The title was irresistible: “Toward a Grand Unified Theory of Psychiatry.” I clicked the link to Emily Deans’s Evolutionary Psychiatry blog, where she posits inflammation as the underlying pathology of all psychiatric illness. This is an interesting idea, although Dr. Deans does not provide supporting evidence, at least in that post. Here I will show that even if something like inflammation is the problem at the cellular level, in practice we need more robust theories to understand how to help our patients. After a look at the conceptual models mental health professionals use today, I will explore the strengths and weaknesses of relying on a flexible combination of theories.
When a psychiatrist sits down with a patient, she is confronted with a deluge of information: his story, his behavior in the interview, information from his family and other clinicians, and records. Our minds organize this information by identifying patterns. Ideally, each pattern functions as a hypothesis to be tested by comparing it with information from our memories and from outside sources in a continuous process of understanding the patient and refining treatment.
I entered medical school with the hope of understanding how people work. (This may reflect my family’s engineering background.) I was familiar with great theories of physics, chemistry, and biology—Newton’s laws, the periodic table of elements, and evolution by natural selection. I assumed that similarly elegant principles governed how people think, feel, relate, and behave as individuals and in groups.
Psychiatry attracted me with similarly ambitious aspirations. Psychoanalysis was the dominant theoretical system, and analysts’ interest in unconscious phenomena extended from individuals, to relationships, groups, and social movements. While Freud recognized that his constructs were metaphors for physiological processes yet to be discovered, his theories of the id, ego, superego, and the ego’s mechanisms of defense provided a set of ideas for thinking about patients. But I found concepts themselves a bit slippery, and as a neophyte, it was very difficult to relate them to what I observed and experienced with patients. Frustrated, I attended lectures by a senior neurology resident who promised to explain how the brain worked, but I quickly learned that for all its intellectual rigor, the neurology of the time had little to say about how people experienced life or about psychiatric illness.
With residency came responsibility for actually taking care of patients rather than speculating about what make them tick. What a relief it was when I first understood the concept of psychiatric diagnosis. I was with a group of residents and a supervisor puzzling over a patient’s family dynamics, and it suddenly came to me that if I simply listed his symptoms—depressed mood, loss of interest, irritability, suicidal ideation, insomnia, reduced appetite, lack of energy, and difficulty concentrating—I could diagnose him as depressed, which predicted some hope of recovery and guided me toward treatment. For some years thereafter my reliance on diagnosis was limited—I was immersed in the lives of my disturbed patients and listening to conflicting and often highly idiosyncratic opinions from teachers and colleagues. Psychiatry was turmoil then: the psychoanalysts who had dominated the field for decades were being displaced by upstart psychopharmacologists. I tried to learn both approaches and others as well, including family systems and existential psychiatry. Sensing that each of these had something valuable to contribute and none had a picture of the whole patient, I avoided committing myself to one approach. But it was troubling not to have a clear way to think about my patients and their problems.
With the revolutionary publication in 1980 of the DSM-III, the third edition of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, I finally grounded myself in diagnosis as the first step toward understanding my patients. The DSM is a remarkable compilation of a century’s research by psychiatrists and other scientists from around the world. Its latest edition, the DSM-V, categorizes almost every imaginable form of deviant behavior and suffering into 157 diagnoses, each of which can be further specified by severity, time course, remission status, and other qualifiers. Like most psychiatrists, I find it immensely useful for conceptualizing patients’ problems and recommending treatment, particularly medication.
Relying too much on diagnosis as a conceptual model is problematic. Psychiatric diagnoses are based on what a patient says, how he behaves, his family history, and the course of his illness. So far, none of many attempts to include blood tests, scans, or other biomarkers has passed the test of clinical usefulness. But recent neuroscientific discoveries are casting doubt on the validity of our diagnostic categories: many do not correspond to how the brain actually functions and malfunctions.
Further, a psychiatrist who limits himself to diagnosing his patients cannot capture the complexity of their problems. To diagnose a patient’s panic attacks, manic episode, oxycodone addiction, or multi-infarct dementia says a great deal, but it leaves out his relationships, lifestyle, and values, as well as whether he will accept or can afford treatment. More fundamentally, it ignores the pathophysiology of his individual brain and how it plays out in his psychology, relationships, and life. So while diagnosing a patient is a good start, by itself it is too limited serve as a real guide to treatment.
Recognizing the limitations of diagnosis, many clinicians use psychological theories to guide their work. The psychoanalytic movement spawned a number of these: unconscious conflict, ego mechanisms of defense, and object relations are examples. Family systems theory, cognitive psychology, and behavior therapy are other effective frameworks for thinking about some patients and their treatment. More recent approaches combine concepts from more than one theoretical model. These include cognitive-behavioral therapy, internal family systems therapy, emotionally focused therapy, and dialectical behavior therapy. In the hands of expert clinicians, each of these is useful for certain groups of patients, but also limited in the types of patients and problems it can effectively treat. And excessive reliance on theory can do real harm: witness the damage that resulted from dogmatic application of theories derived from psychoanalysis to blame mothers for causing their children’s autism and schizophrenia.
George Engel did not publish his biopsychosocial model until after I finished my residency, but I was exposed to his ideas earlier, since two of my medical school teachers had worked with him at Rochester. What a relief it was to find a model that accounted for the multiple dimensions of patients’ problems! A few years later, the DSM-III added “axes” to include the patient’s personality, medical problems, stressors, and severity of illness in the diagnosis. (The axis approach was not popular with clinicians, and the DSM-V abandoned it.) Unfortunately, these multidimensional models are even more limited in clinical utility than diagnosis itself. They function as problem lists, and they remind clinicians not to overlook anything, but the best they can manage to put the problems together is a Venn diagram of overlapping circles. They say nothing about how the various factors interact and are incapable of generating testable clinical hypotheses for use in treatment.
Many experienced psychiatrists, including myself, end up taking eclectic or pragmatic approaches. We combine or switch among theoretical models and draw on our experience to choose what is likely to be most effective for each patient at a particular time. This is often an unconscious process, part of the mysterious human talent for synthesis. The strength and also the limitation of such an approach is that each practitioner, each clinician-patient relationship, each patient encounter, and even each moment within a session are utterly unique, complex, and irreproducible. The clinician must rely on her own creative talents to decide what is most important and how to intervene. This makes it difficult to practice anything like evidence-based treatment. Further, clinicians are influenced by their own personal stresses, limited perspectives, and emotional vulnerabilities. There is always doubt about whether one has made the right choices, but such continual questioning is part of the process. The best decision-making instrument we have is still a well-trained, experienced clinician, with his or her personal insights, talents, and limitations.
What about a “Grand Unified Theory” of psychiatric illness? Inflammation, as proposed by Dr. Deans, is a relatively weak theory, for its recommendation of attention to diet, exercise, stress reduction, and sleep, while sound, is not adequate to treat most psychiatric problems. It will become more robust if we discover the mechanisms by which cellular inflammatory processes contribute to disturbed thoughts, emotions, and behavior, but at this time we have no overall theory. The diagnostic process is a starting point to understanding what is going on with a patient. The clinician must decide what else is relevant from the medical and social history. Treatment with psychotherapy, medication, or both is then a process of using psychiatric concepts to recognize patterns which can guide therapeutic interventions. The patient’s response to each intervention sheds further light on the situation and guides the next intervention.
Conceptual models, including diagnosis, are essential for this process, but we still lack a theory with enough explanatory power to be useful with all patients. We therefore use models from a variety of sources. Recent discoveries in neuroscience are challenging some of the models we use, including our diagnostic categories, but given the complexity of the brain and body, it is unlikely that neuroscience will provide an overall conceptual model in the foreseeable future.