychosomatic illness and disease. The term itself suggests relationships, especially that of causality. Some workers, depending on their focus of examination attempt
Stress and psychosomatic
illness
69
to turn this around, to somatopsychic, in order to imply consequential causality. Others have used such formulations as psychosociosomatic or biopsychosocial in order to emphasize the relationship of the environments in which patients become ill and are treated. As general concepts, these formulations have been useful in identifying for a biological medicine that there are other approaches that are useful in approaching illness and explaining such questions as: “why, at this time, in this place, has (tuberculosis) occurred in this individual.” At best, these are verbal accumulations that sensitize us to thinking in a contextual way about illness. Too often, the emphasis is skewed toward one of causative relationships. Elsewhere, we have dwelled upon the difficulties that result from this preoccupation, if only because our descriptions are frequently of the same thing in different languages [13-181. Therefore, a correlation between specific elements of one langauge with another is to be expected. This does not imply causality, necessarily. Rather, it more often implies concurrence inasmuch as the psychological, biological and socially conceptualized variables are occurring at the same time in the life of the individual. These integrative models tend to imply a sequential linearity as opposed to a contemporary reciprocal one; in other words, a social event is seen as influencing a psychological reaction that results in biological change, rather than a more dynamic back and forth interplay or relationships between the three. The concept of modifying feedback influences of one on the manifestation of the other(s) is submerged. Historically, the term psychosomatic medicine has emphasized one kind of specificity or another. H. Flanders Dunbar’s [19] focus was on personality and disease, not dissimilar from Meyer Friedman’s [20] current emphasis on behavior pattern type A and coronary artery disease. Franz Alexander [5] emphasized the specificity of the conflict and the organ system of response. The illnesses which these investigators chose to study inadvertently became the psychosomatic diseases. So identified, a generation of physicians came to believe that primarily psychological factors caused the somatic pathology and, therefore, that a psychotherapeutic approach was the mode of approach. Needless to say, in a pragmatic medicine, a failure to greatly ameliorate or cure based on this reasoning, led to a reactive disbelief and disregard for these formulations. More important, it would seem that terms such as psychosomatic are misleading. They imply and promise more than they deliver. Rather they are symbols suggesting that an illness phenomenon needs to be approached by several methods of investigation. An attempt to draw temporal and causal relationships will only sometimes succeed and will depend on a host of more subtle factors unique to the individual’s biological, social and psychological development. These cannot be identified in the broad generalities but can only be examined in the particulate which will yield specific patterns for an individual and perhaps to a subset of vulnerable individuals as Herbert Weiner emphasizes in his valuable book, Psychobiology, and other writings [21,22]. There are other difficulties with the term psychosomatic, than the inference that some illness is uni-dimensional. The concept has focused on the individual rather than on the patient as part of an environment in which the individual has developed, experienced and been treated for illness. Only more recently have we come to examine the environments in which illness is first experienced and is subsequently treated and the relationships therein. The milieu and the factors in that milieu are
70
CHASE PATTERSON KIMBALI
very much part of the illness study. It is a milieu in which the individual is a part, his illness and his treatment are a part. There is an essential complementarity between his illness and the significant others that define his illness. A modern trend in the area of psychosomatic is the exploration of these environments, placing increasing emphasis on a social analysis of the environment in which the individual lives, looking for the kinds of events that modify behavior and what the consequences of this modification imply. Understanding behavior requires study of the environment as well as of the individual, using different languages of exploration, social, psychological, biological and their subsets. Illness, which may be viewed as a lesser or greater response than is optimal to cope with events threatening to the stability of the milieu interna requires examination of the psychological and biological as well as the social if we would approach it optimally and completely. This, I believe, is the appropriate definition of psychosomatic. The shift of psychosomatic medicine from a disease orientation in the individual whether it be biological or psychological to a person-environment one is heralded by our changing language, one which attempts the synthesis of phenomenological states by incorporating elements derivative of the different languages that frequently coexist. For example, Meyer Friedman’s original identification of Behavior Pattern Type A included both behavioral and biological variables. George Engel’s and Arthur Schmales Giving-Up-Given-Up state includes social, psychological and biological variables. Grief is another such state. The phenomenon of conversion, another. In medicine, most illness states are labelled on the basis of their biological definitions. What remains to be more specifically identified for these is their social and psychological definitions. These exist in their own right without an obligatory or necessary correlation of elements with each other. Even when such correlations are obvious, the relationship does not necessarily imply a linear or causal one. Most often, perhaps they are only describing the same thing by different languages of analysis. As this deliberate work progresses, we can anticipate the identification of increasingly specific patterns of relationships between social, psychological and biological variables. To get at these, our work will need to address the more difficult and subtle area of coping and adaptational processes, also in terms of social, biological and psychological constructs. As this unfolds, we will define more precisely what is stressful, what is stress, and what the relationships are between these and psychological, biological and social changes.
REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
SELYE H. Sfress in Health and Disease. Butterworths, Boston (1976). BERNARD C. Introduction to the Study of ExperimentalMedicine. Dover Press (1957). CANNON W. Wisdom of theBody. W. W. Norton, New York (1963). FREUD S. CollectedPapers of Sigmund Freud, Vol. V (Edited by STRACHEY _I.), Ch. II, pp. 31-32. Basic Books, Neti York (1959). ALEXANDER F. PsychosomaticMedicine. W. W. Norton, New York (1950). WOLF S. and GOODALL H. Behavioral Science in Clinical Medicine. Charles C. Thomas, Springfield (1976). HOLMEST. and RAHE R. The Social Readjustment Scale. J. psychosom. Res. 11,213 (1967). URSIN H., BOADE E. and LEVINE S. Psychobiology of Stress: A Study of Coping Man. Academic Press, New York (1978). GARDNER A. W. and TAYLOR P. Health at Work. Halsted Press, New York (1975). BERKELEY G. The Works of George Berkeley, Bishop of Cloyne 1685-1753. J. Exshaw, Dublin (1784).
Stress and psychosomatic
11. HUGHES P.,
SENAY
E. and PARKER R. The medical
71
illness
management
of a heroin
epidemic.
Archs
gen.
Psychiat. 21,585 (1972). 12. ENGEL 0. L. and SCHMALEA. H. Psychoanalytic theory of somatic disorder: conversion, specificity and the disease-onset situation. J. Am. psychoanalyt. ass. 15,344(1967). 13. KIMBALLC. P. Conceutual developments in psychosomatic medicine 1939-1969. Ann. Int. Med. 73, . . 307 (1970). 14. KIMBALLC. P. The languages of psychosomatic medicine. Psychotherpy and Psychosotnatics, Vol. 28, Nos. l-4. Karger, Base1 (1977). 15. IMBALL C. P. Changing concepts in psychosomatic theory-I. Historical analysis. Psychotherapy andPsychosomatics, Vol. 31, pp. l-8. S. Karger, Base1 (1979). 16. KIMBALLC. P. Liaison psychiatry: the extension of a concept. Psychosomatics 21, 684-687 (1980). 17. KIMBALLC. P. The experience of open heart surgery. VI. Research and consultation psychiatry. In Psychic and Neurological Dysfunctions After Open Heart Surgery, pp. 215-228. Georg Thieme, Stuttgart (1980). 18. KIMBALL C. P. Liaison psychitry as a general systems approach to behavior. Psychotherapy and Psychosomatics, Vol. 32, pp. 134-147. S. Karger, Base1 (1979). 19. DUNBAR H. F. Emotions and Bodily Changes: A Survey of Literature on Psychomatic Interrelationships. Columbia University Press, New York (1954). 20. FRIEDMAN M. and ROSENMAN H. Overt behavior pattern in coronary disease: detection of overt behavior pattern A in patients with coronary disease by a new psychophysiological procedure.
J.A.M.A. 173,1320(1960). 21. WEINER H. Psychobiology and Human Disease. Elsevier North-Holland, 22. WEINER H. The illusion of simplicity: (Supplement) (July, 1978).
the medical
model
revisited.
New York (1977).
Am. J. Psychiat. 135, 27
Research. Vol. 26, No.
I, pp.63-67.
1982.
0022.3999/82/010063-09$03.00/O c 1982. Pergamon Press Ltd.
STRESS AND PSYCHOSOMATIC
ILLNESS
CHASE PATTERSONKIMBALL* (Received16 April 1981) Abstract-Increasing attention has been addressed
to the relationship of stress and illness during the past several decades. Despite extensive and intensive research in this area leading to statistically significant correlations, the specificity of the relationship remains vague. Some of the difficulty evolves from the definition of stress and illness. What is stress? How is it defined? Can it be defined independent of an identifiable reaction or change in behavior? How is illness defined? How does stress as a phenomenon relate to its perception and cognition and on what levels? To what extent do the physiological-worded concepts of Selye and Levi relate to the social and psychological ones of Holmes, Rahe and others? What do we mean by psychosomatic illness? How specific is it in terms of discrete entities. If we are in agreement that there is something there-something relating to change in behavior, conceptualized in social, psychological and biological terms with an event occurring external or internal to the individual, what greater precision can we strive for in determining and identifying specificity and the particulate? Is it possible to define a new model of relationships that will assist in identifying, determining and intervening in stress-related illness situations?
THE PROBLEM DESPITE the number of new papers written over the last few years, the inquiry into the relationship between stress and the development of illness remains elusive. This in part results from an initial failure to define these terms adequately. As a result, the studies addressing this subject are at once too broad and too narrow. The broadness is directly related to the failure in defining stress. The narrowness relates to the vagueness of the term illness and its limitation to physical complaints that fit a biological disease classification. The latter omits a wider view of illness, as any change of behavior that results in compromised function. By thus limiting, less attention has been addressed to change in other behavior which may be an intermediary factor in a subsequent illness or specific pathophysiological process. The continued exclusion of altered social, psychological, and physiological (short of pathophysiologic) behavior from most studies has ignored potential data that would sharpen our focus in the specific area of the relationship between better defined terms-stress and illness.
TOWARD
DEFINITION
A renewed attempt to define stress is in order. Among the more recent stress and illness advocates, Hans Selye [l] identified stress as something that correlated with and was measurable in terms of a physiologic response, specifically in the adrenal cortical-pituitary axis. He interpreted this as a general response to an identifiable provoking external factor. The response, he saw as an initial defense mechanism and named it, the General Adaptational Syndrome. This was seen as a normal response when sufficient to handle the external factor and maintain a relatively intact and functioning organism. The physiological alteration returned to baseline once the This paper was delivered at the 5th Congress of the International College of Psychosomatic Medicine held in Jerusalem, Israel, in September 1979. *Division of Biological Sciences, University of Chicago, 950 E. 59th Street, Chicago, Illinois 60637, U.S.A. 63
64
CHASE PATTERSONKIMBALL
factor was controlled. Subsequently, Selye became intrigued with a continued response correlating with provocation by an external agent even after that agent was no longer present. The processes once set in motion did not stop. This resulted in altered and usually an irreversible altered physiological state. These processes, Selye called “diseases of adaptation”. Several concepts are of immediate interest in this formulation. First, is that while Selye was interested in external factors as provoking responses in Claude Bernard’s milieu interna [2], his major emphasis came to be on the continued response of the organism to an external factor, even though that external factor was no longer present. In other words, it was still operating in some way. The idea was not unfamiliar to immunological concepts developed as far back as the late 19th Century, i.e. that an external factor had in one way or another combined with or was similar enough to an internal factor that the latter in its own right became a stimulus keeping the first response going, even when the external factor was no longer present. Second, Selye suggested that for some reason there was a deficiency in Walter Cannon’s [3] feedback mechanism, resulting in a failure to modulate or turn off the provoked response. Third, these concepts arc similar to the reasoning and the vocabulary used in the developmental psychology of Sigmund Freud [4]. Freud suggested that an unresolved trauma occurring at a vulnerable time in the life of an organism was recorded as a conflict within the mind, subsequently repressed or less available to consciousness, but arousable under a new social situation reminiscent of the initial one. (Considerations one and three might be subsumed under the common denominator, memory. In some way the organism remembers and reinvokes a response learned earlier. The reason for a particular response or its failure may be, and has been, explained in terms of a given organism’s vulnerability based on either genetic or early epigenetic factors or both.) Thus from the outset, the definition of stress was contingent upon a (measurable) response of (and in) the organism. Since the response was measured in physiological or biological variables, the definition of stress has had a distinctly biological orientation. [This seems to be so whether or not the concept of internal stressors (psychological, immunological) is accepted.] Franz Alexander [5] similarly defined stress as relating to an event external to the organism but correlating with an identifiable measurable response in a vulnerable organ system. The organ system of response was conditioned in early development either genetically or epigenetically, but in either case, depended upon the symbolic nature of the external stimuli. (In other words, the external factor stimulated an unconscious unresolved memory event which was manifested in the physiologic response. Presumably, the vulnerable organ system was that which correlated physiologically with the developmental phase in which the memory event triggered by the symbolic external event was initially recorded.) The Cornell group of Harold Wolff and Stuart Wolf [6], working at the same time from a more phenomenological orientation, correlated identifiable and measurable physiologic events with social factors outside the organism. They were able to reproduce the physiologic event in simulated situations when the amnestic social factor was reintroduced. Subsequently, Thomas Holmes’ and Richard Rahe’s [7] work considerably altered the definition of stress. Rather than measuring an actual response to an event, they asked individuals to define the effect of an
Stress and psychosomatic
illness
65
event in terms of the amount of (re)adjustment it would take to maintain the status quo or to return to a state of equilibrium. This led to a listing of situations or events, ranked in order of severity based on a consensus in a social group. These were called “life crises”. The term “life crisis” was viewed as the experience induced by the event. The latter was viewed as potentially stressful. This definition does not (at least immediately) imply an identifiable and measurable biological response. [In fact, the choices as potential stressors offered to subjects engaged in the research were almost always social ones. What eventually was visualized by the researchers as a response was the amount of (re)adjustment necessary to handle the event. Adjustment, i.e. response, was rarely identified in measurable biological terms. Rather, response seems to be more identifiable in psychological and social terms, i.e. the amount of psychological and social alteration (accommodation), that would be necessary to handle the hypothetical stressful experience.] Thus, the identifiable and measurable biological response of the early researchers defining stress became the anticipated, hypothetical social or psychological response of the latter. It was an “as if” or fantasized response, what one thought a particular crisis would demand of them. (What was measurable and measured was essentially an anticipated psychological or social response. These were developed on the basis of a social consensus, i.e. ranked in order of the amount of readjustment a particular hypothetical event would require on the part of a defined population.) Thus, the definition of stress came to be defined in terms of the identification and amount of social (re)adjustment (measure) a hypothetical event would require. The scales from Holmes’ and his colleagues’ research subsequently were used to correlate the measure of identifiable and measured accumulated life crisis units with illness. It is imperative to recognize that the accumulated life crisis units were measured in terms of the amount of (re)adjustment a hypothetical event (i.e. before it happened) would require. The measure used for the hypothetical event was automatically used for the actual event, without either verification or assessment of whether there was any correlation between the measure of adjustment for the actual event and that hypothetically anticipated for the fantasized event. This is the glaring and grievous dichotomy in the research. It is an unwarranted assumption that there would be a de facto strong correlation between the anticipated and the actual. An attempt to verify the relationship between the anticipated and the actual needs to be undertaken. Were this done, it is possible that the present low level of correlation between life stress and illness for all individuals would result in a high correlation when anticipated and actual measures were close together and an extremely low correlation when the measures were greatly dichotomous. Ursin et al. [8], on the basis of their study of paratrooper trainees, have recently suggested that the concept of stress requires refinement. They suggest that activation per se is not sufficient to define stress. Rather it is a maladaptive activation, one that is less than (hypo-) or more than (hyper-) sufficient to handle the provoking external stimulus. While they adhere primarily to identifying activation (and consequently what is seen as stress) in measurable biological processes, they also examine the modifying variables affecting the relation of the provoking external stimulus and the response. They emphasize that what comes to be identified as stressful is idiographic, as do Gardner and Taylor [9]. This includes a host of variables including the perception and cognition of the provoking stimulus-the defences against
66
CHASEPATTERSONKIMBALI
these; the affects generated and the defences against these; memory and learning of what is stressful; memory and learning of what is useful in handling (coping with) and intactness of the individual; motivation; role stress; native intelligence identification; the relationship between achievement motivation and actual performance (feedback).
SUMMARY
OF STRESS
In summary, then, even as we begin to address the relationship between stress and illness, there are some intrinsic difficulties deriving from the definition, identification and conceptualization of stress and what is stressful. The studies considered herein for the most part address stress in terms of its response (Selye) or what would be likely to necessitate a response (Holmes, Rahe). Ursin et al. attempts to refine this a little by suggesting that what is stressful has to do with organism’s incapability of responding at an optimal level to that situation. In either case, we are left with the Berkeleyian dilemma [lo], of whether we can call something stressful if there is not an identifiable and measurable response in the perceiver. (Response depends upon a perceiver. Further, we might inquire whether an event is stressful if it is only anticipated as being stressful but is not actually experienced as stressful, in terms of a non-response. We are left with the question whether or not a failure to respond identifies a stress. The answer will be in terms of the consequences to the organism.) More seriously, the identification of what is stressful and what is stress in terms of response practically begs the question of what the relationship is between the two of them. It is simply reductionistic. This is especially so in terms of what has been identified, and what we will here consider as illness. A stress is something that causes a response. The response is identified as a change in the behavior of the organism as identified and measured on some level, most often biologically. The degree of stress or how stressful something is, is measured in the adequacy of the response to handle (cope with) that stress. If the response is optimal, i.e. if it is optimally sufficient in handling the stress, the response is soon over and the system(s) of response returns to baseline. When it is less or more than optimal, the return to baseline in the system(s) of response will be aberrant. These considerations add other connotations to the idea of stress and what is stressful; e.g. direction, system(s) of, and duration of response. Various workers refine the definition of stress in terms of the less than, more than, prolonged and/or otherwise inadequacy of the response to handle what is stressful. These refinements add other problems in considering a proper definition of stress inasmuch as they focus increasingly on the response. The response (or lack of response) becomes increasingly dependent for its definition on the individual organism or what internal mechanisms it has which affects its means of coping, the idiographic ones. This property of of stress resides within, rather than outside, the individual. They may be both quantitatively and qualitatively assumed under the general designation of coping processes, those internal variables that determine and tune responses. Perhaps, the study of stress and what is stressful is best attended to in the particulate by focusing on these determinants and modifiers of response intrinsic to the organism. The languages subserving the identification and description of these modifiers include
Stress and psychosomatic illness
67
social, biological and psychological ones. This is the second part of this paper and reflects the major part of our preoccupation in psychosomatic studies. However, processes are not the focus of our concern in this discussion. Having discussed stress and having identified that to a very large extent, its definition is determined by the identification of response, we must address response and its relationship to the second part of our title illness . . . all the time keeping in mind that since by definition stress and response merge, therefore assuredly stress and illness (as one kind of response) will also. ILLNESS
In opening a discussion of illness, we will arbitrarily define illness as a “poor” (perhaps an untoward) response, one that results in an identifiable and measurable change from that which is considered optimal in an organism’s total functioning. (The clause, “from what is considered optimal”, we agree, is fraught with difficulty because it involves a certain amount of subjectivity on both the part of the organism as observer as well as the outside observer. The word “total” is used merely to emphasize that functioning is inevitably defined at different levels in different languages. To address only one is to ignore major areas of research.) For our purpose, we shall address illness response in terms of the social, psychological and the biological. This is a crucial dimension of our discussion. Few stress researchers have undertaken detailed and careful investigation in each of these areas. Without addressing each, it is difficult to gain adequate or complete understanding of the other and/or the effect of change in one on change in the other. [It is important to focus on processes rather than fixed states in our initial discussion of illness. For instance, an alteration in social behavior viewed as a response to a stress may be more damatic than an alteration in psychological processes or in biological responses. It may forever be co-related with these and depend on these for its manifestations inasmuch as it is psychologically and biologically mediated. On the other hand, the social response may modify the degree of identifiable and measurable response observed in the psychological and/or biological. A similar statement could be made for the psychological and biological in turn. Simultaneously, however, the effect of a social change will in and of itself affect psychological and biological behavior secondarily. Thus the behavior we observe on any one of these levels (e.g. biological) will include the response (R,) as primary to what is stressful and response (RJ as secondary to the responses observed in the other two responding systems (e.g. psychological and social). The severity of illness in any one of these responsive systems will be adjudged by the identification and measurement of the deviation of its chronicity from a baseline state. This should allow for the introduction of a behavior not initially observed in the baseline state, as well as for an alteration of behavior observed in the baseline state.] Succinctly, without considering the other two, we shall not be able to understand their role in modifying the observed behavior in the third. The extent of illness is defined by the amount of, direction of, and chronicity of the change from baseline as determined for each of the responsive systems. This definition depends not only on the organism reacting, but the observer of the reacting organism. At one time or another, the response may be more in one system than another. When in one system, it may obscure or even preclude major change in one or both
68
CHASE
PATTERSON KIMBALL
of the others. (Depending on other circumstances, idiographic to the organisms, e.g. past experience, response is selectively and primarily channelled through one system or another. Thus, we can suggest that response under or above a defined level of optimal change is defined as illness in a given system.) Hence, we have social illness, psychological illness, and biological illness. These illnesses relate to the extent to which measurable indices deviate from the norm. The chronicity of the illness is defined in terms of the fixedness of the behavior, i.e. its non-reversibility. It is at the point where response results in permanent changes, based on altered structure or fixed patterns of response that we begin to talk about diseases and disease states. (Again, an illness or disease seen primarily in terms of the extent and degree of the alteration in social behavior may also effect alterations in psychological and biological processes. When one or the other of these becomes greater in terms of change from baseline than the other, it becomes defined in terms of the illness or disease of the second.)
PSYCHOSOMATIC
ILLNESS
The idea of social and psychological illnesses as distinct from biological ones is especially difficult. First, we wish to distinguish these from the illness behavior which is associated with biological disease. There are patterns of social behavior which can be discerned as illness in and of themselves. (One example is drug usage and abuse, once we exclude the biological addictive properties, as Patrick Hughes [l l] has suggested, in his application of an epidemiologic model describing its spread as similar to that of infectious disease. A lesser, but more common example, is loneliness, a major factor among the elderly visiting medical clinics.) Psychological problems as distinct from psychiatric disease may be seen as examples of behavior unrelated to biological pathology. Such forms of behavior may be observed as relating to stress. As such, they may in some way protect against altered biological function and structure. The failure of stress researchers to focus on changes in behavior other than biological ones neglects a rich area of investigation for understanding coping and adaptive processes. The Holmes and Rahe studies of identifying what is stressful in social changes provide a measure for this. However, their studies largely have ignored the next step, i.e. changes in social and psychological behavior, in their emphasis on identifying altered biological functions. In some studies, they have identified possible processes as end-points, while they may be means of coping. Engel and Schmale have proposed one such psychosocial intermediary pathway for reaction to a social situation identified as stressful. This is the GivingUp-Given-Up state [ 121. This is viewed as putting the stressed individual at risk for a subsequent pathophysiological process, and is composed of identifiable and measurable psychological and social variables. It is probable that on a subtle level altered physiological function would be identifiable. This intermediary step moves us more quickly into the idea of psychosomatic illness, whose definition we need to more fully examine. These difficult abstractions lead us directly into the most difficult area, that of psychosomatic illness and disease. The term itself suggests relationships, especially that of causality. Some workers, depending on their focus of examination attempt
Stress and psychosomatic
illness
69
to turn this around, to somatopsychic, in order to imply consequential causality. Others have used such formulations as psychosociosomatic or biopsychosocial in order to emphasize the relationship of the environments in which patients become ill and are treated. As general concepts, these formulations have been useful in identifying for a biological medicine that there are other approaches that are useful in approaching illness and explaining such questions as: “why, at this time, in this place, has (tuberculosis) occurred in this individual.” At best, these are verbal accumulations that sensitize us to thinking in a contextual way about illness. Too often, the emphasis is skewed toward one of causative relationships. Elsewhere, we have dwelled upon the difficulties that result from this preoccupation, if only because our descriptions are frequently of the same thing in different languages [13-181. Therefore, a correlation between specific elements of one langauge with another is to be expected. This does not imply causality, necessarily. Rather, it more often implies concurrence inasmuch as the psychological, biological and socially conceptualized variables are occurring at the same time in the life of the individual. These integrative models tend to imply a sequential linearity as opposed to a contemporary reciprocal one; in other words, a social event is seen as influencing a psychological reaction that results in biological change, rather than a more dynamic back and forth interplay or relationships between the three. The concept of modifying feedback influences of one on the manifestation of the other(s) is submerged. Historically, the term psychosomatic medicine has emphasized one kind of specificity or another. H. Flanders Dunbar’s [19] focus was on personality and disease, not dissimilar from Meyer Friedman’s [20] current emphasis on behavior pattern type A and coronary artery disease. Franz Alexander [5] emphasized the specificity of the conflict and the organ system of response. The illnesses which these investigators chose to study inadvertently became the psychosomatic diseases. So identified, a generation of physicians came to believe that primarily psychological factors caused the somatic pathology and, therefore, that a psychotherapeutic approach was the mode of approach. Needless to say, in a pragmatic medicine, a failure to greatly ameliorate or cure based on this reasoning, led to a reactive disbelief and disregard for these formulations. More important, it would seem that terms such as psychosomatic are misleading. They imply and promise more than they deliver. Rather they are symbols suggesting that an illness phenomenon needs to be approached by several methods of investigation. An attempt to draw temporal and causal relationships will only sometimes succeed and will depend on a host of more subtle factors unique to the individual’s biological, social and psychological development. These cannot be identified in the broad generalities but can only be examined in the particulate which will yield specific patterns for an individual and perhaps to a subset of vulnerable individuals as Herbert Weiner emphasizes in his valuable book, Psychobiology, and other writings [21,22]. There are other difficulties with the term psychosomatic, than the inference that some illness is uni-dimensional. The concept has focused on the individual rather than on the patient as part of an environment in which the individual has developed, experienced and been treated for illness. Only more recently have we come to examine the environments in which illness is first experienced and is subsequently treated and the relationships therein. The milieu and the factors in that milieu are
70
CHASE PATTERSON KIMBALI
very much part of the illness study. It is a milieu in which the individual is a part, his illness and his treatment are a part. There is an essential complementarity between his illness and the significant others that define his illness. A modern trend in the area of psychosomatic is the exploration of these environments, placing increasing emphasis on a social analysis of the environment in which the individual lives, looking for the kinds of events that modify behavior and what the consequences of this modification imply. Understanding behavior requires study of the environment as well as of the individual, using different languages of exploration, social, psychological, biological and their subsets. Illness, which may be viewed as a lesser or greater response than is optimal to cope with events threatening to the stability of the milieu interna requires examination of the psychological and biological as well as the social if we would approach it optimally and completely. This, I believe, is the appropriate definition of psychosomatic. The shift of psychosomatic medicine from a disease orientation in the individual whether it be biological or psychological to a person-environment one is heralded by our changing language, one which attempts the synthesis of phenomenological states by incorporating elements derivative of the different languages that frequently coexist. For example, Meyer Friedman’s original identification of Behavior Pattern Type A included both behavioral and biological variables. George Engel’s and Arthur Schmales Giving-Up-Given-Up state includes social, psychological and biological variables. Grief is another such state. The phenomenon of conversion, another. In medicine, most illness states are labelled on the basis of their biological definitions. What remains to be more specifically identified for these is their social and psychological definitions. These exist in their own right without an obligatory or necessary correlation of elements with each other. Even when such correlations are obvious, the relationship does not necessarily imply a linear or causal one. Most often, perhaps they are only describing the same thing by different languages of analysis. As this deliberate work progresses, we can anticipate the identification of increasingly specific patterns of relationships between social, psychological and biological variables. To get at these, our work will need to address the more difficult and subtle area of coping and adaptational processes, also in terms of social, biological and psychological constructs. As this unfolds, we will define more precisely what is stressful, what is stress, and what the relationships are between these and psychological, biological and social changes.
REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
SELYE H. Sfress in Health and Disease. Butterworths, Boston (1976). BERNARD C. Introduction to the Study of ExperimentalMedicine. Dover Press (1957). CANNON W. Wisdom of theBody. W. W. Norton, New York (1963). FREUD S. CollectedPapers of Sigmund Freud, Vol. V (Edited by STRACHEY _I.), Ch. II, pp. 31-32. Basic Books, Neti York (1959). ALEXANDER F. PsychosomaticMedicine. W. W. Norton, New York (1950). WOLF S. and GOODALL H. Behavioral Science in Clinical Medicine. Charles C. Thomas, Springfield (1976). HOLMEST. and RAHE R. The Social Readjustment Scale. J. psychosom. Res. 11,213 (1967). URSIN H., BOADE E. and LEVINE S. Psychobiology of Stress: A Study of Coping Man. Academic Press, New York (1978). GARDNER A. W. and TAYLOR P. Health at Work. Halsted Press, New York (1975). BERKELEY G. The Works of George Berkeley, Bishop of Cloyne 1685-1753. J. Exshaw, Dublin (1784).
Stress and psychosomatic
11. HUGHES P.,
SENAY
E. and PARKER R. The medical
71
illness
management
of a heroin
epidemic.
Archs
gen.
Psychiat. 21,585 (1972). 12. ENGEL 0. L. and SCHMALEA. H. Psychoanalytic theory of somatic disorder: conversion, specificity and the disease-onset situation. J. Am. psychoanalyt. ass. 15,344(1967). 13. KIMBALLC. P. Conceutual developments in psychosomatic medicine 1939-1969. Ann. Int. Med. 73, . . 307 (1970). 14. KIMBALLC. P. The languages of psychosomatic medicine. Psychotherpy and Psychosotnatics, Vol. 28, Nos. l-4. Karger, Base1 (1977). 15. IMBALL C. P. Changing concepts in psychosomatic theory-I. Historical analysis. Psychotherapy andPsychosomatics, Vol. 31, pp. l-8. S. Karger, Base1 (1979). 16. KIMBALLC. P. Liaison psychiatry: the extension of a concept. Psychosomatics 21, 684-687 (1980). 17. KIMBALLC. P. The experience of open heart surgery. VI. Research and consultation psychiatry. In Psychic and Neurological Dysfunctions After Open Heart Surgery, pp. 215-228. Georg Thieme, Stuttgart (1980). 18. KIMBALL C. P. Liaison psychitry as a general systems approach to behavior. Psychotherapy and Psychosomatics, Vol. 32, pp. 134-147. S. Karger, Base1 (1979). 19. DUNBAR H. F. Emotions and Bodily Changes: A Survey of Literature on Psychomatic Interrelationships. Columbia University Press, New York (1954). 20. FRIEDMAN M. and ROSENMAN H. Overt behavior pattern in coronary disease: detection of overt behavior pattern A in patients with coronary disease by a new psychophysiological procedure.
J.A.M.A. 173,1320(1960). 21. WEINER H. Psychobiology and Human Disease. Elsevier North-Holland, 22. WEINER H. The illusion of simplicity: (Supplement) (July, 1978).
the medical
model
revisited.
New York (1977).
Am. J. Psychiat. 135, 27