I suspect that many people who come to this website are not fans of psychiatry. Although I am also critical of various aspects of psychiatric practice, I continue to go to work every day, where I treat many people. One of my goals with this blog is to try to articulate how I integrate into my work my own and others’ criticisms of the field.
There is much talk today in medicine about evidenced-based practice. This means that one bases treatment decisions on the best data available. I suspect the current emphasis on evidenced-based practice is a reflection to some extent of how so many of us tend to use anecdote or habit to guide us even when research suggests otherwise. In addition, studies – especially double-blind controlled studies – are important, but often in the course of a day a physician needs to make decisions for which there is not an adequate database. Anecdotal experience, on the other hand, can be compelling but misleading. In psychiatry we know that the placebo response in most treatment studies is high. Dr. Irving Kirsch has studied this in detail and his book, The Emperor’s New Drugs, offers an excellent discussion of the placebo response. When someone takes a pill and improves, it is impossible to know if that improvement is due to the specific physiologic actions of the drug or relate to non-specific effects (i.e., those that do not depend on the specific attributes of the drug in question) that we call the placebo response. The person is nevertheless improved – that is not in question. What we do not know is what is responsible for that improvement.
My experience with the drug lamotrigine taught me how easy it can be to be misled about the benefits of a drug. When this drug was being touted as a treatment for depression in people diagnosed with bipolar disorder, I was far along in my skepticism about new medications. Depression in that context can be very difficult to treat effectively. There was one double-blind study to suggest lamotrigine might be effective. It was small but promising. I work with many people who have been ill for many years and who have tried a variety of treatments – both pharmacologic and non-pharmacologic – without success. Many of them suffer enormously. I began to prescribe lamotrigine to appropriate patients. Frankly, I found it surprising that my patients seemed to get better. Colleagues had similar experiences and I decided that lamotrigine was effective. Several years later, I learned that the larger studies that are required for FDA approval did not support the earlier findings. Those results were not widely known or reported. Upon learning this, I talked with each of my patients who was taking that drug. I found that most people were not significantly improved from where they had been before they started taking the drug. What happened?
What I found was that although initially patients reported improvement, over time the majority had either drifted back to their previous states of depression or had tended to experience waxing and waning levels of depression. As this happened, I would discuss various treatment options but I did not stop lamotrigine because in my mind the drug worked. I assumed that if I stopped it the patient would only get worse. I informed my patients of the new information and most elected to stop taking the drug. Although they were no better after stopping the drug, they were no worse.
In future blogs I will describe how I am trying to reduce the effect of these sorts of biases in my clinical work. I cannot eliminate the placebo effect but I hope that with systematic tracking of my patients, I will be less likely to remember the dramatic cases or the ones that fit my hypotheses about what is and is not effective treatment.