Wednesday, September 24, 2008

Science or Scam: Neuro-imaging for ADHD?

More than a dozen years or so ago I was attending a conference in Israel on ADHD when one of the organizers--a neurologist-- asked me to please address the problem of Dr. Daniel Amen's claims about his subtyping of ADHD through the use of the SPECT imaging technology. The problem she said, was that many of her patients were flying from Israel to the U.S. in order to be "subtyped" and then treated by Dr. Amen in California. Single photon emission computed tomography (SPECT)) is a nuclear medicine tomographic imaging technique using gamma rays. It is very similar to conventional nuclear medicine planar imaging using a gamma camera. However, it is able to provide true 3D information. This information is typically presented as cross-sectional slices through the patient. One disadvantage of this technology, in contrast to MRI or fMRI, is that it requires giving a dose of a radioactive tracer. In response to Dr. Amen's talk I asked to see any data supporting his claims. He responded by saying that he had over 12,000 cases on which to base his typology. "What statistical methods did you use?" I asked. He replied that they had not been published yet, but that researchers like me would have to undertake such a huge job. More recently, 11 years later in La Jolla, California I happened to be on a panel with Dr. Amen and the same issue was raised about publications. He responded that there was now a publication, but he didn't recall the name of the journal; but he asked one of his colleagues in the audience for the name. The colleague looked puzzled, threw up his arms quizzically, and said he didn't know. So much for supportive scientific proof. A prominent neurologist and imaging researcher, George Busch, M.D. happened to be on the same panel. He unequivocally denounced Dr. Amen's claims and asserted that no respectable scientist had yet to find a way to use neuro-imaging to make those clinical subtype distinctions, let alone a diagnosis. Work by Jay Giedde, Judy Rapaport, and Javier Castellanos at NIMH with MRI and fMRI have indeed shown that there are important brain differences between ADHD and normal controls, both cross-sectionally and developmentally. But no one claims that any diagnostic rules from those data are capable of the precision required to beat clinical assessments. Here's what Dr. Amen claims about ADHD subtypes: Type 1 — Classic ADHD. Symptoms such as short attention span, distractibility, disorganization, procrastination, poor internal supervision plus hyperactivity and impulsivity.* Type 2 — Inattentive ADHD. Classic ADHD symptoms, but instead of hyperactivity, there is low energy.* Type 3 — Overfocused ADHD. Classic ADHD symptoms as well as negative thoughts and behaviors, such as opposition and arguing.* Type 4 — Temporal Lobe ADHD. Classic ADHD symptoms plus irritability, aggressiveness, and memory and learning problems.* Type 5 — Limbic ADHD. Combines ADHD with depression and low energy and decreased motivation.* Type 6 — The Ring of Fire. Cross between ADHD and bipolar disorder. Characterized by moodiness, aggressiveness, and anger. Now any experienced clinician will undoubtedly agree that these are recognizable forms of presentation at a child clinic. In fact, these are classic descriptions from the literature: the hyperactive/impulsive type; the inattentive type; the overfocused type (e.g. Kinsbourne's type); the hypoactive type, etc. But are these "types" confirmed by an appropriate methodology as variants of ADHD? Where is the cluster analysis or factor analysis of large samples characterized through rigorous clinical documentation? Where are the structured or unstructured interviews and histories to validate the diagnosis? What are the statistical boundaries among these so-called types? What is the evidence that they respond differently to treatments or have other biological or genetic markers to distinguish them? If I had 12,000 cases in my database, I would not waste a day before exploring the typologies that might be hidden there. Amen's work is classic quasi-scientific mystification: the failure to distinguish between anecdotes and data, and between hypothesis and fact. Like all fringe quasi-scientific appeals to a needy public, there are classic signs of when the patients are being fooled: 1) There is an impressive and truly science-based technology, so sophisticated that the ordinary public must take the claims on faith; 2) The proponent of this new method, though possibly trained in traditional clinical and scientific paths, breaks with the majority of scientists and fails to pass the test of peer review; 3) The proponent himself (or herself) is too busy seeing patients and collecting large fees to do the necessary research themselves; 4) The proponent tirelessly appears at conferences and seminars worldwide, and develops an adoring but uninformed following despite repeated criticisms to produce real data; 5) Standard treatments are often the outcome from the elaborate workups and tests, though actual followup studies are seldom provided. I have to admit that personally Dr. Amen is charming, well-informed, and well-trained. He gives a convincing talk, and if I were an uninformed normal patient, I would probably agree that there is no definitive biological test for ADHD, no pathogonomic sign, and a truly complex clinical picture. I might possibly end up in desperation spending thousands of dollars after seeing the lovely colored pictures of the brain, with hot spots where ADHD resides. But fortunately, I have been around long enough to spot mumbo-jumbo when I see it. Let the buyer beware.