Some months ago, I received a stern admonition from my family doctor. My fasting blood sugar of 99 mg/dL was “right on the border”, he said, and I had better work on bringing it down. “But,” I protested, “when I was in medical school (in the 70s), the normal FBS range went up to 110 mg/dL!” "Well,” he replied a bit huffily, “they changed the criteria!”
Some months ago, I received a stern admonition from my family doctor. My fasting blood sugar of 99 mg/dL was “right on the border”, he said, and I had better work on bringing it down. “But,” I protested, “when I was in medical school (in the 70s), the normal FBS range went up to 110 mg/dL!”
"Well,” he replied a bit huffily, “they changed the criteria!”
Indeed, “they” did. The most recent criteria identify “pre-diabetes” (also called “impaired glucose tolerance”) as an FBS between 100 and 125 mg/dL.
But isn’t this what psychiatrists get accused of doing all the time - moving the diagnostic goal posts so as to broaden the inclusion criteria for various disorders? And doesn’t this inevitably lead to over-diagnosis and over-prescription of potentially hazardous drugs? Such, at any rate, is the thinking behind certain claims in a recent petition drawn up by some members of the American Psychological Association, and strongly supported by some prominent psychiatrists.
But the diabetes experts who promoted the new, lower threshold for an abnormal FBS did so based upon sound epidemiological and clinical data. As the American Diabetes Association website notes,
“People with prediabetes are at higher risk of cardiovascular disease. People with prediabetes have a 1.5-fold risk of cardiovascular disease compared to people with normal blood glucose…We now know that people with prediabetes can delay or prevent the onset of type 2 diabetes through lifestyle changes.” (http://www.diabetes.org/diabetes-basics/prevention/pre-diabetes/pre-diabetes-faqs.html)
To the best of my knowledge, the prediabetes criterion change was not based on predictions of how much or how little anti-diabetes medication doctors were likely to prescribe, or how much “Big Pharma” would stand to gain or lose from the lower FBS threshold. Yet some in the mental health field would like the DSM-5 criteria sets to reflect and preempt such entirely hypothetical changes in prescribing patterns. Thus, the psychologists' “Open Letter to the DSM-5” warns,
“The proposal to lower diagnostic thresholds is scientifically premature and holds numerous risks. Diagnostic sensitivity is particularly important given the established limitations and side-effects of popular antipsychotic medications. Increasing the number of people who qualify for a diagnosis may lead to excessive medicalization and stigmatization of transitive, even normative distress.” (http://www.ipetitions.com/petition/dsm5/)
Now, I would agree that several of the proposed changes in the DSM-5 are indeed “scientifically premature”, in so far as they lack a sound foundation in long-term, epidemiological and clinical studies. (Others, such as dropping the “bereavement exclusion” in the diagnosis of major depressive disorder, are founded on quite good science, in my view). I also support an independent review of the proposed DSM-5 changes. And, yes-psychiatrists must consider the overall health risks of their diagnostic system, as Dr. Allen Frances and others have repeatedly pointed out on these pages. But such deliberations must carefully weigh both the risks and the benefits of changes in the diagnostic system, based on well-validated studies-not on the basis of crystal ball gazing!
Objecting to a reduced threshold for a particular disorder based on the cynical prediction that “doctors will use it as an excuse to write more prescriptions” is neither good science nor good public policy. Furthermore, until we have universally accepted criteria for a given disorder, backed up by a veridical biomarker, it simply begs the question to assert that a lower threshold for a diagnosis will lead to “false positives” or “false epidemics.” Without an unimpeachable criterion for “true,” we cannot declare a finding to be “false”!
Absent empirical studies showing a direct connection between the newly-proposed diagnostic threshold and actual harm to patients, the “over-prescribing” argument amounts to social engineering-not scientifically-based nosology. It is asking our diagnostic criteria to do the work of our continuing medical education-which, let us be frank, could certainly use improvement. If, for example, we want to discourage physicians from prescribing antidepressants for the very mildest cases of major depressive disorder (MDD), the best means of doing so is through enhanced training and education-not through manipulation of our diagnostic criteria for MDD. Of course, there may be perfectly sound reasons for “tightening” the MDD criteria, based on considerations of reliability, validity, and response to treatment-but these are scientifically-justified reasons for modifying any medical diagnosis, and have nothing to do with anticipated “misuse” of the criteria by over-zealous prescription writers.
Furthermore, there may be compelling public health reasons for identifying certain “pre-clinical” or “prodromal” psychiatric conditions, based on morbidity and mortality data, and on the longitudinal outcome of these conditions. Yes, the potential for inappropriate or premature pharmacotherapy exists with respect to these more subtle derangements--but that risk alone should not dictate our nosology. Moreover, it is fallacious to equate “treatment” in psychiatry with medication, even though rates of psychotropic prescriptions have been rising substantially in recent years-often owing to prescriptions written by primary care physicians rather than psychiatrists.
Indeed, some sub-clinical or prodromal states may respond well to various non-pharmacological interventions. For example, a recently-released study in the British Journal of Psychiatry (Bechdolf et al, Br J Psychiatry. 2011; Nov 10) examined the construct of “early initial prodromal state (EIPS) of psychosis in which most of the disability and neurobiological deficits of schizophrenia have not yet occurred.” The authors investigated the effects of an integrated psychological intervention (IPI), combining individual cognitive–behavioral therapy, group skills training, and other modalities, compared with “supportive counseling”. The primary outcome was progression to psychosis at 12- and 24-month follow-up. IPI was found to be superior to supportive counseling in preventing progression to psychosis at 12-month follow-up.
Let’s be clear: if confirmed, this would be a very important public health finding. It simply won’t do for critics of DSM-5 to argue that such pre-clinical or prodromal syndromes will inevitably lead to over-use of antipsychotic or other medications-though, of course, this outcome is possible. These critics must also weigh the potential benefits of early intervention, as well as longitudinal studies examining the risks of failing to intervene early in pre-psychotic states.
There may also be a critical role for the FDA, in promulgating stricter prescribing indications for antidepressants and antipsychotics. Rather than jiggering our diagnostic criteria to discourage excessive medication prescribing, I believe we would make more progress with an FDA-mandated statement in the drug labeling information-a sort of “brown box” warning stating, for example,
IMPORTANT: There are no convincing data showing that antidepressants are superior to placebo or non-specific interventions for very mild instances of major depressive disorder (e.g., Hamilton Depression Rating <24). Antidepressants may carry a substantial side effect burden, and in mild cases of MDD, it is preferable to begin a form of psychosocial treatment. Antidepressants appear most effective in cases of moderate-to-severe MDD, and in cases involving melancholic or psychotic features.
Ref: Ghaemi SN, Vhringer PA: Solving the Antidepressant Efficacy Question? Effect Sizes in Major Depressive Disorder, in press, Clinical Therapeutics.
Our diagnostic criteria should be guided by the best science we can muster, not by well-intentioned but misguided attempts to anticipate “bad behavior” on the part of clinicians. In short, if we are to change our diagnostic criteria in psychiatry, we should model our behavior on that of the diabetes experts, not the crystal ball gazers.
Acknowledgments: Thanks to James Knoll, MD, Nassir Ghaemi. MD, and Susan Kweskin for helpful comments on this piece.