A Better Analysis Of The Psychiatrist "Shortage"

A paper in Health Affairs on the "psychiatrist shortage" has been getting a lot of press lately (1).  People are acting like the authors' conclusions are definitive rather than highly speculative, but that is a standard approach in the press.  In the article they use American Medical Association (AMA) Physician Masterfiles from 2003 and 2013 to calculate the number of psychiatrists per 100,000 population for those two dates.  They compare it to similar data for neurologists and family physicians.  Between 2003 and 2013 there was a -0.2% change for psychiatrists as opposed to a +35.7% change for neurologists, a 9.5% change for primary care physicians and a 14.2% change for all physicians in this time period.  They also calculated medians for all groups and coefficients to look at workforce distribution.  As expected psychiatrists and neurologists showed some skew of distribution compared with adult primary care physicians.  That could also be seen in the density of psychiatrists by region:  24.47 per 100,000 in New England to 13.33 per 100,000 in the Pacific area.  They show the geographic distributions by highlighting quartiles of distributions on a quartile map of the United States.  The regions highlighted are 300 - Hospital Referral Regions rather than states.

There appears to be a significant typographical error on page 1275: "Our finding that there was almost a 10 percent decline in the population adjusted mean number of psychiatrists per HHR supports the belief that the supply of psychiatrists likely limits patient access to their services".  They are referring to median numbers here and in their abstract where they use the correct term.  The actual number of psychiatrists in 2003 was 37,968 and in 2013 it was 37,889.  The real numbers just don't seem that dramatic.

In the context of these statistics the authors offer a very inconsistent analysis frequently equating the number of psychiatrists with access to services or imposing severe limitations on treatment as illustrated by their statement: "Since the current supply of psychiatrists is not meeting the needs of people with mental illnesses and is not keeping pace with population growth, policy makers and the medical community must consider ways to address the shortage and improve access to mental health care".  This conclusion is quite a stretch considering data that the authors include in the paper.  They use the figure of 9.6 million adults with severe mental illnesses and only 40% of those people receiving care.  That means if the 37,889 psychiatrists they counted had only 250 people with severe mental illness on their caseload - 100% of these patients would be treated.  I propose that psychiatrists only see patients with the severest forms of mental illness and in today's world 250 patients is a very modest caseload.  In the heyday of psychoanalysis, some analysts did not treat many more patients over the course of their career.   At the maximum this suggests a geographical mismatch between patients and psychiatrists rather than a global shortage of psychiatrists.  Is increasing the supply the best approach to this problem?

In another section of their paper the authors point out that psychiatrists account for only 5% of the mental health workforce; 95% being psychologists, social workers, therapists , and counselors.  They acknowledge that they have no equivalent statistics for those disciplines or nurse practitioners or physician assistants.  Many systems of care these days see a prescriber as a prescriber and selectively hire non-physician prescribers over psychiatrists.  Even without the data it would seem fairly obvious that there has been a proliferation of non physician prescribers over the past decade and no shortage of antipsychotic, antidepressant, stimulant, or benzodiazepine prescriptions.  How can there be a shortage of prescribers in a sea of overprescriptions?

The authors notion that "policy makers and the medical community" are going to provide the solution here is also incorrect on several grounds.  First and foremost - if there is a problem - these are the same people who got us here in the first place.  The authors themselves reference a Graduate Medical Education National Advisory Committee study from 30 years ago predicted the shortage.  Any Medline search looking at "psychiatrist shortage" will also yield papers on this topic dating back to 1979.  In that time frame there have been very modest attempts to expand the workforce in psychiatry.  I made that statement based on expansion of residency slots.  The reality is that there are many International Medical Graduates who are well qualified for residency positions and may have even completed equivalent certifications in their country of origin.  The authors also seem to miss the point that these same "policy makers" have initiated policies to expand non-physician prescribing that has led to decreased staffing by psychiatrists in many settings.  They make the typical mistake that policy makers can't have it both ways and they seem quite intent on reducing rather than expanding the psychiatric workforce.  In the argument the only function a declaration of a psychiatrist shortage limiting mental health treatment is to scapegoat psychiatrists for a problem that may be imaginary but at the minimum is out of their control.  The appeal to policymakers also ignores the fact that policy makers in the US, generally advance pro-business policies that place both physicians and their patients at a distinct disadvantage compared to the business.  I will address some of those points below.

Some additional points not considered by the authors:    

  1.  Inefficiencies in the psychiatric workforce are large - Those inefficiencies are two fold.  First, the practice of psychiatry is notoriously inefficient.  I have done comparisons with both ophthalmology and orthopedic surgery on this blog where comparatively fewer specialists cover an impressive array of serious illnesses.  They do this largely through a much better triage system focused triaging the most serious illnesses.  By comparison, the conditions treated by psychiatrists all receive rationed care and in some cases - the care is completely displaced to a non-medical facility.  In most others there is inadequate infrastructure to address the problem.  The facilities themselves are managed by non-medical administrators who in may cases have caused disruptions in care and severe quality problems.  Care is further fragmented by the fact that managed care companies and governments do not provide realistic reimbursement for the care delivered and incentivize hospitals to provide minimal care.

Second,  managed care and government bureaucrats in their infinite wisdom have made psychiatry even less efficient.  I interject the term "medication management" here as an example and will elaborate below.

2.  The prevailing model of care is antiquated and a throwback to the 1980s - The preferred business and government model of care is the so-called medication management visit also more pejoratively known as the med-check.  It is based on a thoroughly poorly thought out idea that people with severe mental illnesses can be treated with medications for the symptoms of those illnesses.  That model does not work at even the most basic idea that there are social etiologies of symptoms that need to be addressed by social and psychotherapeutic interventions.  There are no medications that treat unemployment, separation and divorce, or the sudden loss of a loved one and yet the entire billing and coding structure for psychiatric visits was based on this model.  Even worse - the productivity scale for employed psychiatrists is still based on this model with a rough correlation between how many people are seen in one day and compensation.

3.   Academic and intellectual approaches to psychiatry are at an all-time low -  An intellectual approach to the field is critical whether considering phenomenology, the conscious state of the individual or all of the medical factors associated with treating the psychiatric disorder.  The environment is also frequently neglected because it is managed by non psychiatrists - at least until there is an incident or violence, aggression, self-injury, or suicide that requires analysis.  The intellectual approach to the field requires study of both the individual and the environment that they are in.  An intellectual approach to psychiatry also requires centers of excellence where people with those problems can go to receive expert care.  Centers of excellence are much less common in psychiatry than other fields.  Over the past 20 years academics and educators in the field have been subjected to the same productivity demands as clinicians.  Academic work of all kinds is devalued in order to increase the number of  patients visits focused on medications.  All incentives in place from the policy makers point toward a continued non-intellectual approach to the field.

4.  Practically all employer based positions are burn-out jobs - Reasonable people will work them for a time and then quit and ask themselves how they got involved in that situation in the first place.  The authors seem to think that better compensation or collaborative care models would increase the participation of psychiatrists in these flawed systems of care where they are "supervised" by unqualified business people.  To me the best insurance against burnout and practicing a higher standard of care is to not accept any payment arrangement that involves your work or professionalism being compromised.

5.  Public health and infrastructure needs are always neglected when it comes to psychiatry and mental health -  The most pressing issue is the dismantling of hospital structures and hospitals with therapeutic environments.  We cannot expect this to be rebuilt with the current paradigm of containment and maximizing profit by discharging people without adequate treatment.  Another way to look at the situation is that we cannot expect intellectually stimulating, state-of-the-art treatment environments when the only admission criteria is business and government defined dangerousness.  We also need therapeutic environments for the psychiatrically disabled rather than psychiatric slums and homelessness.

 The public health measures do not stop there.  America's huge appetite for addictive drugs drives a lot of psychiatric morbidity.  This offers one of the best areas for reducing the incidence of psychiatric problems and the need to see a psychiatrist.  Nobody at the policy level seems to be very interested in this problem.  Perhaps it is a resignation to the political success of the cannabis movement and more recent ideas about psychedelics being therapeutic drugs.  Reducing drug addiction and exposure would not only reduce the incidence of accidental overdoses but it would also reduce the incidence and severity of psychiatric disorders by an additional 30%.  Addictive drugs is just one aspect of prevention that is ignored by policy makers.  I would list violence and homicide prevention as a close second.

I still operate from the basic assumption that physicians are bright, well intentioned people.  That means they operate best when they have a manageable schedule, are not overworked and sleep deprived, and are allowed time for intellectual pursuits in their field.  You don't go into medicine to put in 8 hours, punch a clock and go home.  Ideally there is intellectual stimulation at work every day.  The intellectual stimulation certainly needs to be there if the psychiatrist has any involvement in teaching psychiatric residents.  It can't be there if physicians are managed like production workers especially when the product they are producing is an inferior one.

And practically every psychiatrist knows that the business-managed work product that they produce is markedly inferior to what they were trained to do and what they are capable of.  That is what fuels the private practice movement - NOT financial remuneration.

How can anyone expect to recruit and retain psychiatrists when their practice environment is actively being destroyed?  Why would anyone be interested in the field?



George Dawson, MD, DFAPA



1: Bishop TF, Seirup JK, Pincus HA, Ross JS. Population Of US Practicing Psychiatrists Declined, 2003-13, Which May Help Explain Poor Access To Mental Health Care. Health Aff (Millwood). 2016 Jul 1;35(7):1271-7. doi: 10.1377/hlthaff.2015.1643. PubMed PMID: 27385244.

A Grand DSM critique from Health Affairs

There is a large Health Affairs article that just became available online.  It criticizes (what else) the DSM 5.  The article and its initiatives all seem to flow from the conclusion:

"Inadequate interdisciplinary review and collaboration translate into missed opportunities to increase the accuracy of explanations for mental disorders.  They also lead to suboptimal care and outcome disparities for millions of patients at a time when dramatic differences in psychiatric diagnosis and treatment rates by sociodemographic status, ethnicity, and geography have undermined public confidence in psychiatry.” (p 7)

I hope that anyone reading this blog knows what the factors are in the mismatch between psychiatric diagnoses and care.  I hope that anyone reading this blog knows the biases against psychiatry and how that influences the allegations of overdiagnosis, diagnostic reliability, overprescriptions and conflict of interest that are typically leveled at psychiatrists and their professional organization.  The most obvious example and a point that seems to be completely lost on these authors is the rationing of psychiatric services and the resulting fact that most of the diagnostic disparities that they are complaining about are not due to psychiatrists or the DSM.   I hope that any reader here has also noted my running commentary about the real causes of “suboptimal care and outcome disparities”.  It is directly related to managed care, pharmacy benefit managers, and the adoption of these same rationing practices by local, state, and federal governments charged with the provision of mental health and substance abuse services.

The authors seem to lack an understanding of some of the basic social processes that they believe to be impacted by the DSM.  They cite the New York Times as a source for the issue of whether the DSM committee backed down on diagnostic revisions that would have disqualified “half of those who currently receive benefits for autism spectrum disorders” and various other changes.  As a psychiatrist who is intimately familiar with the disability process, the determination of disability is a political process at the level of the Social Security Administration.  A diagnosis is an entry point but it does not assure a disability award or even ongoing disability payments.  I have seen patients who were hospitalized for severe problems who did not get a disability determination in their favor.  I have seen people who clearly misrepresented themselves, did not believe they have a mental disability, and who received disability determinations that they requested.  As far as I can tell, the system is currently set up to favor people with mental illnesses who have been hospitalized at least three times in two years.  There are companies who facilitate applications.  It generally takes a series of two or three appeals that can drag out over a year or two.  If it comes to a hearing, those hearings are uncontested and they are not adversarial in that the government does not have an attorney present to oppose the application and the decision is made by a judge and not a jury.  The most  significant political event in this process occurred about 15 years ago when the government decided it would not consider alcoholism and drug addiction a disability.  Prior to that alcoholism was a leading cause of disability in many states.  With all of those political variables how can a DSM diagnosis be seen as the rate limiting step in that process?

The authors also conclude “Psychiatric conditions result from a combination of biological and environmental factors”.  The arguments that follow suggest that psychiatrists are basically clueless about these phenomenon.  I did not see George Engel or the biopsychosocial model of illness referenced.  In Engel's seminal 1977 paper in Science, he directly addressed the limitations of the biomedical model and changed the paradigm for the future by proposing a biopsychosocial model.  This paper is dramatic in its intellectual scope and it addresses practically all of the issues brought up in the Health Affairs article including several areas that are not addressed such as the experience of the patient.  Engel also addressed the issue of “When is grief a disease?”, a popular current DSM critique:

“…Hence the physician’s basic professional knowledge and skills must span the social, psychological, and biological for his decisions and the actions on the patient’s behalf involve all three.  Is the patient suffering normal grief or melancholia?  Are the fatigue and weakness of the woman who recently lost her husband conversion symptoms, psychophysiological reactions, manifestations of a somatic disorder, or a combination of these.  The patient soliciting the aid of a physician must have confidence that the MD degree has indeed rendered that physician competent to make such differentiations.”  

A reference to Engel would seem appropriate but it detracts from the authors’ contentions that physicians seem to need to have their biopsychosocial horizons broadened and acknowledging that a physician discussed this definitively 35 years ago would detract from their argument.

The authors more direct arguments about the role of “social and institutional influences on diagnosis” can be similarly addressed.  Although they don’t acknowledge the DSM, they discuss post traumatic stress disorder as an example of environmental exposure.  They cite evidence gathered in the psychiatric literature as their proof.  In fact, any psychiatric evaluation should contain a formulation section that considers social, biological, and consciousness based factors in the overall evaluation of the person seeking help.  This is nothing new and every competent psychiatrist is trained to do this.  The now abandoned oral Board exam, used to test these skills.  The idea that these factors are relevant to psychiatric diagnosis have been taught to psychiatrists for decades.  Do we really need to learn that from a panel of social experts who don't talk with people about that information every day like we do?

The idea that social context,  is a relevant factor has also been obvious to psychiatrists for a long time.  Psychiatrists are routinely asked to evaluate and treat patients from various socioeconomic and cultural groups and frequently work with interpreters in the process.  There is no basis in fact for their speculative comment that “Identifying and understanding the causes of diagnostic disparities can lead to improved diagnostic criteria and their more accurate application.”

On the issue of institutional and policy factors the authors also miss the mark.  They make the previous mistake about diagnosis and Social Security disability by suggesting that a specific diagnosis results in a disability check.  They do not point out how the Social Security process rather than a DSM diagnosis may be more important in the issue of disabilities for mental health. Interestingly they are concerned about the “major consequences for payers and patients" and reference a study looking at the prescription of atypical antipsychotic medications for children.  They ignore the fact that the actual treatment of mental illnesses are outside of the purview of the DSM and that overprescription (if this is actually overprescription) is a widespread problem that extends well beyond the field of psychiatry.  As is the case with all critics of psychiatry and the DSM, they give a pass to the real causes of systemic poor treatment and a focus on medications rather than psychosocial therapies and that is the managed care industry and its supporters at all levels in the government.

Their final focus on publicity and marketing is certainly not a problem specific to psychiatry.  It is also a process that is not DSM dependent.  Restless leg syndrome or insomnia do not need to be in the DSM to end up being treated on a large scale by primary care physicians.  All it takes is a pharmaceutical company web site with a checklist.  They provide no insight into why the political process of direct-to-consumer advertising as determined by lobbyists, politicians, and the associated exchange of money should be part of a DSM oversight process.

The authors proposed Psychiatric Diagnosis Review Body and its potential benefits are equally speculative.  Their idea that there would be “greater sophistication” in the explanations of mental illness is doubtful, especially considering the impact that Engel’s biopsychosocial model has had on both the field and DSM development.  Their idea that the work of a review body would “heighten mental health practitioners’ awareness of population level differences in diagnoses, in some instances improving their ability to tailor diagnoses to patient’s demographic characteristics and cultural backgrounds…” is also problematic.  First off, the DSM is written for psychiatrists and a psychiatric diagnosis and formulation is much more than looking at a list of symptoms that possibly identifies a person as being a statistical outlier in a group.  Any person can pick up a copy of the DSM and presume to make a "diagnosis" based on these criteria, but that is not a psychiatric diagnosis.  Secondly, cultural, demographic characteristics, and demographic factors have already been incorporated into psychiatric evaluations for decades.  An even greater question is what broad scale social data would add to the evaluation of the individual patient given the biases that are usually present in those studies.

The authors suggest that the incorporation of feedback from the review body would “increase public confidence in the manual and psychiatry as a medical profession”.  The single most important factor that would enhance psychiatry’s image would be the recognition that rhetorical negative arguments against the profession abound and need to be corrected.  That could start by recognizing what psychiatrists actually do and what a DSM is actually used for.  It would also take a critical look at why 20 years of rationing of psychiatric services by the managed care industry and the government is the single largest factor in why these services have deteriorated and now operate on the premise that getting people on one medication or another is the best way to treat mental illness.  The authors in this case banter about million and billion dollar amounts that are typically used to suggest the impact of the DSM or significant conflicts of interest in psychiatry.  Nobody is focused on the fact that the managed care industry makes far more money than that by denying medical care.  Psychiatric services make up a disproportionately large amount of denied care.

If you are really interested in improving the care of people with mental illness in this country it would seem logical to attack those who routinely deny them care and interfere at all levels with the provision of care rather than those providing the care and trying to improve it.   That is the most important social problem affecting the provision of mental health services and access to psychiatry.  Social scientists seem to be as disinterested in that fact as the average journalist.

George Dawson, MD, DFAPA

Hansen HB, Donaldson Z, Link BG, Bearman PS, Hopper K, Bates LM, Cheslack-Postava K, Harper K, Holmes SM, Lovasi G, Springer KW, Teitler JO.  Independent Review Of Social And Population Variation In Mental Health Could Improve Diagnosis In DSM Revisions. Health Aff (Millwood). 2013 Apr 24. [Epub ahead of print] PubMed PMID: 23614899.

Engel G. The need for a new medical model: a challenge for biomedicine. Science 1977; 196:129-136.

George L. Engel, MD. JAMA.2000;283(21):2857. doi:10.1001/jama.283.21.2857