The Balanced Rhetoric Against Neuroscience

The New York Times editorial pages continue to be a place where anti-neuroscience rhetoric can be expressed primarily as decreased funding or more accurately portion of the available NIMH funding.  Maybe there has been some pro-neuroscience opinion expressed there and if there was I have missed it.  I recently posted an exciting development in neuroscience teaching for psychiatrists and psychiatric residents.  In that post I reference an opinion piece by Richard Friedman, MD a psychiatrist (1).  Dr. Friedman makes several arguments for psychotherapy as if it is unrelated to neuroscience and based on that premise concludes that there is no substitute for psychotherapy, that people are more than a brain in a jar, and that anyone benefiting from psychotherapy seems to prove  that.  I found that to be an incredible statement considering that (according to Koch in above graphic):  "The brain is the single most complex object in the universe." There is also the fact that with 7.4 billion people on earth - there are 7.4 billion unique conscious states - the vast majority of which are not accurately described by any DSM or psychodynamic diagnosis/formulation.  All the time that Dr. Friedman is mounting this critique he also discusses the importance of clinical research and suggests shifting the funding balance away from neuroscience.

In the recent case John C. Markowitz a professor of clinical psychiatry at Columbia has a more subtle form of the argument.  In this case and the previous opinion piece the authors both endorse the importance of neuroscience to a point.  In this case the argument is - yes neuroscience is important but let's reestablish balance between neuroscience and clinical studies such as looking at the efficacy of psychotherapies.  Breaking it down, Dr. Markowitz makes the following points:

 1.  Under the directorship of Thomas Insel, the NIMH clinical research budget was "strangled" and the resources were diverted to neuroscience research.  The author acknowledges both the need for neuroscience research and the primitive stage of psychiatric diagnostics based on clusters of signs and symptoms.  This was really the basis for Insel's RDoC initiative looking at more reliable markers of psychiatric syndromes.  Any practicing psychiatrist who has seen all of the iterations of the DSM realizes that we are as far as we can go with this manual.  That includes from the standpoint of validity but also in terms of the clinical examination by psychiatrists.  As long as we are all contained by this manual, the clinical method of psychiatry will remain stuck somewhere in the 1940s.  That should be extremely disconcerting to the profession and future psychiatrists.

DSM technology is extremely limiting in terms of the usual clinical trials.  The NIMH sponsored Star*D study is a decade and a half old at this point.  It has defined the response rates for both antidepressant therapies and provided a discussion point for psychotherapy trials of depression.  Clinical trials of antidepressants provide an equally varied result.  Any practicing clinician knows that these studies are all seriously flawed out of the gate by using DSM diagnoses and also an intent-to-treat analysis that does not resemble clinical practice.  The variation in diagnoses from depression to anxiety to depression plus anxiety as seen in clinical practice should point to the fact, that patient selection into clinical trials currently results in very heterogenous patient populations in terms of both therapeutic effects and medication tolerability.  We can continue to spend large sums of money on these trials of mixtures of patient populations and post modest positive results or we can attempt to identify patients who will respond specifically and not experience side effects from a particular therapy.  That is the real promise of neuroscience based research.

2.  The patients who need help are poorly served by current neuroscience research.  The helpful psychotherapies listed by the author like interpersonal psychotherapy (IPT), cognitive behavioral therapy (CBT), and other psychotherapies have been around for decades.  I happen to have copies of Interpersonal Psychotherapy by Klerman, Weissman, Rounsaville, and Chevron and Cognitive Therapy of Depression by Beck, Rush, Shaw, and Emery.  The publication date of the former is 1984 and the latter is 1979.  Both therapies have been out there for over 30 years.  At this point both have been studied hundreds of times.   Looking at clinical trials on Medline yields 1711 for CBT and 261 for IPT.  Not only that but some of the clinical trials that were successful (like IPT for cocaine use) have never made  it into clinical practice.  In fact, in most places getting a therapist who actually practices any of the specific research proven psychotherapies is impossible.  The problem does not seem to be a lack of psychotherapy research but a lack of access to practitioners who use research proven psychotherapies.  Mental health treatment is the most highly rationed treatment resource and additional studies that continue to prove that existing psychotherapies work seems superfluous at this point.  Any current studies are often compared to existing therapies and with the DSM problem contributing to diagnostic heterogeneity.  Any new trials should only be funded for serious conditions where the therapy might be useful.  There is no reason to expect that a new therapy applied using the current diagnostic system or clinical trials technology will lead to any enhanced treatment effects.

3.  Existing treatments are not "good enough".  The author attributes this "good enough" statement to Insel himself.  I understand the point he is trying to make.  The author points to continued suffering, treatment failures and suicides as evidence that more is needed now.  The problem is that there is no assurance that clinical research will add any more at this time.  Certainly a focus on suicide as a stand alone problem (not suggested at all by DSM) and on serious disorders with no treatment like adult anorexia nervosa is warranted.  But even then we are left with a clinical trials technology that consistently produces modest results at best.  More multimillion dollar trials of psychotherapy that we already know is somewhat effective when patients have no chance of ever receiving it against a backdrop of "is this really depression or anxiety" seems like a waste of time and money to me.  It seems like a much better idea to develop a neuroscience method to determine who needs psychotherapy and who might benefit from medications.  But even then, the only treatments that will be readily available will be the medications and even then less than half of the affected patients will get access to treatment.  Good luck trying to find a psychotherapist and an insurance company willing to cover the cost of the number of sessions used in the psychotherapy research. Research proven therapies are only as good as the number of practitioners using them and access to those practitioners.

4.  The placing all of your eggs in one basket argument.  This is basically saying that if the ratio of clinical to neuroscience funding is 10% to 90% the risk is missing something big in the clinical research and not getting any useful results from neuroscience.  Given the history that I have provided, there needs to be a clear advance on the clinical side in order to fund large trials.  It does not make any sense to continue to  fund more of the same  or slight modifications of treatment for common disorders.  Our eggs have been all in one basket and I would call that treatment as usual.  In the 30 years that I have been in practice, there is nothing that I would call a major breakthrough.  Clinical research results come and go.  Effective psychiatrists are effective psychiatrists not based on breakthroughs but how they approach clinical practice.  Even that mode of treatment is threatened by widespread support for "collaborative care" that is being justified using the same kind of research that justified managed care in the first place.  In the end there has been nothing more destructive in terms of access to care for mental disorders than managed care.

In many ways these ongoing arguments resemble the arguments of the biological psychiatrists and psychotherapy psychiatrists that I trained under in the 1980s.  Many programs were split under this artificial division with the residents left to identify with biological or psychotherapy faculty.  It is interesting to note that this division occurred at a time when Kandel wrote a paper on how psychotherapy is neuroscience in action (3).  That may have been missed because the biologically based psychiatrists at the time were really focused on pharmacology and neuroendocrinology rather than a comprehensive neuroscience.  Neuroscience and the old diagnostic technology and clinical methods seem to be the current points of division.

A lot of the criticism is directed at Insel.  I have heard him talk about the initiatives and the rationale sounded clear to me.  I think that rationale is very similar to what I have discussed so far, but for clinical psychiatrists it is also the realization that as long as we live in an approximate world - we will get approximate results.  The inertia to stay in that place is always puzzling to me.

But - it is time to move out of the 1950s.

Clinical psychiatry the way it is currently researched and practiced holds no promise for understanding the most complex known object in the universe.  Neuroscience is one of the big ways out of that predicament.



George Dawson, MD, DFAPA      



References:

1:  Friedman RA. Psychiatry's Identity Crisis. New York Times July 17, 2015. p SR5.

2:  Markowitz JC.  There’s Such a Thing as Too Much Neuroscience.  New York Times October 14, 2016. p A21.

3:  Kandel ER. Psychotherapy and the single synapse. The impact of psychiatric thought on neurobiologic research. N Engl J Med. 1979 Nov 8;301(19):1028-37. PubMed PMID: 40128.

              

National Neuroscience Curriculum Initiative - a brighter future for psychiatry

As any reader of this blog probably knows, I am a big proponent of neuroscience education for psychiatrists and always have been.  I have suggested in the past that it would take broad collaboration.  I posted some examples of an NIMH initiative on neuroscience  education.  I teach neuroscience (also known as neurobiology) myself and that has led me to be acutely aware of the lack of educational resources on the field.  That background is what has made this the happiest day at a CME conference that I have ever spent.  I am currently at the University of Wisconsin 4th Annual Update and Advances In Psychiatry - a conference that has really been in place for the past 41 years.  After watching a comprehensive update on eating disorders I settled in to listen to Melissa Arbuckle, MD, PhD; Professor of Clinical Psychiatry, Director of Residency Training; Department of Psychiatry, Columbia University Medical Center.  The title of her presentation was Discussing the Neuroscience of Mental Illness with Your Patients.  I turned to the section in the syllabus and there was one page with a case report on the front and a crude drawing of the brain (two views) on the back.  Being a traditional conference guy who likes a ton of technical information, no audience participation, and no role playing - I was prepared to be disappointed.

I was not prepared for what would transpire in the next 90 minutes.  I have posted here many times why I thought that every psychiatrist should know neuroscience and ways to do it, specifically the need for widespread collaboration due to a lack pf neuroscience manpower in most departments.  Dr.  Arbuckle started out explaining what the National Neuroscience Curriculum Initiative was.  It was started by a collaboration of like minded residency directors  to come up with a program to teach neuroscience to psychiatry residents.  She showed the explosion in neuroscience papers in psychiatry just over the past decade.  She referred to an article in JAMA Psychiatry (1) with her collaborators on why neuroscience needs to be integrated into psychiatry right now.  She discussed the New York Times editorial that showed up three weeks later with the criticism of their viewpoint (2).  Although she did not mention it, like a lot of articles it as written from the perspective of a psychiatric identity crisis.  Whenever I see that term it seems like the authors are firmly behind the curve and don't seem to understand what neuroscience encompasses.  Dr. Arbuckle said that the article was critical of the criticism that the brain was the basis of human behavior.  Quoting the article:

"Indeed an article in May in one of the most respected journals in our field JAMA Psychiatry echoed this view: "The diseases we treat are diseases of the brain."........ Even if this premise were true - and many would consider it reductionist and simplistic - an undertaking as ambitious as unraveling the function of the brain would likely take many years."  The author is a psychiatrist and goes on to say that he is all for neuroscience and even talks about some recent research techniques he (implicitly) just doesn't think psychiatrists should study it?  He also seems to conflate psychotherapy as being independent of neuroscience when in fact we have known just the opposite since since Kandel's 1979 seminal lecture Psychotherapy And The Single Synapse.  

I am equally incredulous when people seem to argue about the importance of neuroscience in psychiatry.  I find reductionism and reductionist approaches to be perfectly understandable and acceptable.  It is an interesting form of rhetoric to use these terms pejoratively.  Most people go into medicine because they want to know how things work.  If they did not enter with that goal, it soon becomes apparent that knowing mechanisms whether they are theoretical or not is an important aspect of studying medicine.  Some of the first mechanisms I studied in medical school involved cholera and diphtheria toxin. How is it possible to determine these mechanisms and recent significant epigenetic mechanisms without taking a reductionist approach?

The exercise that Dr. Arbuckle introduced to the audience was the diagnosis and treatment of complex cases.  The case vignette involved a young woman with borderline personality disorder.  The task for the audience was to pair up and role play discussing the relevant neuroscience concepts of treatment with the patient using the brain diagram as an aid.  Eliciting responses from 300 people in a room slows things down.  After the audience was done, she showed a film of an expert presenting this information to a patient and what presentation materials might be available.  It went very well and it presented the rationale for dialectical behavior therapy and not a medication.  It was a clear example of a neuroscience based discussion that provides a rationale for psychotherapy.  There are numerous materials on the web site (9 modules, 56 sessions, 40 authors) and wide scale participation is encouraged.

The information up to that point was quite exciting.  Dr Arbuckle had plenty of enthusiasm in her closing remarks.  In those remarks she pointed out the goal of "getting the entire field up to speed" in neuroscience.  She pointed out that everything on the site is free but at some point they may ask physicians to pay for CME.  She said that she realizes that this is literally "changing the world and that is what we are going to do."        

This was the most exciting commentary from a psychiatrist about teaching the entire field and the future of psychiatry that I have ever heard.  I have never been this impressed by any development in the field during my career.  And I am a psychiatrist who is as pro-neuroscience as anybody.  How is it that I am just hearing about this initiative right now and only because I am attending a conference?  That is why I am posting my experience here and a link to the NNCI web site and materials.

Dr. Arbuckle and her collaborators are one of the few bright spots for the future of psychiatry.

But they are very bright.


George Dawson, MD, DFAPA


References:


1:  Ross DA, Travis MJ, Arbuckle MR. The future of psychiatry as clinical neuroscience: why not now? JAMA Psychiatry. 2015 May;72(5):413-4. doi: 10.1001/jamapsychiatry.2014.3199. PubMed PMID: 25760896.

2: Friedman RA.  Psychiatry's Identity Crisis.  New York Times July 17, 2015. p SR5.


Attribution: 

The graphics at the top are two slides from one of my lectures.  I like to present data on the unique aspects of every individual brain and why that can happen.  The slides are not from the NNCI program and I am not affiliated with that program.  Click on each slide to enlarge.

Public Service Announcement - Moderation Policy

This is just a reminder and to establish an anchor point about moderation on this blog.  If you attempted to post a response on the blog and it did not show up it is most likely due to one of these reasons.  It is always possible that I made a mistake or that your response ended up in my spam file.  In fact, several posters clued me in about that happening and I did find and post their responses.  It should be pretty clear that you do not have to agree with me but there are some limits that are fairly basic to a small barely read blog like this one.  These limits follow:

1.  No trolls - I am not fascinated by trolls or the troll culture.  One of the reasons that I started this blog was that I used to post in what I thought was a professional forum that was dominated by a troll.  There was no moderation and eventually all of the interesting posters left and went elsewhere.  So spare me the First Amendment arguments - that kind of rhetoric just doesn't fly here.  My reasonable test is what would happen to a troll in any medical staff meeting?  Consider this a similar atmosphere.

2.  No advertising - This is a strictly non-commercial enterprise.  I type it up on my own and there is no sponsor or source of revenue.  As I have pointed out to people who want to advertise here, I have many copyright permissions at this point that are all contingent upon the non-commercial status of this blog so at no point is it possible for me to accept advertising.  The Creative Commons license allows anyone to repost the contents of this blog with the proper attribution.  If you like the content - just repost it or link to it.  No advertising also includes posts with hyperlinks to product advertising and that kind of link results in a rejected post.  It also includes posts with no content and apparent compliments that are hyperlinked to advertising through the name of the poster.

3.  No anti-psychiatry rhetoric - There can be discussions of anti-psychiatry from a philosophical standpoint, but if the main point of the post is to bash psychiatry or psychiatrists this is not the place and you probably already know that.  There are many anti-psychiatry blogs that you can flock to.  Some of them are even contain active posts by psychiatrists.  Psychiatry attracts a lot of haters and my original analyses of the problem in 2012 still stands as well as a clear anti-psychiatry bias in the media.  At least part of that media bias is that they are just being provocative to draw a crowd and of course they have no responsibility to people with clear problems.

4.  No personal attacks - ad hominem is the poorest form of criticism.  If the argument cannot be addressed, attacking the person is not a substitute - at least here.  There are many other blogs and media sites where that is acceptable.  The standard I use for physicians is the expected behavior in professional settings with colleagues.  If a post does not meet that standard it is on shaky ground.

5.  No medical advice - this blog is not the place to seek personal medical advice.  Nothing here should be construed as medical advice or a discussion of an actual patient.  My opinion stated here in many places is that only your personal physician or physicians knows enough about you to make suitable recommendations.

6.  No interminable arguments - I don't have to have the last word on anything and frequently just stop posting.  I will not post repeated posts making the same argument or with very slight modifications.  One of the most interminable arguments has to do with the link between psychiatrists and pharmaceutical companies.  I have debunked that argument on this blog many times.  I would refer you to blogs where they have interminable arguments about how bad psychiatric medications are and how corrupt psychiatrists are for their connections to Big Pharma.  Neither argument is accurate but this is the place to read about how medications are actually used.

7.  Stay on topic - It is a reasonable requirement to address the topic.  Newspaper web sites are good examples of what can happen when this simple rule is not followed.  A corollary is to address the topic in a timely manner.  I have had people notice posts from 3 years ago and write a response.  Blogger does not allow me to terminate discussions like some web sites.  My only option is to not post untimely responses.  Six months is a reasonable time frame.

That is what I have so far.  I reserve the right to add more as they come to me.  The goal is to maintain coherence and rational discussion.  Any inspection of the blog posts shows that there is not a lot of discussion and stimulating discussion is not one of my primary objectives. I appreciate the people who have taken the time to read this blog and post their comments - many of which make excellent points and do stimulate further discussion.

My goal all along has been to produce opinion and analysis consistent with the way that real psychiatrists train, work, and practice.  I do not see myself as unique in any way.  I routinely have contact with excellent psychiatrists who I have trained with and who are colleagues.  They approach problems in psychiatry and think about those problems in very similar if not identical ways that I do.    I see this blog as a creative outlet as well as one of the few places on the Internet where this kind of content is available.


George Dawson, MD, DFAPA




Attribution:


The photo:  John Schneider.  RCA 40A Ribbon Microphone. August 19, 2007.
Attribution-NonCommercial 2.0 Generic (CC BY-NC 2.0)

Big Data - What Is It Good For?

Big Data and Data Science have been buzzwords in science and industry for over a decade.  A Medline search shows over a thousand current references to Big Data in healthcare.  A good starting point is consider what is meant by Big Data and then discuss the implications.  A quick scan of the references shows that they vary greatly in technical complexity.  A standard definition from Google is: "extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions."  These techniques developed because of the widespread availability of digitized data and the ease with which sets of behavioral choices (in the form of mouse clicks) can be collected on web sites.  In many cases specific data collection paradigms can be used to elicit the information, but there is also a wealth of static data out there as well.  In health care, any electronic health record is a massive source of static data. Financial, real estate, educational records and records of all of these transactions are also a significant source.

Most  Americans logging in to set up a Social Security account online (ssa.gov)  in the past couple of years would be surprised at what it takes to complete the job.  After the preliminary information there is a set of 5 security questions.  Four of those questions are about your detailed personal credit history - home mortgage information and credit card history.  When Social Security was initially set up in 1936 there was widespread concern that Social Security Numbers would become national identifiers.  At one point Congress had to assure the electorate that the number would not be used for that purpose.  Since then the SSN has been used for multiple identification purposes including credit reporting.  At this point it seems that we have come full circle.  Congress invented the SSN and told people it would not be an identifier.  They mandated its use as an identifier. Congress authorized and basically invented the credit reporting system in the United States.  The federal government currently uses the credit reporting system to quiz taxpayers wanting to set up a Social Security account online.  In the meantime, large amounts of financial, legal, and health care data are being collected about you under your SSN in data systems everywhere.  At this point the full amount of that data and the reasons why it is being collected for any person in the US is unknown because it is all collected without your knowledge or your consent.  It is impossible to "opt out" from this data collection.  The federal government does have an initiative to remove SSNs from health records, but there are so many other identifiers out there right now, this effort is too little and too late.

Additional sources of data include your online foot print including sites that you may have visited and what you seem to be interested in.  A visit to Amazon for example and a quick look at an expensive digital camera may result in that same camera with a link to Amazon in the margin of very other web page you see for the next two weeks.  Expensive digital cameras of a different brand than the one you originally looked at may start showing up.  You may notice product ads showing up in your Facebook feed that you mentioned casually to your friends during a conversation there.  The conversation could be as generic as bicycle seats and suddenly you are seeing a flurry of ads for bicycle seats.  Any number of web sites encourage to sign in with other accounts and then share your account information with them.  All of this data provides companies with what they need to fuel their predictive algorithms to sell you a product.  It provides the major advertisers in this space like Google and Facebook with a huge revenue source because based on the scale and personalization of these ads - they are effective.  Big Data seems to be very good for business.  But is there a downside?        

That brings me to a current resource on the nefarious uses of Big Data written by an expert in the field.  Cathy O'Neil is a PhD in mathematics.  Her PhD work was in algebraic number theory.  She started work as an academician but subsequently worked for a hedge fund, work as a data scientist for several firms and currently heads the Lede Program in Data Journalism at Columbia University.  I am familiar with her work through her blog MathBabe.  Her newly released book Weapons of Math Destruction takes a look at the dark side of Big Data specifically how data collection and biased algorithms can be good for administrators, politicians and business but bad for anyone who falls under the influence of those agencies and their work.  In the introduction she leads of with the example of teacher assessments.  I was familiar with a scattergram that she had posted on her web site showing that year to year teacher assessment scores were essentially uncorrelated or random.  In the book she describes the human toll in this case a teacher fired because of this defective algorithm.  In another example later in the book, an experienced teacher scored a 6 out of 100 on a "value-added" teacher evaluation.  Only tenure kept him from getting fired.  The scoring algorithm was opaque and nobody could tell him what had happened.  The next year he scored the 96 out of 100.  But the algorithm was so flawed he knew that score was no more legitimate than the last one.  With the politicized environment surrounding teaching the proponents of teacher "accountability" like this variation since it fits their ideas about the system retaining incompetent teachers that need to be weeded out.  In fact, the algorithm is defective and like many is based on erroneous assumptions.

I personally know that physicians are subjected to the same processes as teachers, but so far it is less technologically advanced.  O'Neill points out that there is nothing magical about algorithms.  That they frequently incorporate the biases of the people who design and contract for them.  Opacity and a lack of correction by feedback is another feature.  I worked for the same employer for a number of years when physician "accountability" measures were put in place.  The "algorithm" for salary went something like this RVU Productivity + Outside Billing + Citizenship = Pay.  RVUs were the total number of patients seen according the the biased government and managed care billing schemes.  Outside billing was any consulting work done outside of the clinical work that was billed through the department.  Citizenship included teaching and administrative duties as well as any Grand Rounds or CME lectures that were done.  In other words apart from the subjectively based billing scheme all of the inputs are almost totally subjective and influenced by all kinds of pseudoaccountability measures along the way.  For example, in parallel with the teacher ranked on the algorithm, I was told one year that I had achieved the top rank in terms of documentation in a group of about 25 physicians.  The next year - making no changes at all in terms of that documentation - I was dead last.  My conclusion, like the teacher in the example was that the rating scheme was completely bogus and with that kind of a scheme who cares about the results?

The number of based algorithms applied to physicians has eerie parallels to those mentioned in WMD.  Here are a few that I picked out on the first read:

1.  The algorithm is based on faulty data - the teacher evaluation algorithms were based on a faulty interpretation of data in the Nation at Risk report.  The report concluded that teachers were responsible for declining SAT scores between 1963 and 1980.  When Sandia Labs reanalyzed the data 7 years later they found that an great expansion in the number of people taking the test was responsible for decreased average score but subgroup analysis by income group showed improved scores for each group (p. 136).  The only reason that teachers are still being blamed is political convention.  I posted here several years ago that the top ranked students in the world in Finland are taught by teachers who are assumed to be professionals and who are not critiqued on test results.

The parallel in medicine was the entire reason that medicine is currently managed by the government and the healthcare industry.  It was based on criticism in the 1980s that doctors were lining their pockets by performing unnecessary procedures and that work quality was poor.  That should sound familiar because that criticism has been carried forward despite a major study that showed it was completely wrong.  The massive Peer Review Standards Organizations (PRSO) in each state in the 1990s conducted rigorous reviews of all Medicare hospitalizations and concluded that there was so little overutilization and so few quality problems that it would not pay to continue the program.  The only reason that managed care companies exist today is by political convention.

2.  An effective teacher like an effective doctor is too complex to model - When that happens only indirect measures or "crude proxies" (p. 208) can be used to estimate effectiveness.  In medicine like teaching - the proxy measures are incredibly crude.  They generally depend on diagnosis, poorly account for comorbid illness, and the outcome measures are heavily influenced by business rather than medical decision making.  The best examples are length of stay parameter and readmission parameters.  Every physician knows that there are set payment schedules based on the supposed ideal length of stay for a particular illness.  The business influence in the discharge decision is so malignant these days that non-physician case managers are present to pressure physicians into discharging patients.  If the discharge beats the length of stay parameter - the hospital makes money.  I sat in a meeting at one point and asked the obvious question: "OK - we have completed the discharge checklist - do we know the outcomes?  How do the patients do when they are discharged by this process?  How many of them die?"  Dead silence followed.  Most people would be shocked to hear that what passes for evidence based medicine is often a checklist that has no meaning in the real world.  Making the points on the checklist is good for advertising though.

3.  There is a lack of transparency in the overall process - The teachers in WMD who were blindsided by the algorithm were never told how that conclusion was reached.  I encountered the same problem in a managed care organization when it was clear to me that administrators with no knowledge of psychiatry were telling us what to do.  In some cases, "consultants" were brought in to write reports to confirm the most recent administrative edicts.  When I asked my boss if I could talk with the people sending out the edicts I was informed that there was a "firewall" between clinicians and upper management.  This lack of feedback is another critical dimension of algorithms gone astray.  If you are writing an algorithm biased toward a business goal - why would you want feedback from clinicians?  Why would you want any humanity or clinical judgment added especially in the case of psychiatric care?  Let's just have a dangerousness algorithm and leave it at that.  Those are the only people who get acute treatment, even though it is patently unfair relative to how the rest of medicine works.             

Big Data is good for science.  We can't do elementary particle physics or genomic analysis very well without it.  Big Data is also good for business is much different ways.  There are clearly people out there who cannot resist buying items online if the Amazon algorithm shows it to them enough times across a number of web pages.  Big data in business can also come up with billing algorithms that have less to do with reality than making a profit.  Similar programs can be found for employee scheduling, performance analysis, and downsizing.  The problems happens when the business biases of Big Data are introduced to science and medicine.  Those techniques are responsible for an array of pseudoquality and pseudoaccountability measures for physicians, hospitals, and clinics.

Unfortunately physicians seem to have given up to the political conventions that have been put upon us.  Some administrator somewhere suggests that quality care now depends on a patient portal into an electronic health record and a certain number of emails sent by patient to their physician every month.  Across the country that will result in hundreds of millions of emails to physicians who are already burned out creating highly stylized documentation that is used only for billing purposes.  Terabytes of useless information that nobody will ever read again - the product of a totally subjective billing and coding process that started over two decades ago.  Is there any data that email communication is tied to the effectiveness or technical expertise of the physician?  I doubt it.  I worked with great physicians long before email existed.

It is about time that somebody pointed out these manipulations provide plenty of leverage for the management class in this country at the expense of everyone else.  It is well past the time that doctors should be confronting this charade.  


George Dawson, MD, DFAPA




References:

1.  Cathy O'Neill.  Weapons of Math Destruction - How Big Data Increases Inequality And Threatens Democracy.  Crown Publishing Group.  New York, NY, 2016.  I highly recommend this book for a look at the other side of Big Data.  It is written in non-technical language and is very readable.

Attribution:

The photo at the top is a Server Room in CERN By Florian Hirzinger - www.fh-ap.com (Own work (Florian Hirzinger)) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0) or GFDL (http://www.gnu.org/copyleft/fdl.html)], via Wikimedia Commons"   href="https://commons.wikimedia.org/wiki/File%3ACERN_Server_03.jpg"><img width="512" alt="CERN Server 03" src="https://upload.wikimedia.org/wikipedia/commons/thumb/d/d7/CERN_Server_03.jpg/512px-CERN_Server_03.jpg"/></a>

  

Psychosis Idealized

I thought I would provide a counterpoint to the New York Times editorial entitled "Medicating A Prophet" written by Psychiatrist Irene Hurford (1).  The opinion piece is available free online and I encourage anyone interested in the topic to read the article rather than accepting my summary here as adequate.  I will say from the outset that I am not a stranger to any of the issues that Dr. Hurford discusses either clinically or personally.  The bulk of my career was spent treating people with severe mental illnesses and addictions.  Once you have worked in that setting, it is clear that many people who are severely ill need involuntary treatment and that is one of the decision points that she addresses.

In her essay, Dr. Hurford describes an early call experience during her residency.  She was asked to assess a man in the emergency department (ED) who had been delusional for 30 years.  The delusions were religious and grandiose in nature.  He was a college graduate but was homeless living on the street in Philadelphia.  He also had AIDS and the complication Kaposi sarcoma.  His reason for being in the ED was "to preach".  Dr. Hurford encourages him to come in for voluntary treatment but he refuses.  At that point she ponders involuntary treatment but in the essay decides to discuss the patient's right to psychosis. Later we learn that she made the decision but has decided to analyze that decision in retrospect based on factors that she has encountered since.

One of those factors was the influence of a professional colleague who based on her own experience with psychosis and that colleague's mother's experience suggested that thoughts about living "in psychosis" and outside of psychosis need to be challenged.  She basically states that the problem may be within the beholder rather than the identified patient. Following that logic, it makes sense to show up in an ED to preach while ignoring serious health problems.  It also makes sense to make decisions about the person's "in psychosis" experience knowing so little about them.  In my experience, nobody in the ED calls the parents or family of a 50 year old street person, to get a clear picture of how the psychosis has truly affected him.  When I have treated these people on an inpatient unit and made those calls, I have never heard that the patient was well served by the psychosis.  Not a single time.  In many cases, family members were surprised to hear that person was still alive.

Dr. Hurford advances a number of other arguments that I call into question.  She uses a very loose definition of insight as a "failure to accept an alternative view of reality".  She turns this around to suggest that anyone who does not accept this premise (implicitly the treating physician) also lacks insight.  I don't think that you can practice psychiatry and not be comfortable with alternative realities.  I would suggest a more appropriate definition of insight as a decision-making process.  Can I accurately assess how I am doing in the world?  Am I making decisions in my best interest?  Are those decisions consistent with my ability to survive?  If I realize that I am not doing well can I get help?  Pretty basic decisions.  Not a question of lifestyle choices.  To have a lifestyle you have to live.  That is the kind of insight that I am used to dealing with.

Dr. Hurford discusses a case of a young patient with a psychotic disorder who stopped taking his medications and started using cannabis on  daily basis.  He dropped out of college and became progressively incoherent and then mute.  She is concerned about traumatizing the patient by "enforcing" treatment even though he cannot "eat, sleep, and talk."  I don't follow the logic that some treatment intervention - even basic detoxification from cannabis is somehow more traumatic than not eating, sleeping, or being able to communicate.  How is that a preferred alternative existence?

At that point she digresses to a very brief overview of the usual comments about mental illness being only peripherally associated with violence and the lack of evidence that forced treatment led to fewer hospitalizations, arrests, or a better quality of life.  She cites a meta-analysis of three randomized-controlled studies of more than 700 people.  There are a lot of reasons why meta-analyses are not superior to the actual trial data.  There are also a lot of reasons why truly clinical samples with these problems cannot be ethically randomized or included in the studies.  There are also reasons why I would expect the entry points into these studies to be highly variable as well as the treatment resources that are involved.  In Minnesota, we have 87 counties and the rule is that there are 87 interpretations of the commitment act for involuntary treatment.  There are two corollaries operating here.  The first is that the courts will be very liberal in terms of dropping commitments until something bad happens.  At that point the pendulum swings back in the direction of more frequent commitment.  The second is that only the wealthiest counties in the state can afford to provide adequate resources to treat the severely mentally ill.  Even then there is no assurance that the counties that can afford it will actually provide the care.  Some currently function like managed care companies and ration the care.  They can end up rationing care and commitments in order to save the county money.  The lack of evidence that forced care does anything may be more of an indictment of the lack of quality or consistency in delivering care and interpreting the law  and rationing care more than anything.  I have personally treated many times the number of people with forced treatment than in the meta-analysis and there is no doubt that the outcomes were better than with no treatment.

The outcome variables cited by Dr. Hurford are also dreadfully lacking compared with what can be seen routinely in clinical settings.  They include very adverse outcomes in encounters with the police including getting shot, dying from a treatable illness, suicide, loss of relationships with spouses and children, loss of a job and income, and acute loss of life due to poor insight and judgment.  In Minnesota, all that takes is going outside in the winter time without adequate protective clothing and you are dead or in the Burn Unit with frostbite.
     
Right now we are in the midst of a sweeping cultural change that idealizes psychosis and suggests that hallucinogens and cannabis are therapeutic drugs.  That will put the next generation or two of people with psychoses, mood disorders, and substance use disorders at risk for chronicity and every possible negative outcome.  A point that should not be lost on anyone is how no care for psychosis is "cost-effective" care when the total impact on the patient is ignored.  My point in writing this rebuttal is really advice for the people in these generations.  Ask any psychiatrist treating you or your family member where they stand on this issue.  

Especially if you value psychotic symptoms a lot less than your psychiatrist does.



George Dawson, MD, DFAPA





References:

1.  Irene Hurford.  Medicating A Prophet.  New York Times.  October 1, 2016.

Cancer Care In Psychiatry - Yes It Happens

Great Case Records of the Massachusetts General Hospital in the New England Journal of Medicine this week.  A psychiatrist is presenting a case of complicated disorder and coordinating care with oncologists after a cancer diagnosis is made.  It must have taken an editorial change in the Journal to get psychiatry more front and center in this prominent medical journal.  This article has a lot of meaning for me, because the bulk of my career was spent on these issues.  When you are the inpatient doc or the psychiatrist staffing the case management teams - either Assertive Community Treatment (ACT) or some other case management model a lot of this important work falls to you because there is nobody else out there.  Contrary to the popular hype about collaborative care - people with severe mental illnesses generally avoid medical clinics and non-psychiatric physicians.  There are always some exceptions, but as I have said many times there are reasons that people do not go into psychiatry.  Talking to people with communication problems, irrational thought processes, and atypical social behavior are high on the list.  Some of the best primary care physicians recognize this and like the oncology clinic described in this paper give people with severe mental illnesses wide latitude in terms of appointments and treatment schedules,  but they can only do so much.

What is needed on the front end is a psychiatrist with strong medical skills to figure out the problem and get other medical staff involved.  I don't mean looking at PHQ-9 scores and suggesting medication adjustments.  I mean actually talking with the patients.  Very frequently a person with impaired judgment due to a psychiatric disorder will be fully cooperative in one setting but not another.  Consider a patient who is acutely admitted with very high blood pressure (260/140).  He has a history of schizophrenia and hypertension requiring moderate to high doses of two different antihypertensives.  In this case the patient has gone off of both medications and maintenance antipsychotic medication.  He is agitated and paranoid.  It is impossible to determine if he is also delirious due to the presence of cognitive disorganization from acute psychosis.  He will not allow testing or physical examination beyond the blood pressure determinations and eventually stop cooperating with that.  The inpatient psychiatrist makes an assessment and swings into action.  He tries to establish rapport with the patient to convince him that this is a medical emergency.  At the same time he has contacted the hospitalist service and there is agreement that he needs to go to an acute care bed as soon as possible.  The hospitalist reminds the psychiatrist that they cannot touch the patient unless he consents "or it is assault".  His advice is to call when he is ready to cooperate or call when he gets obtunded by encephalopathy or a stroke and they will treat him acutely.  In this case the psychiatrist eventually convinces the patient to check out the ICU and walks him over there.  Once he is in the bed - he is fully cooperative with all recommended measures including a complete physical exam.

In addition to the lack of a logical progression to the care of severe medical problems there can also be obstacles at the level of presenting the diagnosis to the patient.  I have presented diagnoses of cancer and diabetes mellitus to patients with psychiatric disorders only to be told that the diagnosis is impossible.  "I could not possibly have diabetes doctor, because I don't have a pancreas." comes to mind.  In terms of evidence, holding up an image of a lung or brain tumor may get the response: "I don't think that is my x-ray doctor.  That is somebody else's x-ray."  Those responses and the lack of ability to cooperate can be very frustrating for primary care physicians and specialists.  These patients are always very easy to get rid of.  All it takes is a comment like: "If you want me to treat you - you are going to have to stop smoking.  If you can't stop smoking - I can't treat you."  At the other end of the spectrum I have had Internists coach me over the phone on what to do for a patient, because they knew the patient would never come into their office or pick up a prescription.

The most frustrating cases are the ones that I saw too late.  The mistakes of informed consent as in:  "Mr. Smith you have colon cancer and need to have surgery to have the tumor removed.  At this stage you have an excellent chance for recovery but we have to operate in the next few months."  Mr Smith has schizophrenia.  He smiles and seems to understand everything.  He just wants to get out of the clinic and away from doctors back to his home where he will be much less anxious.  He never returns until two years later when concerned relatives call the police because of their concern about him.  The police find him alone at home.  The house is in disarray.  There is blood everywhere.  Mr. Smith is emaciated and has lost 35 lbs.  He is brought to the hospital and admitted to inpatient psychiatry.  He is seen by the oncologist who originally consulted on his case.  He now has widely metastatic colon cancer and no chance for survival.  That whole sequence of events can be prevented by a psychiatrist willing to discuss these potential outcomes long before the clinical picture worsens.

An infrastructure that allows for continued outreach and rapport building is also useful.  I had many patients with terminal cancer diagnoses admitted to my inpatient unit, not only because they were mentally ill and medical services would not admit them, but because there was no place else for a mentally ill person with terminal cancer to go.  Pulling all of the necessary resources and teams together with an initial acute admission potentially saves lives.  This paper is a good example of that, but acute care psychiatrists may still be held to the "acute dangerousness" standard for care and these admissions are actively discouraged.    

In the case of the NEJM article, the patient is a 63 year old woman with a history of lung cancer, chronic obstructive pulmonary disease (COPD), hypertension, hypercholesterolemia, asthma, insomnia, and restless leg syndrome.  She has also had two previous cancer diagnoses.  A previous diagnosis of lung cancer had been treated surgically with lobectomy and adjuvant chemotherapy six years earlier.    She had a past history of stage IIa estrogen-receptor positive and progesterone-receptor positive, HER2/neu negative invasive ductal carcinoma of the left breast.   The breast cancer was treated with lumpectomy, whole breast irradiation, and chemotherapy.  The patient had a previous psychiatric diagnosis of Attention Deficit-Hyperactivity Disorder treated at times with various stimulants and modafinil.  She was a admitted to the inpatient psychiatry unit for treatment of depression with electroconvulsive therapy.

In the process her psychiatrist comes up with a list of 13 factors that affect cancer care in patients with severe mental illness and cancer and strategies to approach them.  The factors follow initially what is known about cognitive and social behavioral deficits in the population with severe mental illness.  That would include an inability to understand the diagnosis or treatment.  In many cases, the patient is unable to provide informed consent due to cognitive deficits.  System wide deficits are identified at the level of the provider, the health care system, and society and culture.  Any physician who tries to provide medical or psychiatric care to these populations has seen most of these deficits.

In addition to the factors affecting cancer care there is a separate table of Differential Diagnosis of Depression in a Cancer Patient that every psychiatrist working in these settings needs to be aware of.
      
The psychiatrist in this case provides psychotherapy focused on the patient's understanding of the illness and their decisions to cooperate with care.  That included but was not limited to biological interventions for depression.  ECT and lurasidone were the main identified treatment modalities.  An enhancing mass was noted in the right breast on chest CT scan to follow up on previous cancer treatment.  That was subsequently diagnosed as ductal carcinoma of the  right breast.  In this case, the radiation oncologist modified the radiation treatment to maximize the flexibility of treatment for the patient.  She was able to complete all 5 treatments without difficulty.  

From a psychiatric standpoint she was discharged as improved after 19 days taking lamotrigine, gabapentin, quetiapine, and modafinil.  She had received 6 ECT treatments.  But as importantly, she had follow up oncology care, identification of a new cancer diagnosis, and coordination of that care also occurred in that setting.  This is a very compelling study at that level and a clear departure from the rationed inpatient care that people have come to expect when psychiatric units are run by business people.


George Dawson, MD, DFAPA


References:

1:  Irwin KE, Freudenreich O, Peppercorn J, Taghian AG, Freer PE, Gudewicz TM. Case 30-2016. N Engl J Med. 2016 Sep 29;375(13):1270-81. doi: 10.1056/NEJMcpc1609309. PubMed PMID: 27682037.

2: Shtasel DL, Freudenreich O, Baggett TP. CASE RECORDS of the MASSACHUSETTS GENERAL HOSPITAL. Case 40-2015. A 40-Year-Old Homeless Woman with Headache, Hypertension, and Psychosis. N Engl J Med. 2015 Dec 24;373(26):2563-70. doi: 10.1056/NEJMcpc1405204. PubMed PMID: 26699172.

Supplementary:

From the New England Journal of Medicine earlier this year:

New England Journal of Medicine Discovers Assertive Community Treatment. link

The Reality Of Burprenorphine Therapy

It is increasingly popular for politicians and healthcare businesses to discuss their ideas about how to end the opioid epidemic that they started.  One of the common themes is widespread availability of both buprenorphine maintenance therapy and naloxone opioid antagonist therapy for acute overdoses.  I am certainly not opposed to either and in fact work in an addiction treatment environment where these are two of several medication assisted therapies used to treat addictive disorders.  I am skeptical of the idea that broad prescribing of these therapies in either primary care clinics or some treatment settings will ever occur.  Naloxone will be more readily available because there is a movement to create easy access without a prescription.  That will never happen with buprenorphine.  Last week - an article in JAMA backs up my skepticism (1).

The JAMA article looks at 3234 buprenorphine prescribers in the 7 states with the most buprenorphine prescribers.  In their introduction the authors talk about the policy initiatives to increase the maximum patients per prescriber from 30 to 100 patients after a year.  The average monthly patient census per month varied from 7 - 22 patients and a median monthly patient census of 13 patients.  The duration of treatment episode was 53 days.  This illustrates that the monthly census was well below the allowed limits and the duration of treatment was well below the recommended maintenance guideline of 12 months.  They cite evidence that novice prescribers wanted more access to substance use counselors or other prescribers with more experience as potential limiting factors.

The authors of this article do not offer other explanations for the low rate of buprenorphine prescribing.  I have a few.  I really do not like stigma arguments.  To me stigma seems like an excuse for not being able to overcome societal biases toward a particular problem.  I don't see how you can train to be a physician and not have most of these biases wrung out of you.  With addictions and mental illnesses there may be a stronger bias based on personal experience.  Some physicians may have come from a family where the the father was an alcoholic or a heroin addict living homeless on the street and everybody was used to that idea.  Some physicians may have come from families where the father was still drinking and dying of cirrhosis and the familiy opinion was that he "has a right to drink himself to death" rather than get treatment that he did not want in order to stop drinking.  Other physicians may have come from families where father and his father both had severe alcoholism.  Grandfather drank himself to death by the time he was 50.  Father got treatment for his alcohol problem and was in stable recovery for years.  All of these personal experiences and the reactions to them will affect how a physician approaches alcoholism and addiction.

Those biases are all part the the inevitable decision-making process that leads physicians down specific career paths.  I have lost count of the number of times that another specialist told me that they really liked psychiatry and were considering the residency except for certain features of the field.  A couple of examples include needing to try to predict suicide and aggression and live with the consequences or dealing with a certain diagnostic group like patients with severe personality disorders.  People are less specific about addictions, probably because as medical students and interns we all see the severe effects.  Most of the acute care hospitals where physicians train have 30-50% of their admissions based on the acute effects of alcohol or drug use.  That includes many admissions for acute hepatitis, hepatic encephalopathy from cirrhosis, acute alcohol poisoning, acute overdoses on addictive drugs, and various psychiatric morbidities like delirium and psychosis from the acute effects of addictive drugs.  It is less obvious but addictive drugs and alcohol are also overrepresented as reasons for admission to surgical trauma units and burn units.  Most interns and residents see these effects first hand and develop both short term and long term perspectives on these problems.

This seems like another case of managers and politicians not appreciating the intense interpersonal aspects of medicine.  Physicians are all not foot soldiers just waiting for the next assignment from a policy maker.  Physicians have probably carefully selected the type of practice they want to be in and there are more than the technical aspects of the speciality that were considered.  It takes a unique skill set to treat people with addictions.  Treating and maintaining an opioid addict in treatment long enough with buprenorphine maintenance for them to realize any benefit is a very unique skill.  Being affiliated with other buprenorphine prescribers is also a necessity to provide cross coverage for patients.  Speciality care centers for addiction seem like an idea to me that does not get a lot of consideration.  Trying to run a buprenorphine maintenance program in a practice environment that is rationed to the degree it currently is does not seem feasible to me.  Adding buprenorphine maintenance as just another task for a busy primary care physician practicing primary care medicine is not likely to work.  It should be obvious that these physicians have  more than enough to do right now.

There is a lot more to it than increasing the maximum numbers of opioid addicted patients on buprenorphine maintenance and trying to treat as many people as possible.  The data from this paper illustrates that.  There is also the issue of the preventing the pool of opioid users from increasing while trying to treat those who are currently dependent on these drugs.  That seems like the best long term option to me.

Addressing this complicated problem takes more than a licensed buprenorphine prescribing physician sitting behind a desk who is willing to prescribe it.  It takes better infrastructure including managers who are enlightened enough to get that physician the kind of resources they need to do the work.  I never hear politicians or policymakers talking about that.


George Dawson, MD, DFAPA


Reference:

1: Stein BD, Sorbero M, Dick AW, Pacula RL, Burns RM, Gordon AJ. Physician Capacity to Treat Opioid Use Disorder With Buprenorphine-Assisted Treatment. JAMA. 2016 Sep 20;316(11):1211-1212. doi: 10.1001/jama.2016.10542. PubMed PMID: 27654608.