Review by Neil Vickers

Nassir Ghaemi’s book The Rise and Fall of the Biopsychosocial Model (2009) is about the biopsychosocial model of health and disease as it is used in psychiatry. In Ghaemi’s view, at least in the English-speaking world, psychiatry has always been implicitly eclectic. Of course, some psychiatrists rejected eclecticism in favour of a single form of treatment such as psychoanalysis, or psychopharmacology, or electroconvulsive therapy. But psychiatry’s status as a medical specialty has always been predicated on the idea that it was broadly-based: it was not necessary to attach oneself to any particular school of thought about mental illness in order to use the title of psychiatrist. More positively, this diversity was often said to correspond to the peculiar characteristics of mental illnesses. On this view, there is no unified field theory that unites all the concepts and information needed in psychiatric practice because some mental illnesses are thought to stem from the patient’s outlook on his life, some the result of genes, and others still of the interaction of genes and experience. To complicate matters further, psychiatrists disagree amongst themselves about which disorders belong in which categories. Ghaemi sees the biopsychosocial model as a set of truisms that enables psychiatrists to ignore the significance of these disagreements. Everybody believes that biology, psychology and the social world have a bearing on mental health. The biopsychosocial model is a way of not mentioning the war.

Ghaemi is at his most persuasive when setting the biopsychosocial model in the context of earlier solutions to this problem which has dogged psychiatry more or less since its inception. Not only are there major disagreements about mental illnesses in general and in particular, there is no way of bridging the gaps that commands sufficiently wide support within psychiatry, other than by saying that everyone’s on to something. In Ghaemi’s narrative, the rot set in with Adolf Meyer’s psychobiology. Meyer believed that in psychiatry you have to treat the whole person. You took a detailed case history paying close attention to the biological, psychological, and social factors relevant to the patient’s symptoms. He was especially interested in the ways in which the patient’s social and environmental background might predispose him to react pathologically to certain life events. Most mental illnesses, in his view, were reactions to life events. Ghaemi points out that Meyer did not reject biology; but Meyer didn’t pursue the biological causes of mental illness through his own research, even though he encouraged others to do so. Phillip Slavney and Paul McHugh note that ‘for Meyer the person was a level of biological organisation just as important as the molecular or the cellular. Functioning at each level has its own rules and should be studied on its own terms.’ It should be borne in mind that in Meyer’s psychobiology, functioning at each level would be studied by distinct disciplines. Biologists and physiologists would study anatomical lesions and toxic factors, psychologists would study personality traits, cultural influences, environmental stressors and the like. In some respects Meyer was a forerunner of Karl Ludwig von Bertalanffy’s general systems theory (GST), which was very compatible with Meyer’s outlook. GST was the study of systems in general. Perhaps what distinguishes GST most from Meyer’s theory was a more flexible account of the subsystems that defined the whole. In Von Berlalanffy’s work, the hierarchical levels of functioning were not necessarily coterminous with specific academic disciplines. Nevertheless, Ghaemi is right to echo a point made by Paul McHugh that the biopsychosocial model is ‘Adolf Meyer’s concept of psychobiology renamed and reanimated for the contemporary era’ (2006). Roy Grinker and George Engel, the two men who Ghaemi thinks did most to establish and disseminate the biopsychosocial model in its current form, were strongly committed to Von Bertalanffy’s ideas about systems and both acknowledged Meyer’s foundational role. For Grinker, the psychiatrist and the psychoanalyst were trying to understand ‘the psycho-somatic-environmental systems as processes in transaction, within a particular universe or field.’ Engel too presented health and disease in terms of a hierarchy of systems culminating in the biosphere into which were subsumed society, culture, community, family, the person all the way down to sub-atomic particles by way of the nervous system, the vital organs and body tissue.

The biopsychosocial model was not devised in response to a problem in mainstream psychiatry at all. It came out of psychosomatic medicine, a discipline in which there was great openness to psychoanalytic ideas but which was not exclusively psychoanalytic. On page 53, Ghaemi concedes that ‘the main source of the BPS model is psychosomatic medicine, concerned with medical illnesses with psychological components’. He continues: ‘It has been little discussed whether a model derived from this small corner of the psychiatric profession is appropriate for the entire broad range of mental illnesses.’ I will come to the second point later on. The first warrants more detailed exploration than Ghaemi gives it. The biopsychosocial model was elaborated as a way of conceptualising the vicissitudes of internal experience of illness. Above all else, it was an attempt to study the ways in which the patient’s personhood contributes to the unfolding of pathological processes. It is a major weakness of Ghaemi’s book is that he refuses to engage with this idea. By turns he says it’s something doctors know already or he treats it as a form of eclecticism that puts medicine on the same level as psychology and sociology. But that is not how the founders, all of whom were medically qualified, saw it.

This brings me to an aside. On the subject of the founders, Ghaemi is eccentric, to put it mildly. He thinks the term ‘the biopsychosocial model’ ought to be credited to Roy Grinker because Grinker almost used the term in 1952. Moreover, Ghaemi insinuates that when he wrote his famous paper of 1977 Engel must have known about Grinker’s priority but chose not to acknowledge him. Ghaemi awards Grinker priority on the basis of a paper entitled ‘Training of a psychiatrist-psychoanalyst’ which was given as a presentation to the Chicago Psychoanalytic Institute but which wasn’t published until 1994. (Engel wasn’t present when it was first read out.) In this paper Grinker describes body, mind and society as a unity and invokes the ‘psycho-somatic-social’ as a near synonym. Here is the passage in full:

[The psychiatrist and the psychoanalyst] are practitioners in a field of behavior in which they try to understand the psycho-somatic-environmental systems as processes in transaction, within a particular universe or field. The psychiatrist or analyst is usually interested most intensely in varying levels of the psychic system. The physiologist or physician penetrates into the depths of activities of the somatic system. The sociologist is more concerned with the interaction of individuals as total persons within various social or environmental settings. It is not possible for any person to fully understand a system from its structural analysis attained by working inside that system alone, however. One can learn more about interrelations between somatic and psychic or between psychic and social systems by making observations at the boundaries of their interactions. In order to understand more adequately the processes at work in the total psycho-somatic-social field, however, one must understand the processes that go on in transaction among at least three systems by assuming a more distant position outside the system but within the field.

The point seems well made. But in 1945 George Engel and John Romano called for a ‘more comprehensive frame of reference or conceptual scheme of disease [than that] with which the student had heretofore been ... familiar ... [a] conceptual scheme ... in which psychologic and social factors exist or coexist with more impersonal biologic factors, eventually to cause, provoke, or otherwise modify variations in the total human biologic behavior.’ If near misses count, why do Engel and Romano not get priority? If Ghaemi had given the history of psychosomatic medicine its proper weight, he would have known that it was founded on the idea of a link between the biological, the psychological and the social. Grinker first used the term ‘the biopsychosocial model’ in 1962, some eight years after Nathan Ackerman used it in 1954 (in a paper entitled ‘Some Structural Problems in the Relations of Psychoanalysis and Group Psychotherapy’ (International Journal of Group Psychotherapy (Jan 1, 1954): 131-146). It was also used in two papers by F. A. Weiss in 1958. In short, the case for Grinker’s priority is not credible.

Focusing who was first to use the word or the phrase impoverishes the historical context. But the context is both rich and relevant. The biopsychosocial model was not just an aspiration, it was embodied in a programme of research and it is as a research programme that Ghaemi should have considered it. He mentions a few projects, notably Engel’s study of a baby named Monika who was born with an oesophageal atresia, which meant she had to be fed through a surgically produced gastric fistula during the first two years of her life. In the United States, psychosomatic medicine in the 1930s was associated with Walter Cannon who was no Freudian. By the forties, psychoanalysis had made it its own. At this point the leading figures were Helen Flanders Dunbar and Franz Alexander, who was one of three psychoanalysts to psychoanalyse Roy Grinker, the other two being Freud himself and Therese Benedek. In 1932 Alexander set up the Chicago Institute of Psychoanalysis, independently of the American Psychoanalytic Association. Unlike every other psychoanalytic institute in the world, the Chicago Institute was not specifically focused on mental health. Instead, trainees were invited to study the role of psychic factors, especially emotions, in bodily disturbances. Grinker’s 1952 paper in which he posits a unity of body, mind and social world was given to Alexander’s Institute and begins with a fulsome tribute to him. Alexander believed there were psychological determinants of duodenal ulcers, a research programme Ghaemi ridicules, citing the discovery of the Heliobacter Pylorus. Grinker practised conventional five-times-a-week analysis for many years. Engel never did. Ghaemi concedes that Engel came to psychoanalysis slowly but he pays almost no attention to his pre-psychoanalytic research in psychosomatics. In the above-mentioned paper, Theodore M. Brown has documented the extent of Engel’s involvement in psychosomatic medicine between 1938 when he got his first hospital job and 1946 when he began his training analysis. He shows that until 1942, Engel was hostile to Freudian ideas.

If Ghaemi had followed Engel through a single one of his studies he might have understood the biopsychosocial model better. Ghaemi likes to quote recent writers’ definitions but Engel’s contribution is presented in a few summary sentences. Even less defensible is the attribution to Engel of positions he never advances such as the claim – culled from Giovanni Fava in 2005 - that all disorders have multiple, co-occurring and distinct bio, psycho and social determinants. At a minimum, it is essential to distinguish the biopsychosocial model in clinical practice from the biopsychosocial model in research. The two are of course related but the second cannot be read off from the first. Like all theories rooted in a notion of the whole person, the biopsychosocial model in clinical research looks at the course of an individual’s life. Engel and Franz Reichsman – another internist-psychoanalyst – carried out a longitudinal study of Monica that lasted 40 years. They were interested in understanding better the physiology of the mother-infant relationship. In this their work invites comparison with Bowlby’s work on attachment and maternal deprivation or even Harry Harlow’s experiments on monkeys separated at birth from their mothers. When Monica was 15-months’ old she was readmitted to hospital having failed to bond with her mother who was then 20. She stayed in hospital for nine months during which time she became attached to Reichsman and one of the nurses, both of whom in turn became quite attached to her. During periods of separation, Engel and Reichsman did cry but became unresponsive and withdrawn. They observed a loss of muscle tone, profound immobility, a sad facial expression. Monica’s gastric secretions ceased and could not be stimulated using histamines. When Reichsman or the nurse reappeared, Monica’s muscle tone returned, she moved about, showed joy and her gastric secretion rates increased. Engel and Reichsman named the complex of psychophysiological withdrawal a ‘depression-withdrawal reaction’. In a fine article on the evolution of Engel’s psychoanalytic thought, Graeme J. Taylor (who trained with Reichsman) notes that Engel was wary of talking in terms of psychogenic causation. However, in common with many researchers who have carried out studies of attachment patterns since then Engel did think that ‘hidden within the interactions between infant and mother are a number of processes by which the mother serves as an external regulator not only of the infant’s behaviour and autonomic physiology, but also of the neurochemistry of its maturing brain’ (Taylor, 451). Ghaemi claims that ‘There is not and never has been much ‘Bio’ in the biopsychosocial model in psychiatry’ (49). The Monica studies contradict him.

Ghaemi presents Engel’s resistance to Alexander’s search for psychic causes of physical ailments as an instance of his theoretical impotence. Unlike Grinker, says Ghaemi, Engel hadn’t a clue how to put together the three parts of the biopsychosocial model but pretended he had through his version of the biopsychosocial model, a name he took from Grinker without acknowledgment. There is nothing fraudulent about noticing an association before it can be understood in causal terms. And when associations are demonstrated in areas that were hitherto not thought to be in play, they deserve notice. Modern stress research – a very biopsychosocial field – finds itself in a similar position. It demonstrates the ways in which certain mental dispositions trigger physiological changes that begin in the autonomic nervous system and the endocrine system and often affect the whole body. Stress doesn’t cause heart disease or diabetes. It exacerbates other biological weaknesses that make one more vulnerable to those conditions. A physician treating a patient with metabolic syndrome does not have to choose between the causes but can attack them all, biopsychosocially.

Ghaemi’s determination to interpret the biopsychosocial model narrowly as a treatment protocol enables him to withhold from consideration biopsychosocial research in psychiatry. The entire field of gene-environment epigenetic studies is biopsychosocial. The interaction between serotonin transporter genotype and adverse environment in depressive disorders is well established. But the association does not shed much light on the causes of depressive disorder. Many researchers have suggested that chronic stress results in overproduction of corticotropin-releasing hormone by the hypothalamus and of the stress hormone cortisol by the adrenal gland and that their combined action may underpin the interaction between serotonin and adverse environment.

Monkeys are born with either a long or a short version of the serotonin transporter gene (5-HTLLP). Stephen Suomi and his colleagues at National Institute of Child Health and Human Development (NICHD) in Bethesda, Maryland, looked at the impact of mothering on serotonin metabolism in a large group of rhesus monkeys. Half of the monkeys had been separated at birth from their mothers and been reared by their peers. The other half were left with their mothers. At six months, the peer-reared group were returned to their mothers. When the monkeys were eight years old, Suomi and his colleagues identified the version of the serotonin transporter gene that each monkey carried. They found that peer-reared monkeys with the shorter version of the gene had poorer serotonin metabolism than monkeys with the longer version but that mother-reared monkeys with the short version did just as well as those with the longer version. In other words, good mothering appears to ensure that monkeys with the shorter version of the gene still metabolise serotonin properly. This maternal buffering is an eminently biopsychosocial example of a gene-environment interaction. With aggression it’s even more interesting. Monkeys with the short version of the serotonin transporter gene were more aggressive than monkeys with the longer version but mother-reared monkeys with the short version were least aggressive of all. If you look at alcohol consumption the same thing is found. Monkeys with the short version of the serotonin transporter gene were more likely to drink alcohol to excess than monkeys with the longer version but mother-reared monkeys with the short version were less likely to drink alcohol to excess than mother-reared monkeys with the long version of the gene. In both these cases, what may be a genetic risk factor for individuals with poor early experience may actually be a genetic advantage in individuals with good early experience. For some time, researchers looking at human mental health looked for associations between the two variants of the gene and major mental health conditions. In the Dunedin Longitudinal study (led by researchers at King’s College London and the university of Otago!), Avshalom Caspi and colleagues demonstrated that that humans with the shorter version of the gene were more likely to suffer from depression than those with the longer version but only if they had higher levels of concurrent stress or if they had a history of being maltreated. I could offer many more examples of biopsychosocial research of this sort.

On page 60 of his book, Ghaemi quotes George Davey Smith who stated that ‘while he thought that psychosocial factors were relevant to the aetiology of many illnesses, especially psychiatric, they were not yet shown to be “direct causes” but rather influences on the distribution of known exposures”’. In support of this claim, Ghaemi considers the example of H. Pylori. Psychoanalytically-inclined practitioners of psychosomatic medicine were convinced that there was a strong association between certain kinds of personality and ulcers. But in 1983, Robert Warren and Barry Marshall discovered H. Pylori and discovered that it was the most common bacterium found in the human gut: about half of all humans have it. Marshall conjectured that H. Pylori actually causes the ulceration. Indeed, he demonstrated that it could by infecting himself with it and suffering gastric inflammation as a result. It is at this point that Ghaemi applies Davey Smith’s insight. ‘Perhaps social factor Z sometimes increases exposure to H. Pylori; even so, H. Pylori is still key to the causation of the illness and a major point for intervention, while social factor Z is secondary and only contributory to the extent that it that it affects exposure to H. Pylori’, he writes (61). ‘To ignore the ultimate biological cause may be counterproductive’. (Why would a biopsychosocially-oriented physician want to ignore an ultimate biological cause?) However, 15 per cent of those with duodenal ulcers don’t have H. Pylori in their gut. And only 10 per cent of those that do go on to develop duodenal ulcers. So what else might be involved? Associations have been found between ulcers and lifestyle habits such as alcohol consumption, smoking, and not eating breakfast in the morning, all of which might be proxies for stress. But according to Robert Sapolsky, there is evidence that the biology of the stress response may also play a part. And what initiates the stress response? Ghaemi’s social factor Z. As Ghaemi surely knows, the field of epigenetics of early life suggest that psychological experience can actually change a person’s biology. Suomi and his colleagues found that the structure of the brains of their monkeys differed markedly at only 6 months of age. In 4,400 genes the two groups of monkeys exhibited highly differentiated patterns of methylation. In other words, some genes were switched on and some were switched, for the long term. The idea that the psycho and social always play second fiddle to bio does not frame epigenetic research on the long-term consequences of childhood trauma.

Ghaemi asks whether a theory derived from a ‘small corner of the psychiatric profession is appropriate for the entire broad range of mental illnesses.’ But it’s not a small corner now. The biopsychosocial model has broadened out considerably since the 1930s to encompass attachment theory, most aspects of human behavioural biology, phenomenology. What Ghaemi really seems to object to is any departure from biomedically-oriented medicine. He waxes lyrical about Osler’s ‘clinico-pathological method’ in medicine in which symptoms were related to underlying morbid anatomy post-mortem. Is that really a way forward for psychiatry? An answer that a supporter of the biopsychosocial model might make is that the only way to do good biomedical research is to take account of the impact of psychological and social factors on biology.

Finally, there is the issue of the levels. It is not a false problem but I can’t see how it invalidates the biopsychosocial approach generally. It may that the only way to disentangle the bio from the psycho and the social is by doing research that considers all three. This is what has happened in epigenetic research.