Verified by 
There has lately been a loud cry in the fields of psychology and health care policy for advancing a scientific basis for therapy. Articles in Newsweek and recently The Washington Post have advised therapy consumers that their therapist better be using techniques that have been "scientifically proven" to be effective. If not, the therapist may be behind the times, misinformed, or worst of all unethical.
Stop. Let's think for a minute. What the proponents of evidenced-based treatment are talking about is randomized placebo controlled studies of treatment modalities. For example, take a population of patients with a simple phobia. Randomly assign them to three treatment categories: a placebo, a drug treatment or exposure therapy. See which group does best, and there's your science, your evidenced based treatment. Let's imagine, hypothetically but not improbably, that the results of the well run, placebo controlled, randomized study show that exposure therapy is most effective for the most people suffering from a simple phobia.
Keep in mind that in this hypothetical population, there is a minority for whom exposure therapy, at least in the conditions of the study, is not effective.
So now imagine a patient comes along - let's say a 25 year old woman; to the primary care physician, the therapist, psychologist, psychiatrist, or the psychoanalyst. She has a simple phobia. She also has OCD, a hint of ADD, chronic low level depression, and just had her third failed relationship with a boyfriend, each of which has followed the same self-destructive path. She describes her mother as "loving but cold". She feels, in some vague way, that she is different from other girls of her age.
What does the clinician (who does not live in the cleanly dichotomous world of health policy and enthusiasts for evidence based randomized control studies) do now?
She treats the patient - not a symptom or a set of disorders. Ideally, I would hope, with a mix of approaches that fluidly adapts itself to the needs and capacities of the patient at any given moment. I am a psychoanalyst, but I know that gradual exposure is the best way to treat a phobia. I would consider myself quite ridiculous if I prescribed a course of psychoanalysis for that.
On the other hand, I also know that problems related to having grown up with a "cold but loving mother" (whatever that may mean to this patient), can be complex, subtle, pervasive and potentially crippling for later life relationships. I might indeed recommend psychoanalytic psychotherapy to this patient for the sequelae of this complex early relationship environment. And I might suggest psychoanalysis if the psychotherapy ran into roadblocks and impasses after a time.
Once again, I find myself wanting to plead, "Can't we live with complexity and layers and multiple possibilities and nuance in mental health care, and avoid the dangers and false comforts of simple-simplistic-dichotomies?" Like the tempting but false dichotomy, there is evidience or there is not evidence. I understand and appreciate the scientific method, and value the data obtained by systemic, well designed studies with appropriate controls. But of necessity, these studies can generally test one element of a real person in real life, and their results must be used appropriately, in the context of a larger and more complex picture.
And finally, can someone point me to the basis that underlies the propositions that evidenced-based studies are the best way to reach policy and clinical decisions that help people the most? What's the control group?
