July 17, 2013

Improving the Accuracy of Psychiatric Diagnosis

Author Picture

Allen Frances, MD

Professor Emeritus, Duke University School of Medicine, Durham, North Carolina


Psychiatric diagnosis faces a distressing paradox of simultaneous underdiagnosis and overdiagnosis.

About 5% to 10% of the general population has a clear-cut mental disorder that can be diagnosed reliably and treated effectively.1,2 Unfortunately, these people are massively underdiagnosed—only one-third of those with serious mental illness see a mental health clinician,1,2 and we have over 1 million psychiatric patients in prisons and jails3 because they didn’t receive adequate care in the community.4 These are people who desperately need the treatment they don’t get—they have a very low placebo response rate and suffer from grave disability.

Meanwhile, about 15% to 20% of the population is on the fuzzy boundary between having and not having a mental disorder, based on my clinical experience. Currently, these people are massively overdiagnosed and overtreated.5 People with mild and transient symptoms have a placebo response rate of up to 50% and will often get better on their own with time, education, advice, support, and reduction of stress.6 Once an inaccurate diagnosis is made, it often sticks for life and leads to stigma and unnecessary treatment.

My recommendation to doctors is to be extremely alert to severe mental disorders and extremely cautious and patient before diagnosing mild ones. An accurate diagnosis for mild conditions cannot be made on a first visit because often this visit occurs on one of the worst days of the patient’s life. Watchful waiting will most likely lead to a very different impression on subsequent visits. If medicine is given prematurely, patients will credit it for what is usually placebo effect and may therefore stay on the pills for long periods—deriving no benefit but accruing side effects.

Children are especially hard to diagnose on a first visit. They have a short track record, are developmentally labile, respond to stress, and change dramatically from week to week. Teenagers present with all of the above obstacles to accurate diagnosis with the frequently added complication of drug abuse. The geriatric population is also challenging—it is difficult to parse the impact of age, medical illness, medication side effects, and drug-drug interactions.

For mild conditions, it is almost always better to underdiagnose than to overdiagnose. It is easy to step up a diagnosis but very difficult to step down. False diagnoses are made quickly but can last a lifetime. A psychiatric diagnosis made well is often the beginning of a wonderful therapeutic relationship and effective treatment. Made badly, it can lead to disaster.

My advice to patients is to be informed, ask questions, and expect clear answers. A diagnosis requires the same kind of time and effort you would put into buying a house.

There has been a lot of talk lately about paradigm shifts in psychiatry and the hope for quick and sudden breakthroughs. Don’t hold your breath. The brain is defying simple answers, and we are no closer to understanding the cause of mental disorders than we were 30 years ago.

But we do already have effective diagnostic and treatment tools, if only we apply them judiciously to those who will be more helped than harmed.

Financial disclosure:Dr Frances is the author of Essentials of Psychiatric Diagnosis: Responding to the Challenge of DSM-5 and Saving Normal: An Insider’s Revolt Against Out-of-Control Psychiatric Diagnosis, DSM-5, Big Pharma, and the Medicalization of Ordinary Life.

Note: Dr Frances was Chair, DSM-IV Task Force, and is Former Chair, Department of Psychiatry, Duke University.


1. Kessler RC, Berglund PA, Bruce ML, et al. The prevalence and correlates of untreated serious mental illness. Health Serv Res. 2001;36(6 Pt 1):987–1007. PubMed

2. Alonso J, Codony M, Kovess V, et al. Population level of unmet need for mental healthcare in Europe. Br J Psychiatry. 2007;190:299–306. PubMed

3. James DJ, Glaze LE, for the US Department of Justice. Mental health problems of prison and jail inmates. Bureau of Justice Statistics Special Report. 2006; NCJ 213600.

4. Erickson CD. Using systems of care to reduce incarceration of youth with serious mental illness. Am J Community Psychol. 2012;49(3–4):404–416. PubMed

5. Mojtabai R. Increase in antidepressant medication in the US adult population between 1990 and 2003. Psychother Psychosom. 2008;77(2):83–92. PubMed

6. Stein DJ, Baldwin DS, Dolberg OT, et al. Which factors predict placebo response in anxiety disorders and major depression? An analysis of placebo-controlled studies of escitalopram. J Clin Psychiatry. 2006;67(11):1741–1746. Abstract

Category: Mental Illness
Link to this post:
Related to Improving the Accuracy of Psychiatric Diagnosis

10 thoughts on “Improving the Accuracy of Psychiatric Diagnosis

  1. Psychiatrist are diagnosing Depression ,ADD and Borderline PD in an excessive rate ,often for reasons that are not based on assessment but on expediency .
  2. I am a psychiatrist in Mexico. Psychiatrists live in Urban communities, Specialists and GP have limited training on mental disorders diagnosis and treatment. Abscence of diagnosis is the rule. Excesive treatment is the rule when they are detected. I agree with your arguments. Their implementation is limited in other contexts.
  3. I respectfully disagree with assumption that we need to know “the cause of mental disorders” in order to treat them effectively. I also disagree that we are no closer to understanding the cause than we were 30 years ago, but that is a separate issue.

    Knowing “the cause” is a luxury without which the rest of medicine has been doing rather well. We don’t know the cause of diabetes (we understand pathology, which is not quite the same) and yet it never stopped the internists from treating it. What causes cancer? Does it mean we can’t treat cancer?

    Alternatively, knowing the cause might be of little help. We do know the cause of TBI or Creutzfeldt–Jakob disease, now what?

    The problem with psychiatric diagnosis is not ignorance about “the cause” but outdated classification system. Given an opportunity to create a classification based on psychobiology of the brain’s functional units (there is enough data for a model), psychiatrists would be moving into the 21st century together with the rest of medical world. As it is – thanks to mavens in APA – we are stuck in the 19th for another decade and a half, at best.

  4. I consider myself fortunate that my residency training exposed me to a lot of learning opportunities. I trained in a 1000 bed general hospital; in a state hospital, in a very busy emergency room, in a children and adolescent clinic, in a geropsychiatric ward; in a substance abuse program, in an outpatient clinic that served an outpatient community with over a million consumers, and that we were trained in psychotherapy, family therapy, group therapy, pharmacotherapy, surgery rotations, internal medicine rotations, pediatric rotations,OB-GYN rotations, Orthopedic rotations, Consultation/Liason rotations, Neurology rotations, and we spent many hours a week in supervision and being coached/critiqued byour peers and our supervisors while our interviews with patients were observed through two-way mirrors. I must have seen hundreds of thousands of patients with very close supervision the entire time. My mistakes were pointed out to me early and one can learn things in a lot more lastingly impressive manner when their mistakes are seriously but benevolently pointed out to them.
  5. Kind of impossible to do since the psychiatric nosology is no more than a system of sanctioned name-calling endorsed by a committee of so-called experts who confuse assumption and opinion with valid evidence.
  6. i frequently see bipolar d/o diagnosed when the mood lability described is more likely something else, because nonpsychiatrist clinicians don’t paay attention to the duration of mood changes. Also “racing thoughts” has become a pet peeve of mine.
  7. First, let’s be clear here, the vast majority of mental health practitioners (myself among them) are NOT medical doctors. Rather, like me, they are MSW/LICSW folks and psychologists and counselors of various and sundry ilk. And, then there are the educational personnel who often are woefully unprepared and/or ill-trained. Especially when it comes to children and adolescent, Dr. Frances is right-on – this population is difficult to parse as they “morph” rather quickly. Taking whatever time is necessary to ferret out a young person’s situation is crucial to their LIFE. Try not to be bullied to engage in a rush to judgement.

    From my perspective, it is managed care that demands a “DSM Number” in order for compensation to occur. So practitioners spend their time chasing an “appropriate diagnosis,” rather than spending time acknowledging, understanding and appreciating THE EXPERIENCE one has when challenged by a mental health condition.

    It’s no wonder that such under/over-daignosing takes place given the state of the overall systemic craziness.

    I’ll leave they myriad of the many other implicated issues inherent in Dr. Frances’ composition for another time.

  8. Mental health professionals are over-diagnosing psychotic disorders. First “schizoaffective” is so complex that I believe it cannot be diagnosed off a research unit. One can use “Bipolar II of which the major depression component is psychotic.” Clinicians are overly-impressed by reports of “hallucinations” but I very rarely see a careful characterization of these symptoms to see if they are due to a psychotic disorder. But most importantly the mental status exam must document evaluation of the capacity to maintain cognitive coherence given an ambiguous challenge, and this can be done ONLY by first making sure that the subject understands what youw want when you present a proverb, establishing this with the use of familiar proverbs and then introducing UNfamiliar proverbs. Only then can cognitive coherence in response to an ambiguous challenge be adequately evaluated. Shockingly, this point is not made in any relevant article or text to my knowledge, hence a huge false positive rate for “schizoaffective disorder” and “schizophrenia.”

Leave a Reply


Browse By Author



Browse By Author