Saturday, July 20, 2013

Is the FDA objective enough to assess treatments in psychiatry - or is this just politics as usual?

The American Psychiatric Association (APA) feed posted a link to this FDA news release regarding a new biological test for Attention Deficit Hyperactivity disorder.  The device is essentially a quantitative EEG (QEEG) machine.  The QEEG heyday was back in the mid 1980s to 1990's.  Devices were designed that could take the standard output of an EEG montage and look at the frequency bands and how that activity fluctuated topographically within the individual.  There were two major manufacturers at the time and both of those technologies allowed for a comparison of the subjects QEEG with a standardized groups.  The difference could be determined as a t or z score and that was plotted relative to the electrode placements.  The final analysis would yield maps consisting of frequencies and mathematical operations on those frequencies.

There were several articles on this methodology including an impressive article in Science on the diagnostic capabilities of these instruments.  One manufacturer provided an algorithm of clinical features and EEG features that purported to diagnose major psychiatric disorders.  You could actually analyze the data both ways - with or without the clinical features.  There was enthusiasm to the point that a new psychiatric subspecialty in electrophysiology was made to meet the requirements of psychiatrists who wanted to use QEEG technology.

In 1988, I was so impressed with the technology that I approached a potential employer and struck a bargain that I would take a salary cut if they would buy me the machine and the deal was struck.  I was fortunate enough to be affiliated with a certified electrophysiology lab with an outstanding electrophysiologist and EEG technologists.  This was critical in order to collect standardized data and select numerous 2 second epochs of EEG data for computerized analysis.  The epochs had to be completely free of artifact in order to provide valid data for analysis and anywhere from 30 to 60 of these epochs needed to be selected per patient.

If you think about it for more than a few minutes, what is wrong with the idea that EEG frequencies should point to a specific psychiatric diagnosis?  The short answer is a lack of specificity.  There are literally hundreds of conditions that can lead to fast or slow frequencies including normal fluctuations of conscious states.  During my QEEG work we had to collect EEG epochs for analysis in the "eyes closed but alert" state.  Quantitative EEGs can demonstrate significant fluctuation in that state.

After several hundred QEEGs with and without the computerized algorithm, it was apparent that the diagnostic abilities of QEEG were low.  There were literally a handful of analyses that seemed to match the clinical diagnosis and at that point we shut down the project.  As far as I can tell from their web site, that company no longer sells a QEEG machine claiming to make psychiatric diagnoses.

I have not been able to locate the specific reference for this FDA approval.  The FDA press release states:

"In support of the de novo petition, the manufacturer submitted data including a clinical study that evaluated 275 children and adolescents ranging from 6 to 17 years old with attention or behavioral concerns. Clinicians evaluated all 275 patients using the NEBA System and using standard diagnostic protocols, including the Diagnostic and Statistical Manual of Mental Disorders IV Text Revision(DSM-IV-TR) criteria, behavioral questionnaires, behavioral and IQ testing, and physical exams to determine if the patient had ADHD. An independent group of ADHD experts reviewed these data and arrived at a consensus diagnosis regarding whether the research subject met clinical criteria for ADHD or another condition. The study results showed that the use of the NEBA System aided clinicians in making a more accurate diagnosis of ADHD when used in conjunction with a clinical assessment for ADHD, compared with doing the clinical assessment alone."

From that appears to be this registered clinical trial.  No results are reported and there are no publications in peer reviewed journals that I can find.  The concerns about this technology should be apparent from the history outlined in the above narrative and the same application suggested by the FDA.  This is not a diagnostic procedure but one that is a supplement to the clinical evaluation for ADHD.  It reminds me what Russell Barkley - noted ADHD expert and scholar said in a seminar I attended last fall.  There are no gold standard tests for ADHD any more than there are for any other problems of executive function.  He pointed out that hours of neuropsychological testing (he is a neuropsychologist) is no more accurate than standard ADHD checklists.  Neuropsychological testing is important because of the high prevalence of learning disorders in ADHD.

My prediction at this point (pending an actual published research paper) is that this QEEG machine will not be that clinically useful and if it is a question of neuropsychological testing versus the QEEG, neuropsych testing should be the the option because it can detect and allow for treatment planning for any associated learning disorders and QEEG cannot. One of the risks here in an age where insurance companies deny diagnostic costs is that neuropsychological testing is denied and the QEEG substituted depending on cost.  That would not allow for the recognition or treatment planning for a learning disorder.

The larger question is how competent the FDA is to make decisions on devices for psychiatric disorders?  The FDA came out with a notice in 2011 that electroconvulsive therapy devices may need to be reclassified (Class II to Class III) resulting in the need for additional testing, clinical trials, and regulation.  That occurred after two generations of psychiatrists were trained on the current devices and have clinically demonstrated that it is a safe, effective and in many cases life saving therapy.  They completed their own study and meta-analyses and it is unclear to me what they concluded.  I consider the FDA web site to essentially be unnavigable.  Available information in the psychiatric literature suggests that they are still is the process of coming up with a formula for reclassification of ECT devices to a more restrictive category and that their analysis of the efficacy of ECT may have been seriously underestimated.  The concern of the authors is that reclassification will restrict availability of ECT to patients who have clear indications for its use much in the same way that poor Medicare reimbursement restricts the availability in some hospitals now.

The even larger question is there some kind of systematic bias operating here?  Both the ECT and QEEG decisions seem mismatched with the available science and clinical experience.  The FDA has the appearance of transparency, but you can never find what you need in the thousands of web pages that are linked to the agency.  In the ECT example, I could not find a clear statement, vote or conclusion about the ECT decision until I read the article by Weiner, at al.  In the case of the QEEG device there is no publication of the study supporting its use.  Independent review suggests that there have been no advances in the past 16 years.

George Dawson, MD, DFAPA

FDA Executive Summary.  Meeting to Discuss the Classification of Electroconvulsive Therapy (ECT) Devices.  January 27-28, 2011.

Weiner R, Lisanby SH, Husain MM, Morales OG, Maixner DF, Hall SE, Beeghly J,Greden JF; National Network of Depression Centers. Electroconvulsive therapy device classification: response to FDA advisory panel hearing and recommendations. J Clin Psychiatry. 2013 Jan;74(1):38-42. doi:10.4088/JCP.12cs08260. PubMed PMID: 23419224.

Sand T, Bjørk MH, Vaaler AE. Is EEG a useful test in adult psychiatry? Tidsskr Nor Laegeforen. 2013 Jun 11;133(11):1200-1204. English, Norwegian. PubMed PMID: 23759782.

Nuwer M. Assessment of digital EEG, quantitative EEG, and EEG brain mapping: report of the American Academy of Neurology and the American Clinical Neurophysiology Society. Neurology. 1997 Jul;49(1):277-92. Review. PubMed PMID: 9222209.

"E. On the basis of current clinical literature, opinions of most experts, and proposed rationales for their use,QEEG remains investigational for clinical use in postconcussion syndrome, mild or moderate head injury, learning disability, attention disorders, schizophrenia, depression, alcoholism, and drug abuse." (from Nuwer 1997)

1 comment:

  1. I have been finding the endorsement of this for diagnosis and treatment (at least by the American Academy of Pediatrics) kind of questionable as well. The studies that I've seen don't point to any efficacy for diagnosis, and as a treatment it's a lot of money and effort for something that provides less benefit than medication and behavioral interventions. I wonder how much of the information is being pushed by alternative practitioners - but I'm always suspicious of that.