Military brain-chips to cure psychiatric disorders?

Sounding like something straight out of science fiction, DARPA recently announced grants to fund research and development of implantable brain-stimulation chips aimed to relieve, or even cure, mental disorders.  The Defense Advanced Research Projects Agency thinks big, and it has the money, i.e., our tax dollars, to back it up.  Decades ago, DARPA brought us the internet.  In comparison, revolutionizing psychiatry ought to be a walk in the park — right? Find a need and fill it: "Current approaches — surgery, medications, and psychotherapy — can often help to alleviate the worst effects of illnesses such as major depression and post-traumatic stress, but they are imprecise and not universally effective."  You can say that again.  So DARPA created a program called SUBNETS (Systems-Based Neurotechnology for Emerging Therapies) "to generate the knowledge and technology required to deliver relief to patients with otherwise intractable neuropsychological illness."   SUBNETS aims to create an "implanted, closed-loop diagnostic and therapeutic system for treating, and possibly even curing, neuropsychological illness."  In other words, computer chips in the brain.

SUBNETS will pursue the capability to record and model how these systems function in both normal and abnormal conditions, among volunteers seeking treatment for unrelated neurologic disorders and impaired clinical research participants. SUBNETS will then use these models to determine safe and effective therapeutic stimulation methodologies. These models will be adapted onto next-generation, closed-loop neural stimulators that exceed currently developed capacities for simultaneous stimulation and recording, with the goal of providing investigators and clinicians an unprecedented ability to record, analyze, and stimulate multiple brain regions for therapeutic purposes.

SUBNETS is hedging its bets.  With an overall budget of $70 million, it is funding both a diagnosis-based arm, in the manner of the DSM5 of the American Psychiatric Association (APA), as well as a “trans-diagnostic” approach, in the manner of the Research Domain Criteria (RDoC) of the National Institute for Mental Health (NIMH).   The ideological rift between the APA and NIMH last year was awkward and impolitic; fortunately,  SUBNETS has the resources to avoid choosing sides.  A research team at the University of California San Francisco (UCSF) will receive up to $26 million to study diagnostic groups, specifically post-traumatic stress, major depression, borderline personality, general anxiety, traumatic brain injury, substance abuse and addiction, and fibromyalgia/chronic pain.  Another team at Massachusetts General Hospital (MGH) will receive up to $30 million to tackle trans-diagnostic traits, such as increased anxiety, impaired recall, or inappropriate reactions to stimuli.  Both groups will include public and private partnerships, including with device manufacturers Medtronic, Draper Laboratory, and the start-up Cortera Neurotechnologies.

What to make of this?  Well, it's certainly ambitious.  As I read it, the effort relies on several unproven premises.  First, that psychiatric diagnoses, as currently construed, can be differentiated by monitoring activity in specific brain pathways.  This has been tried before without success, and it isn't clear that sensor technology was the reason.  An alternative model would suggest that mental states are an emergent property of widely integrated brain states.  If so, chips implanted in specific areas could no more capture this complexity than carefully listening to the trombone section could capture a symphony.

Another assumption is that carefully focused electrical stimulation can treat a variety of mental disorders.  The efficacy of transcranial magnetic stimulation (TMS) to treat depression provides some support for this idea.  In contrast, typical comparisons to deep brain stimulation to treat seizures and severe obsessive-compulsive symptoms only go so far.  Analogous stimulators may quell a panic state or chronic pain.  It is less clear how complex interpersonal patterns, such as those seen in borderline personality or substance abuse, could respond to this type of intervention.  Of course, we shall see.

A central tenet of SUBNETS is that implanted technology can promote healthy (or curative) neural plasticity.  Plasticity is a popular concept at the moment, highlighting the fact that brain wiring is not static, as was previously assumed.  "Neurons that fire together wire together" — that is, synaptic connections change dynamically in response to input, i.e., life experience.  This property underlies the hope that implanted stimulators may change the activity of neural pathways in a permanent way, "firing" the pathway together to make it "wire" together, and allowing the device eventually to be removed.  Again, we shall see.

Of course, there are many stumbling blocks ahead.  Implanting brain chips is no small matter, and this approach is unlikely to be used in the foreseeable future for anything short of the most severe, treatment-resistant disorders.  Initial public commentary immediately honed in on the "military mind control" aspect of the project, with visions of soldier drones being controlled on the battlefield via implanted chips.  The potential abuse of such technology is manifest and terrifying, and careful controls and standards are needed to assure freedom, not to mention safety.

At the most mundane level, the technology will only work if the science behind it is sound, and that remains to be seen.  Nonetheless, if even a portion of the SUBNETS agenda comes to pass, it would represent a monumental leap for psychiatric treatment.

Psychiatric uncertainty and the neurobiological buzzword

A few years ago I wrote that uncertainty is inevitable in psychiatry.  We literally don't know the pathogenesis of any psychiatric disorder.  Historically, when the etiology of abnormal behavior became known, the disease was no longer considered psychiatric.  Thus, neurosyphilis and myxedema went to internal medicine; seizures, multiple sclerosis, Parkinson's, and many other formerly psychiatric conditions went to neurology; brain tumors and hemorrhages went to neurosurgery; and so forth.  This leaves psychiatry with the remainder: all the behavioral conditions of unknown etiology.  Looking to the future, my fervent hope that researchers will soon discover causes and definitive cures for schizophrenia, bipolar disorder, and other psychiatric disorders comes with the expectation that these conditions will then leave psychiatry for other specialties.  We will always deal with what is left.  At minimum we psychiatrists should accept this reality about our chosen field.  After all, there appears to be no alternative.  Some of us go beyond this to embrace uncertainty as intellectually attractive.  We like that the field is unsettled, in flux, alive. Yet many of us clutch at illusory certainty.  Decades ago, psychoanalysis purportedly held the keys to unlock the mysteries of the mind.  It later lost favor when many conditions, particularly the most severe, were unaffected by this lengthy, expensive treatment.  Now the buzzword is that psychiatric disorders are "neurobiological."  This is said in a tone that implies we know more than we do, that we understand psychiatric etiology.  It's a bluff.

Patients are told they suffer a "chemical imbalance" in the brain, when none has ever been shown.  Rapid advances in brain imaging and genetics have yielded an avalanche of findings that may well bring us closer to understanding the causes of mental disorders.  But they haven't done so yet — a sad fact obscured by popular and professional rhetoric.  In particular, functional brain imaging (e.g., fMRI) fascinates brain scientists and the public alike.  We can now see, in dramatic three-dimensional colorful computer graphics, how different regions of the living brain "light up," that is, vary in metabolic activity.  Population studies reveal systematic differences in patients with specific psychiatric disorders as compared to normals.  Don't such images prove that psychiatric disorders are neurobiological brain diseases?

Not quite.  Readers of these exciting reports often overlook two crucial facts.  First, these metabolic differences only appear in group studies and cannot be used to diagnose individual patients.  As of this writing there is no lab test or brain scan to diagnose any psychiatric disorder.  Attempting to do so would be like diagnosing malnutrition based on height.  While malnourished people are shorter than the well-nourished on average, there is wide overlap and height is not diagnostic.  Second, etiology — the cause of these differences in brain function — remains unknown.  Differences in brain function (and structure) are not necessarily inborn.  Brain anatomy can change as a result of life experience, and metabolic activity (function) from experimental manipulation of cognitive effort, induced mood, guided imagery, etc.  Just as multiple factors affect a subject's height, multiple biological and psychological factors affect brain findings as well.  Thus, learning that patients with borderline personality show decreased metabolism in the frontal lobes (hypofrontality) is neither surprising nor indicative of a neurobiological etiology.  We already know the frontal lobes inhibit impulsive activity, and we already know borderline personality is characterized by impulsivity.  What else would we expect?

Genetic studies consistently show both heritable and environmental factors at play in psychiatric disorders.  Since the 1960s, psychiatry has called this combination the diathesis-stress model: an inborn predisposition meets an environmental stress, leading to an overt disorder.  The model helped shift the field from "nature versus nurture" to "nature and nurture" — and no research discovery or neurobiological rhetoric so far has shifted it back.  Patients and their doctors still contend with diathesis and stress: recreational drug use tips one patient into psychosis, sudden abandonment tips another into borderline rage.  Indeed, clinicians remain much more able to influence stress than diathesis.  A dispassionate assessment of what we currently know should lead to humble agnosticism about psychiatric etiology.  Genetics, biology, and environment all play a role, but beyond that there isn't much we can say.  This is why all current psychiatric medications treat symptoms and are not curative.

In this light, the popularity and zeal of neurobiological language is startling.   The American Psychiatric Association (APA) subtly changed the wording in its new Diagnostic and Statistical Manual, DSM-5, to imply that all psychiatric conditions are biological in nature.  The National Institute of Mental Health (NIMH) assumes that "Mental disorders are biological disorders...."  The National Alliance on Mental Illness (NAMI) says, "A mental illness is a medical condition...."

A more ground-level version is expressed by editor-in-chief Henry A. Nasrallah, MD in the latest edition of Current Psychiatry.  In an editorial not-so-subtly titled, "Borderline personality disorder is a heritable brain disease," Dr. Nasrallah proclaims BPD a "neurobiological illness" and "a serious, disabling brain disorder, not simply an aberration of personality" — as though these were distinct alternatives rather than two terms for the same thing.  After citing a number of biological findings which fail to prove etiology (e.g., the hypofrontality mentioned above) and which show partial heritability, Dr. Nasrallah concludes that "the neuropsychiatric basis of BPD must guide treatment."

Of course, it already does.  We already treat borderline personality disorder the best we know how, with psychotherapy (shown by functional imaging to modify brain metabolism, by the way) and often with adjunctive medication to treat symptoms.  What more do breathless declarations of brain disease buy us, other than reduced credibility?  It's not as though any of us currently withhold neurobiological treatment as a result of outmoded ideology.  On the contrary, the moment the FDA approves a cure for borderline personality disorder based on an established neurobiological etiology, I will gladly refer my patients to the neurologist, virologist, or genetic counselor who would thereafter treat such patients.

Third-party payment for psychotherapy: (2) Medical necessity

In my last post I outlined some complexities of third party payment for office psychiatry, and especially for psychotherapy.  As my example I used Medicare, the only third party payer I bill.  Some of the problems include complex billing (i.e., collecting from multiple parties), partial reimbursement, unrealistic documentation requirements, loss of patient confidentiality, and a misplaced emphasis on medication "evaluation and management" over psychotherapy.  There are also challenges specific to dynamic psychotherapy, such as obscuring the transference.  But I saved the most fundamental issue for this post: Does third party payment for psychotherapy make sense in general? This may seem a puzzling question, coming from me.  I not only value deeply what psychotherapy offers, I make my living from it.  Shouldn't it go without saying that psychotherapy should be paid for somehow, no matter where the money comes from?  My experience with public and private health insurers tells me otherwise.

"Medical necessity" is the linchpin, and frankly the problem.  The more a therapeutic encounter fits a medical model and is arguably "necessary" in that framework, the more readily it is covered by health insurance.  Psychotherapists of all stripes tiptoe uncomfortably around this issue.  Medication management fits the medical model very well, so psychiatrists who incorporate this into their psychotherapy sessions enjoy outsized reimbursement (or their patients do).  Talking about anything else, no matter how central to the patient's presentation, does not fit the medical model nearly as well.  Nonetheless, psychotherapists who offer a step-by-step approach aimed concretely at relief of symptoms emulate medical evaluation and treatment much more than those who employ open-ended, exploratory approaches to tackle dysfunctional family dynamics, chronic self-sabotage, and many other concerns for which people seek psychotherapy (and later report benefit; see Consumer Reports, November 1995, Mental health: Does therapy help? pp. 734-739, and this analysis of the Consumer Reports survey by Martin Seligman).  Note that the crucial variable for coverage is not what helps more, or relieves more agonizing misery.  It's what seems more "medical."

Using “medical necessity” as the criterion to treat human misery that often isn’t medical at all leads to much inconsistency and even cruelty.  As mentioned in my last post, insurers demand that I code my "procedure" (i.e., the session) depending on what we talked about.  If we spend the hour discussing medications, even if this focus can easily be understood as a symbolic, unconscious appeal by the patient for care-taking or some other emotional need, it's worth far more to the insurer than if we spend the same hour explicitly discussing the patient's experiences and reactions to actual caretakers.  (As added irony, the latter discussion can obviate the former in future sessions, a detail lost on insurers and most everyone else.)  Since private insurance partly reimburses many of my non-Medicare patients based on how their sessions are coded, an agitated, marginally employed, chronically suicidal patient with severe personality issues is reimbursed far less over time than a high-functioning, stably-employed patient with a medication obsession.  This makes no sense and is blatantly unfair.

The truth is, I'm the same expert — and put bluntly, worth the same amount of money — no matter what I'm discussing with the patient.  That is, as long as I have the integrity to focus on the patient's central issues, not to provide or bill for unneeded services, not to offer hand-waving in lieu of explanation, not to mindlessly prescribe medication after medication, not to casually chat and call it psychotherapy, and so forth.  In other words, I need to be a good doctor instead of a sloppy or unethical one.  I need to know when to be "medical" and when not to be.

Traditional dynamic psychotherapy fits the medical model especially poorly.  It is not primarily focused on symptom relief.  The treatment is not tailored to diagnostic categories.  It follows no step-by-step sequence.  Even expert practitioners often cannot estimate treatment duration.  After many decades of published studies the evidence base for treatment efficacy still triggers heated debates.  Arguing "medical necessity" for such treatment is at best unnatural, at worst contrived or even misleading.  (It's even more absurd to argue the medical necessity of one specific session in an ongoing treatment; to me, this is like asking whether the 10th note in a piano concerto is "musically necessary.")  Those of us who recognize the value of dynamic work and have seen patients change in important, fundamental ways are kept busy trying to pound this square peg into a round hole.  But CBT doesn't avoid this problem either: it's more like a square peg with rounded corners.

Faced with the struggle to show medical necessity, it's tempting to wonder whether psychotherapists should refuse to play this game.  However, opting out isn't easy.  Even if I chose not to be a Medicare provider — I admitted my mixed feelings about this last time — self-pay patients with private insurance would still seek maximal reimbursement for seeing me.  I can hardly blame them.  I see no way out of participating, at least indirectly, in this misapplied standard of medical necessity.

It's hard enough to assure that all Americans have access to basic health care.  Assuring that all have access to mental health care is one step harder, even when that care accrues only to the seriously mentally ill and fits the medical model very well.  It will be a very long time indeed before America deems it worthwhile to offer psychotherapy to the so-called worried well: those who have all their faculties but are miserable due to inner conflicts, self-defeating beliefs, or a traumatic past.  If that day ever comes, it will be when medical necessity is supplanted by a more fitting standard, one that judges mental distress and its treatment on their own merits, and not by borrowing legitimacy from medicine.

Third-party payment for psychotherapy: (1) "Do you take Medicare?"

From late 1996 to early 2007 I was medical director of a low-fee mental health clinic where psychiatry residents and psychology interns receive training.  Since the clinic accepted Medicare for payment, I did as well.  I signed on as a Medicare "preferred provider" and have remained on the panel ever since, even though I left the clinic for full-time private practice nearly seven years ago. I never joined private insurance panels for several reasons.  As an inveterate do-it-yourselfer, I've always handled my own billing and bookkeeping.  This is considerably harder when multiple health plans are billed, co-payments collected, and so on.  I like the straightforward way I provide a service, and the person receiving the service pays me directly.  Somehow it feels more honest than contracting with health plans to funnel referrals my way.  Private health plans also pay less than usual-and-customary fees and require doctors to share patients' private details with corporate reviewers to document "medical necessity."  Moreover, since dynamic psychotherapy has always been a big part of my practice — increasingly so over time — I'm sensitive to arguments that third-party payment complicates transference and countertransference, obscures acting-out around payment, and detrimentally takes payment out of the treatment frame.  Last but not least, as I'll discuss mainly in my next post, insurers base reimbursement on a medical model that fits poorly with dynamic work.

The upshot is that I have a cash-only (or "self pay") practice, with the exception of my Medicare patients.  Until this year, Medicare "allowed" 65% or so of my full fee.  (Medicare sets an allowed fee for a given service, and then pays 50-80% of that.  I can collect the rest, up to the allowed amount, from a secondary insurer or from the patient.  This works more or less automatically for secondary insurers, and rather awkwardly when I try to collect from patients.)  In 2013 the CPT codes for psychiatric office visits were revamped.  This made billing more complicated, and introduced odd, often illogical variations in Medicare and private insurance reimbursement — sometimes paying more than before, sometimes less.

As one of the few private-practice, office-based psychiatrists in San Francisco still on the Medicare panel, I've become a magnet for these patients.  A local medical center with which I have no affiliation used to refer several callers to me every week, until I sent a letter asking them to please not kill me with their kindness.  Medicare callers request to see me for medications only, even after I explain this is not the nature of my practice.  It's more tricky when patients claim to want therapy to get a foot in the door, and then once in my office and now my medico-legal responsibility, confess that they only wanted medication refills all along.  Some callers ask to be added to a non-existent waiting list, or to call me every month or two to see if I change my mind about accepting them as patients.  Clearly, the demand is there, the economic incentive is not.

Medicare and other third-party payers have a valid need to assure their money isn't wasted.  Sometimes my claims are rejected, as when I received a notice this week that one patient's diagnosis (Depression Not Otherwise Specified, 311) "is inconsistent with the procedures" I billed (three weekly sessions of moderate-complexity medication management, 99213, combined with 50-minute therapy sessions, 90836).  It's tempting to protest this, as there's absolutely nothing inconsistent about treating atypical depression with medication and psychotherapy.  I could take the time to marshal my arguments, compose a letter, and reveal personal details about my patient to present my case.  But it's far easier to resubmit the claim with a slightly upcoded diagnosis, e.g., Major Depression, recurrent, mild severity, 296.31, and get paid.  This uncomfortably clashes with my usual tendency to downcode slightly to protect my patient's confidentiality.  (Since pressures to upcode and downcode routinely distort the documentation of diagnoses in clinical practice, I'm skeptical of all research that uses these diagnoses to derive conclusions about psychiatric practices, disorder incidence, and the like.  Garbage in, garbage out.)

Upcoding and downcoding in such cases is not criminal mischief, but an attempt to fit traditional, mainstream psychiatry into a procrustean bed of medical-model diagnosis and procedure coding.  Public and private insurers alike sacrifice ecological validity for documentation that appears, but really isn't, "evidence based."  To take one example, as of this year we must code medication "evaluation and management" separately from the provision of psychotherapy, even if in practice these are done simultaneously and inseparably.  A 50-minute psychotherapy session (90836) that includes brief attention to medication (99212) is reimbursed at a much lower rate than the same 50-minute session with more time devoted to meds (99213 or 99214).  This makes little sense when in many cases the psychotherapy is far more clinically significant than the medications being discussed.  (You'll note that I think of the psychotherapy code first, but actually it is an add-on to the primary medication "E & M" code.)  If medications are not mentioned or evaluated at all, there is yet another code to use for psychotherapy (90834), with an "allowed fee" of $89 for 50 minutes, well below what any psychiatrist or psychologist actually charges.  If this isn't bewildering enough, some of my colleagues are now doing 52-minute sessions, an insignificant increase in duration that qualifies for a different code with much higher reimbursement.

Since cash-only practice excludes all but the affluent, I view my taking Medicare as a modest concession to avoid elitism.  I also support a single-payer health care system, also known as "Medicare for all," so participating in Medicare feels like practicing what I preach.  At the same time, it's easy to see why most of my office-based colleagues opt out of Medicare: lower pay for more paperwork, rules that don't make sense, and various factors that make dynamic psychotherapy harder to conduct and be paid for.  So far I still answer yes, albeit hesitantly, when asked whether I take Medicare.  In my next post I'll expand these ideas into private insurance for outpatient psychiatry, including whether dynamic psychotherapy resembles a medical intervention enough to fit a "medical necessity" model.

The APA annual meeting: a photo essay

As posted previously, last month I attended the American Psychiatric Association's (APA's) annual conference.  Straying from my usual format, I thought I'd post pictures from the meeting and, of course, offer comments. The meeting took place in Moscone Center, a conference center complex located just south of Market Street in downtown San Francisco.  Depicted here are anti-psychiatry protesters who held a rally in front of the main entrance at noon on the first day.  There was also an exhibit of psychiatry's cruelties (psychosurgery, shock treatment, inhumane conditions in asylums, etc) running all five days in a tent across the street from the conference.   The conference was also a block from Yerba Buena Gardens, where I caught a very pleasant Balinese gamelan concert at the same time as the protest rally. This simultaneity — two events scheduled to coincide, forcing a choice — was a constant in the conference as well. The "scientific program" consisted of  numerous overlapping talks, such that attending any presentation meant missing five or more other good ones.  I'm not sure why the APA opted for such frustrating redundancy.  Nor can I explain why predictably popular talks were scheduled into small rooms, with the result that dozens of registrants were turned away once the room filled.  For instance, the crowd for Otto Kernberg's psychoanalytic talk on love and aggression was several times larger than the assigned room.  In this unusual case we were all moved to a cavernous hall at the last moment, where Dr. Kernberg gave a warm and very engaging presentation on the necessity and creative consequences of aggression in romantic love.  (I like how this photo depicts the renowned psychoanalyst Kernberg representing the APA in an era of biological ascendancy.)

The same huge auditorium was to hold the keynote address by Bill Clinton.  However, Mr. Clinton was ill and could not be there in person.  Several hundred (a couple thousand?) conference-goers nonetheless waited over an hour to see him on video.  Mr. Clinton was pleasant, thoughtful, and charismatic, but didn't offer much specifically about psychiatry or mental health.  Mostly he spoke about public health needs in general.

I didn't take many photos in the talks themselves.  Officially it was forbidden, although this rule was routinely ignored by attendees.  The quality of the presentations was high — I mostly chose "mainstream" ones this time, not the many off-beat and generally smaller meetings.  I attended presentations on suicide, personality disorders, PTSD, sexual compulsions, DSM-5 and mood disorders, the controversy over antidepressant efficacy, psychiatrists writing and blogging for the general public, teaching psychotherapy to residents, and assessing the capacity of demented patients to make medical decisions for themselves.  There were dozens of others I would have liked to attend, had they not coincided with the ones I chose.

I skipped the industry-sponsored, free lunch or dinner, non-CME presentations.  But I did wander through the exhibit hall, both to see the "new investigator" scientific posters, and to peruse the brand-new DSM-5. In contrast to the last time I went to this conference, the industry booths seemed less garish and "over the top."  Of course, there were still a lot of them.  Several had raffles where valuable prizes such as an iPad Mini could be won by those who gave the company their contact information.  One booth offered a pocket digest of the new DSM-5, MSRP about $60, to everyone who watched a 12 minute presentation and coughed up a mailing address.  I was tempted... but no.  (It's interesting to ponder how much a single psychiatrist contact is worth to a drug company.  Much more than $60, I'd venture.)

The DSM-5 itself is $200 in hardcover, $150 in paperback — an unabashed moneymaker for the APA.  Despite the incredible controversy it stirred up, my impression is that the changes from DSM-IV-TR are relatively minor.  In particular, the personality disorder section hasn't changed much, although the new edition is no longer multi-axial, i.e., there is no "Axis 2".  Some language has been made more precise, as well as more "biological" in some passages, and some disorders have been expanded to include more that would previously have been considered normal.  Whether this is good or bad depends on one's perspective in several respects; mostly I find it unfortunate.  DSM classifications often matter more to insurers and disability officers than to practicing psychiatrists, who in David Brooks' words are "heroes of uncertainty" (echoing an earlier post of mine, but I'll forgive him for not quoting me).  We deal with individuals, not disease categories.

I will end with a slide from the talk on antidepressant efficacy that summarizes this tension in my field.  As I've discussed previously, randomized controlled trials (RCTs) are the gold standard for scientific rigor in psychiatry; however, a lot of psychiatry is not scientific in this sense.  DSM categories help define the "average" patient with a particular disorder, leaving a lot of wiggle room since the categories are not based on etiology.  RCTs say which treatments best help this "average" patient, represented by the computer composite in the center of this slide.  However, I don't see "average" patients,  I see one of the 12 individuals who contributed to the composite.  Thus, for me, the new DSM was a sideshow at the conference.  The most insightful presentations, whether on PTSD, suicide, or capacity assessment, combined science and the nuanced human communication of meaning.  They recognized that our work is informed by science but goes well beyond it.  Anti-psychiatrists don't like this, insurers don't like this, neuroscientists don't like this, even many psychiatrists don't like this.  But it's true and inevitable for the foreseeable future.  I like it.  As for the APA annual meeting, I'm glad I went, and equally glad I won't feel the need to go back for several years at least.