IS BEHAVIORISM BECOMING A PSEUDO-SCIENCE?: POWER VERSUS SCIENTIFIC RATIONALITY IN THE ECLIPSE OF TOKEN ECONOMIES BY BIOLOGICAL PSYCHIATRY IN THE TREATMENT OF SCHIZOPHRENIA

Behavior and Social Issues, Fall 2006 by Wakefield, Jerome C

Quoting a critic of biological psychiatry, Wong observes:

Results of over 1,300 outcome studies published since the mid-1950's reveals that 21% of schizophrenic patients on maintenance drug therapy relapse as compared to 55% on placebo (Cohen, 1997). Given the bias of scientific journals for publishing positive results (Dickerson, Chan, Chalmers, Sacks, & Smith, 1987) this probably represents an optimistic estimate of the impact of these drugs. By subtracting the relapse rate for neuroleptics from that for placebo treatment, we may compute a "net" drug effectiveness rate of 34%. Undoubtedly, preventing relapse in one third of treated patients is a significant effect; yet, it is probably inadequate justification for past assertions that these drugs were essential or past practices of prescribing these drugs universally. (2006)

Accepting the cited figures, in my view this is a bewildering assessment. We are talking about life-destroying and family-destroying disorders. In any other such medical domain (e.g., cancer treatment), a one-third improvement rate over placebo would be cause for joy, and would be considered justification for recommending universal administration of a therapy, if there is no test that distinguishes the one-third who would benefit (of course allowing for alternative effective therapies and considering patient preferences and other such factors; no treatment should literally be reflexively "universal"). Indeed, much smaller percentages are often reasons for celebration. (As it happens, as I write this there is an article just published that recommends universal chemotherapy, despite its negative side effects, for estrogen-insensitive breast cancers over a certain size, given that a meta-analysis established that 23% of the cases will benefit; the medical literature is replete with smaller benefit percentages that are considered breakthroughs.)

Wong notes that biases in publishing positive results suggest that the effectiveness of medication is less than we know. It is true there are distortions in the system of studying drugs, not only publishing bias but many other issues as well (e.g., selecting proper dosages, large numbers of nonresponders, spontaneous remissions due to the cyclic nature of many disorders, and false positive misdiagnoses) that may lead to complexities of interpretation and inflated apparent effectiveness. However, Wong fails to consider factors that could suggest that drug effects are more powerful than we know. These include, for example, the PDA-mandated "intent to treat" methodology that may not even report the results for those actually completing treatment despite high drop-out rates, the high rate of spontaneous remissions that weaken apparent effectiveness, and the fact that drug trials almost always try one drug at set doses whereas in clinical practice it is well known that individuals often respond very differently to given dosages and to different drugs within a class, and the clinician will often try several different drugs or a combination of several drugs before finding the right prescription that works for a given client (some recent antidepressant studies are starting to take this into account, and seem to be showing higher rates of remission than single-drug studies).