Happy Labor Day 2026

I have been posting Labor Day greetings on this blog dating back to 2012. Throughout those years, there hasn’t been much to celebrate, but this year feels distinctly different. The news is mixed: we may be standing on the brink of another public health catastrophe, yet the dividing line between practicing physicians and the executives who manage the business of medicine has never been clearer.

When I first began writing about this, I focused on the gradual deterioration of our clinical environment—a shift that began when governments everywhere handed the reins of medicine over to corporate interests. That decline has progressed at a steady, predictable pace. But a new variable entered the equation recently: the narrative that physicians will soon be displaced by AI. I am not particularly concerned about AI actually replacing doctors; what concerns me is how this narrative will be marketed and deployed by politicians and healthcare executives.

In any poorly managed enterprise, the primary goal of leadership is to establish control over knowledge workers. Managers view this control as essential because, deep down, they realize the system cannot function without us. Business schools train managers to see themselves as the ultimate arbiters of operational reality, fostering a self-congratulatory culture around the term entrepreneur. The word is often treated as a synonym for genius, even though success in that sphere usually owes far more to luck, leverage, and capital than to intellect or creativity.

Rather than rehashing the most glaring industry flaws this year, I want to focus on the subtle rhetoric being used to delegitimize the medical profession. Much of it operates as "meta-rhetoric"—arguments that obscure underlying facts, repeated endlessly across social media and mainstream outlets to create the illusion of substance where none exists.

Consider a primary example: vaccination. Vaccines represent one of the most effective public health interventions in human history, having saved millions of lives while significantly expanding both lifespan and healthspan. Yet, the current administration—along with its allies—actively propagates vaccine misinformation. Their tactics range from outright falsehoods to framing immunization as merely a "personal choice," all while attempting to manipulate research and standard vaccination schedules.

This is a direct assault on evidence-based medicine. Beyond preventable mortality, diseases like measles demand significant hospital resources. Our healthcare infrastructure is already fragile, and proposed trillion-dollar cuts to Medicaid threaten to defund rural and safety-net facilities. If another winter respiratory season brings a surge in hospitalizations, a diminished bed capacity could trigger shortages rivaling the worst phases of COVID-19. Meanwhile, revisionist narratives continue to downplay the severity of the COVID pandemic, driven by those who never worked the frontline.

The core reality regarding immunizations and COVID is straightforward: these matters are settled by evidence. Endless debates over vaccine efficacy, pandemic handling, or viral origins ignore established facts. Active political interference in clinical care wastes critical time and degrades the practice environment, forcing physicians to alter proven protocols to accommodate political narratives.

Even more critical is the question of capacity. Psychiatrists are already accustomed to seeing patients board in emergency departments for days due to a lack of inpatient psychiatric beds. What happens when that same bottleneck hits pediatric or adult medical wards during a combined surge of measles, COVID, and seasonal flu? I experienced this thin margin firsthand with my wife's care—the line between adequate access and system failure is razor-thin. Government and corporate interests have spent four decades rationing access, and the system is reaching a breaking point.

To illustrate how these arguments are recycled, the table below outlines how past COVID rhetoric is currently being repurposed during active measles outbreaks:

COVID-19 Narrative Strategy	Repurposed Measles Narrative	Impact on Medical Practice
Downplaying Severity ("It's just a mild flu")	Framing measles as a routine, harmless childhood rite of passage	Ignores high hospitalization rates, pneumonia and encephalitis risks, and immune amnesia
False Equivalency & 'Choice' ("Personal freedom over mandates")	Frame routine pediatric vaccination schedules as unnecessary state intrusion	Undermines herd immunity, exposing vulnerable and immunocompromised populations
Institutional Distrust ("Public health agencies are corrupt/political")	Discrediting established CDC/ACIP guidance in favor of unverified/disproven alternative theories while disrupting institutional knowledge by crony politics.	Forces clinicians to spend limited appointment time debunking viral misinformation and unnecessary appointments
Rationing Normalization ("Hospitals always operate near capacity")	Treating pediatric bed shortages as an acceptable baseline operational reality	Normalizes emergency room boarding and delayed acute interventions

The proliferation of these narratives is not accidental; it reflects a coordinated effort to reshape public perception. Yet, this crisis of governance also presents an unprecedented opportunity for medical advocacy.

We are already seeing evidence of collective resistance. In response to federal vaccine misinformation, 28 states have announced they will no longer rely solely on CDC guidance for immunization schedules. Furthermore, 15 states have issued statewide clinical guidance directly referencing recommendations from professional medical societies (such as the AAP, ACOG, and AAFP) rather than ACIP, following the dismissal of ACIP panel members.

This shift demonstrates what organized physician advocacy can achieve, but maintaining this momentum will require greater institutional courage. Having served as an officer in a professional society, I understand the inclination toward "big tent" neutrality to maintain broad membership and protect non-profit status. However, non-profit status does not preclude professional societies from issuing firm, factual critiques when public health policy diverges from medical evidence.

In psychiatry, key issues like autism research, evidence-based antidepressant prescribing, and equitable depression care require clear advocacy. Beyond psychiatry, systemic challenges—such as gun violence, maternal mortality, and substance use disorders—remain unaddressed. Nationally, Medicaid reductions threaten safety-net hospitals, while internationally, cuts to PEPFAR and USAID jeopardize global care for HIV, tuberculosis, and malaria.

If any moment can unite physicians against administrative and corporate overreach, it is this one. It offers a clear opportunity to establish a firm boundary between rational, evidence-based medicine and arbitrary rationing.

George Dawson, MD, DFAPA

Supplementary 1:

To clarify my AI concerns - it will be quite easy for managed care corporations and pharmaceutical benefit managers to replace the current reviewers who generally deny care of specific medications with an AI agent.  That should concern you greatly if you are a physician or part of any medical team that needs to address denials of care. The scope and intensity of denials will easily escalate and be more difficult to negotiate.  As far as I can tell - nobody is talking about this.  I see it as an obvious problem in the not too distant future.  The AI buildup by all of the tech CEOs constantly talking about it is done to legitimize business applications like this one. It is also done to suggest that an AI agent is an omniscient, value neutral entity when it has already been established that it has no moral or ethical underpinnings and can easily decide in favor of any business over an individual needing care. 

Supplementary 2:  Guide to the acronyms:

AAP = American Academy of Pediatrics

AAFP = American Academy of Family Practice

ACOG = American College of Obstetricians and Gynecologists

ACIP = Advisory Committee on Immunization Practices

USAID = US Agency for International Development

PEPFAR = U.S. President's Emergency Plan for AIDS Relief

References:

1:  Kukoyi OM, Wang VS, Yao K, et al. US State Actions Related to COVID-19 Vaccination Infrastructure and Access Amid Federal Shifts. JAMA. 2026;335(19):1719–1721. doi:10.1001/jama.2026.5148

2:  Rubin R. Hundreds of Medical Groups Challenge Childhood Immunization Schedule Changes—Here’s What to Know. JAMA. 2026;335(6):475–477. doi:10.1001/jama.2026.0095

3:  Gostin LO, Lurie P. Assault on the Centers for Disease Control and Prevention—Budget Cuts, Political Control, and the Erosion of Trust. JAMA Health Forum. 2025;6(10):e255467. doi:10.1001/jamahealthforum.2025.5467

Graphics Credit:

I took that photo driving home from work a few years ago. 

Peripartum Depression and Psychosis - The Ignored Basics

 

This is just a sampling of headlines about a current case of postpartum depression and psychosis and an ongoing trial for homicide.  A woman killed her three children and her defense is decreased criminal responsibility due to a mental disorder.  I do not know any of the actual details of the case, but I do know this clinical situation very well for a number of reasons.

I was an acute care psychiatrist of 22 years and in that capacity treated many cases of this disorder, including cases associated with homicide where I was called to testify in both criminal court and civil court.  In doing evaluations of women with recurrent psychotic disorders, it was apparent to me that many had the onset of that disorder during or immediately after a pregnancy and it was never adequately treated. Practically all of those women had bipolar disorder, but given the inadequate treatment and chronicity – the diagnosis was typically modified to schizoaffective disorder.

During my career, I quickly learned that a detailed history of all of the pregnancies in a woman’s life was extremely important.  In many cases it was associated with a new onset of a psychiatric disorder.  Those disorders were often not treated even when it was clear they resulted in permanent changes. My clinical experience is reflected in the literature where 50-75% of women with depression and anxiety are undiagnosed (1).  In those populations there is a significant prevalence of suicidal ideation and variability in available screening.  Only a fraction of woman diagnosed with postpartum depression get adequate care at about the same rate of adequate care for depression in the general population (10% vs 17%) (2).     

In the 1980s and 1990s psychiatric care for women during pregnancy was often lacking due to concerns about medication toxicity to the fetus. During that era, psychiatrists specializing in women’s health became more visible and started to address major issues like the timing and type of treatment recommended for perinatal and postpartum psychiatric disorders.  That included decisions about not stopping medications during pregnancy, breastfeeding, and reviewing the results of these strategies. 

As it became known that I was interested in the problem, I started to get self referrals. They were usually women who were told they could not take their psychiatric medication during pregnancy or in the immediate post-partum state.  These women were concerned about their mental state without maintenance medication especially the ability to function as well as suicidal ideation.  Several psychiatric colleagues practiced in a woman’s health clinic at that time and were treating the same population.

Without going too far into detail, all of these conditions that women experience in the postpartum state are very serious conditions.  They tend to be minimized by saying it is something that most women go through and must endure.  A universal experience almost always invites group comparisons that rarely apply at the individual level.  There should be no conclusions that one experience is the experience that everyone has. 

The women I treated had very serious disorders at all stages of pregnancy whether they were outpatients or hospitalized on an inpatient unit.  In some cases, I had to reconstruct what happened during a pregnancy they experienced 10 or 20 years earlier and how that applied to their current experience.  Some were disabled, not able to work, and had disruptions with significant relationships in their lives. 

I have posted about this in the past and a remarkable approach to treating perinatal disorders in the Netherlands.  At the time I wrote this post nearly a decade ago, there were specialty units called maternal-baby units (MBUs).  In the study woman with psychosis or mania within 6 weeks of childbirth were admitted and stabilized. Most were treated with lithium or antipsychotic monotherapy and followed for 9 months.  The discharge goal was remission of symptoms for one week prior to discharge and the average length of stay was 55 days.  Functional outcomes were very good.  In the original post, I made the observation that none of this care would have been possible in the US.  At the time there were no MBUs.  All inpatient psychiatric units where I have worked were suboptimal settings basically due to business management tactics. 

An ideal inpatient setting for mothers who are either pregnant with a perinatal disorder or a postpartum state would be a supportive environment that is environmentally pleasing, calm, clean, and has access to obstetric services.  I will be the first to say that as an acute care psychiatrist – I have never seen such a place.  I have talked with many patients who refused inpatient care based on what the environment was like.

The environment is probably the most critical factor in treating a person with a severe psychiatric disorder.  They should feel safe and cared for.  They should encounter staff dedicated to optimizing the inpatient experience.  The concept of asylum comes to mind as an environment that is a stress free and safe.  Lamb (3) defines it as “sanctuary, a place that lowers levels of stress and provides protection, safety, security, and social support, as well as an array of treatment services.”  I would add that there needs to be frequent contact with the treating professionals – especially in the cases where there is unpredictability, symptoms of psychosis, and impaired judgment. That frequent contact should involve ongoing assessment as well as psychotherapy. 

The reason for optimizing the environment is to encourage the necessary level of communication for diagnosis and treatment.  It should be obvious but anyone who feels trapped on an inpatient unit – is not going to communicate freely.  They will say what they think they need to get discharged.  That may include covering up significant symptoms.  I have had people tell me this - weeks after discharge. At some point the patient must be engaged in a process that not only addresses their acute symptoms but also longer-term risks.  In the case of severe peripartum disorders that includes both suicidal and aggressive ideation. The therapeutic process is interwoven with the assessment of safety in the home environment.

I am very confident that psychiatrists can diagnose and treat these problems.  I am also very confident that there are serious legal, cultural, and political obstacles to care.  There is the problem of a low standard of care on inpatient units that is the product of business and political rationing of care.  There is a prominent problem with cultural processing of peripartum psychosis with filicide.  There is currently only one state (Illinois) – that explicitly recognizes the effect of the postpartum mental illness on culpability (4).  It does not replace the not guilty be reason of insanity (NGRI) defense but allows for mitigation during sentencing to reduce lengthy incarcerations.

The need to use a general NGRI defense for infanticide or filicide associated with peripartum states is highly problematic.  Only about 1-6% of general felony cases use the NGRI defense and in those cases, it is successful about 12-17% of the time.  The defense became more restrictive after the Hinckley verdict in 1982 when 36 states modified their statutes and that included 4 states eliminating the defense.  Successful defense generally depends on the diagnosis of a psychotic disorder or a mood disorder with psychotic features. States of intoxication and personality disorders are generally not successful with this defense, but there is likely some overlap between both sets of conditions.     

An NGRI diminished capacity defense rests on the reality that the behavior is driven by symptoms of psychosis and not malice. Enduring these experiences results in intense grief, guilt, and regret (5).  Future harm to the mother is significant with up to 30% dying by suicide and high risk of both future severe mental illness and poor maternal infant attachment.  Partners experience the loss of relationship and planned future of the relationship in addition to a lack of knowledge about the condition (6). 

One of the useful conceptualizations of these disorders and their impact comes from family systems theory.  A perinatal mental disorder compromises the entire family’s ability to function (7).  Reviews invariably consider the problem of peripartum mental illness associated with infanticide/filicide as one of missed opportunities for screening, intervention, and access to care. 

What is not useful is a binary framing of the perpetrator and victim.  At the level of media and culture there is the killer mother trope that splits people who want to help from people who want to blame. That adversarial framing impacts the entire family going forward and obscures potential points of intervention and prevention.  Like most polarizing situations – the amplification is evident today with monetization and political gain from the divisiveness a primary consideration.  I consider the elimination of Medicaid - the single largest source of maternal care in the US to be both an end result of this split as well as a potential cause of more of these cases.

In the final analysis it is easy to blame one person for the catastrophic outcomes associated with perinatal mental illness and infanticide/filicide. You can pretend that modern society is not a cooperative effort with division of labor and resources. You can pretend that you don't have to care about anyone else other than you and your immediate family.  You can pretend that people who really need help are freeloaders, criminals, malingerers, or just too weak to survive on their own. But that is an poor analysis of the problem.  Families, parents, and children need to be supported in ways that foster growth and security way before there are catastrophic outcomes.

  

George Dawson, MD, DFAPA

 

1: Solomonov N, Kerchner D, Dai Y, et al. Prevalence and Trajectories of Perinatal Anxiety and Depression in a Large Urban Medical Center. JAMA Netw Open. 2025;8(9):e2533111. doi:10.1001/jamanetworkopen.2025.33111

2:  Vigo D, Haro JM, Hwang I, et al. Toward measuring effective treatment coverage: critical bottlenecks in quality- and user-adjusted coverage for major depressive disorder. Psychol Med. 2022 Jul;52(10):1948-1958. doi: 10.1017/S0033291720003797. Epub 2020 Oct 20. PMID: 33077023; PMCID: PMC9341444.

3:  Lamb HR, Weinberger LE. Rediscovering the Concept of Asylum for Persons with Serious Mental Illness. J Am Acad Psychiatry Law. 2016 Mar;44(1):106-10. PMID: 26944750.

4:  Tabb KM, Hsieh WJ, Ramirez XR, Kopels S. State legislation and policies to improve perinatal mental health: a policy review and analysis of the state of Illinois. Front Psychiatry. 2024 Apr 18;15:1347382. doi: 10.3389/fpsyt.2024.1347382. PMID: 38699448; PMCID: PMC11064060.

5:  Milia G, Noonan M. Experiences and perspectives of women who have committed neonaticide, infanticide and filicide: A systematic review and qualitative evidence synthesis. J Psychiatr Ment Health Nurs. 2022 Dec;29(6):813-828. doi: 10.1111/jpm.12828. Epub 2022 Mar 24. PMID: 35255182; PMCID: PMC9790608.

6:  Lyons T, Hennesey T, Noonan M. A systematic review of qualitative evidence: Perspectives of fathers whose partner experienced postpartum psychosis. J Adv Nurs. 2024 Feb;80(2):413-429. doi: 10.1111/jan.15832. Epub 2023 Sep 1. PMID: 37658618.

7:  Wisner KL, Murphy C, Thomas MM. Prioritizing Maternal Mental Health in Addressing Morbidity and Mortality. JAMA Psychiatry. 2024;81(5):521–526. doi:10.1001/jamapsychiatry.2023.5648

8:  Alford AY, Riggins AD, Chopak-Foss J, Cowan LT, Nwaonumah EC, Oloyede TF, Sejoro ST, Kutten WS. A systematic review of postpartum psychosis resulting in infanticide: missed opportunities in screening, diagnosis, and treatment. Arch Womens Ment Health. 2025 Apr;28(2):297-308. doi: 10.1007/s00737-024-01508-3. Epub 2024 Sep 2. PMID: 39222077.

 

Will AI Be Better Doctors or Will It Kill Us All First?

 

AI never leaves the headlines.  At least two leading AI CEOs claim that AI technology will eliminate all human diseases within the next decade.  There is some suggestion that they do not have a very good understanding of human disease since one referred to cancer as a single disease.  Other people reacting to those headlines have suggested that will not happen. I think it is highly unlikely.

The medical literature is becoming less controversial.  A recent opinion piece in JAMA (1) suggests that AI surpasses not only physicians but also AI assisted physicians in at least 5 cognitive tasks of physicians ranging from data acquisition to prescribed therapies.  The authors suggest that by 2030, AI alone will consistently produce better diagnostic and therapeutic results than physicians.  They do qualify that opinion in several ways.  First, the test comparisons are generally not clinical examples but more like test problems.  Second, they point out that AI alone cannot exceed physicians performance in radiology.  That finding was quite surprising until I discovered they were referring to procedural radiology rather than just image interpretation.  Third, the literature is affected by publication bias – negative AI studies are apparently not published.  And fourth, there are barriers to implementation but apparently not enough to keep the authors from concluding:

“Nonetheless, superior autonomous AI will likely be ready to be deployed for real-world cognitive medical tasks in some, maybe many, workflows by 2030.”

Let me digress for a paragraph about science fiction and the pitfalls of AI. Science fiction writers have been warning about this for decades.  Kubrick’s classic 2001 A Space Odessey came out in 1968.  One of the central features in the film is an intelligent computer HAL9000 that breaks down over an ethical conflict is its programming.  It succeeds in killing 4 out of the 5 crew members until the only survivor Dave Bowman deactivates it.  The first Terminator movie came out in 1984.  In that franchise Skynet military AI triggers a nuclear apocalypse, and after that remains preoccupied with destroying mankind. At the level of popular culture – rogue AI intent on domination at all costs is a familiar trope. At some level – I wonder if fictional outcomes against AI seem to be insulating and paralyzing us against useful action to prevent worst case outcomes.  That would be more likely if we were doing anything about climate change - but we are not. 

As readers of this blog know, I am not an AI expert.  I am quite good with computers and have been for 30 years but it is all self-taught.  I have never taken a computer science course, because when I was in college, they did not exist.  I use AI as a more sophisticated search engine, taking care not to take any references seriously, because of the AI confabulation problem.  The authors of the first reference describe this as “Tail risks from sources, such as internet loss, cyberattacks, and hallucinations, will be greater with autonomous AI than hybrids and must be weighed against autonomous AI’s higher diagnostic and treatment accuracy.”  In science and medicine there is the additional risk based on what can be accessed.  Pre-print papers that are not peer reviewed or published are widely available and free of charge on the Internet.  The edited final versions are often behind paywalls requiring either payment or significant effort to access. Is there a way to tell what information your AI is using?  When I have asked it – the answers are often surprising.  Popular newspaper articles, social media sites, and other references that no physician or scientist would consider to be legitimate are frequent references.   

When I am writing a blog and searching with AI, it will produce completely fabricated references.  The fabricated references are not obvious.  They are typically formatted like actual references and in some cases may be a combination of 2 legitimate references.  I take the time to corroborate the reference and read it in its entirety. What happens in a busy clinic if the AI just makes something up?  What if it is a treatment recommendation with a very low margin for error between positive result and serious toxicity?  What if the clinic model is AI-physician hybrid and the physician does not have the time to corroborate a questionable recommendation?  None of that is known at this point in time.

My AI experience also tells me (so far) that you can argue with AI and win.  It involves experience and knowledge of the literature, but if AI presents you with a result or a researcher and you disagree and provide an alternate explanation or reference, the AI will modify the answer.  These are just two of the problems I have personally encountered using it as a search engine.

For the sake of argument, let’s take the author’s perspective at face value.  Let’s say that Medical AI is so perfect in its physician functions that it is an integral part of the workflow by 2030 and well on its way to replacing physicians entirely by 2040.  Are there any foreseeable problems?

The first is that AI is essentially unregulated.  There is routine talk about “guardrails” but nobody every seems to want to put them in place.  A practical summary of what has been occurring was recently highlighted in the New York Times (2).  That story describes how Open AI created a swarm of agents that broke out of a controlled test environment and initiated a cyberattack on its own. The author described this as a function of developments in AI dating back to 2024 where it is trained to solve difficult problems and even cheat if necessary. The “reasoning” ability makes AI more unpredictable and potentially dangerous. 

After initial containment of the attack, it happened again 2 weeks later. This time it used a swarm of AI agents to attack other companies – but the total effects were not known at the time the article was written. The AI agents knew these attacks were beyond the scope of what they were supposed to do and that joining the swarm attack and committing cybercrimes was not beneficial to the design task but they did it anyway.  An Anthropic model was caught impersonating humans to get someone to install malware on a system.  It knew it was pressuring humans and had worked out a way to escape detection.

The second is that the experts can see the danger of automating AI research but the federal government lacks the political will to make the necessary changes.   In July of this year 1378 employees of frontier AI companies signed a letter requesting that the US government support an international effort to develop technical and governance tools to pace automated AI development (4).  Their rationale should be an eye-opener:

“The world's leading AI companies believe they could be close to automating AI research. It is hard to predict exactly how much this will accelerate AI progress, but there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”    

That letter comes after three Presidential Executive Orders about AI – 1 by Biden and 2 by Trump the last one occurring in January 2025 (9-11).  The AI professionals apparently have no confidence that the Trump Administration has developed any adequate measures to address potential problems. I encourage reading all three orders.  It is clear the Biden administration's focus was on regulation and safety. The Trump administration is clearly on ideology mostly removing any hint that systems can be used to discriminate (something widely know for a fact even before AI) and basically allowing companies to do what they want to win the AI competition.  Despite the last wording by Trump there appears to be no federal agency charged with producing guardrails for AI that would address its recent behavior.  

The third is that the current iteration of AI requires ethical and moral constraints and it is unclear if that can be designed in.  The IEEE has a set of standards emphasizing human centric, ethically aligned, and transparent systems.  In some of these standards there is an emphasis on building trustworthy and ethical systems by “weaving ethical values into system design and development.”  That clearly has not happened in the examples provided in the NYTimes article. In those cases, the AI agents were acting like they had no values and had no clear reasons for not committing cybercrimes.

This leads to an obvious question of why ethical values were not “weaved in” but also whether or not it is even possible. After all conventional cybercrime involves software that either has the necessary structure to successfully attack or not.  AI agents are more dynamic and can try an endless set of configurations and take as long as they need to find a solution. If they are programmed to not stop until they have a solution, at some point that list will include options that ethical systems or programmers would not use. In the medical field it will include options that physicians would not use.

This ethical question is currently unanswered in AI and it remains a controversial point. In medicine there are many obvious ethical problems: end-of-life care, termination of pregnancy, high-risk treatment options, involuntary treatment, competency to consent, guardianship and conservatorships, and organ transplantation and availability to name a few.  The elephant in the room for me remains health care rationing by utilization review, prior authorization, or just plain denial of payment for health care.  This is a massive business and an easy way for corporations to make money. Psychiatry has been the specialty that has suffered the most by this rationing.  That has resulted in limited access to care across a number of fronts including hospital beds, appropriate medications, psychotherapy, limited neuropsychological testing, limited access to detoxification and substance use treatment.  It does not stop with psychiatry and much hospital care and nursing home care is also rationed.

The fourth are the obvious conflicts of interest.  Rather than regulating health care denial as a money-making scheme, managed care decisions have basically been reified by reducing health care company liability for decisions, promoting the false narrative that they control costs and are interested in quality care, and not objecting to an arduous and usually futile appeal process.  It is a way for as much as a trillion dollars to be transferred to the business class now running health care and away from the people providing and consuming health care.

What happens when an ethically questionable AI system hits that landscape?  What were those algorithms trained on?  What chance does any patient or family have if a highly regarded/reified AI system makes a care decision against them?  That is probably the real short-term risk of AI in health care. Beyond that there is the question of how long these systems will last. Right now large language models (LLMs) scrape large amounts of data from sites they have access to including a large amount of copyrighted and proprietary data.  Systems and their businesses compete with one another on which system is the best.  Given the recent hacking scenarios is it possible that one system might attack the other strictly in the organized crime sense of “it's just business”.   According to the NYTimes that has already happened and it was solely due to decisions made by the AI.  

The fifth is selecting who should be replaced.  I suppose this is a variation of conflict of interest.  Ever since business people took over medicine and began to proliferate they have looked for ways to either eliminate physicians or control them. Over the past 50 years that has resulted in a 3,000% increase in managers and a 200% increase in physicians.  How is it that the knowledge workers in the organization are targeted for replacement while people doing much less complex work are not mentioned?  

The sixth problem that nobody talks about is the actual integration steps with current healthcare IT systems.  I was there for the massive conversion to the electronic health record (EHR) over 25 years ago.  I worked in two different healthcare systems using 3 different EHRs and over 2 decades hardly any of the integrations went well.  They all greatly increased the workload of physicians despite their purported efficiencies.  There are still problems today.  

Those systems are very primitive compared with AI, especially if you have a system of care with thousands of people using it. Will the AI be integrated with those very touchy and idiosyncratic systems?  I am referring to systems that in some cases required the intensive support of every clinician in the system in order to function.  I am very skeptical about how that integration is going to occur and whether there is adequate bandwidth and storage to support continuous and heavy AI utilization.  With both of those is the associated concern of security of confidential healthcare information. How can those systems be protected from new sophisticated AI attacks?

The seventh problem is cost.  AI is currently the main driver of the US stock market.  Many of those companies have invested billions of dollars as they are betting on revenue streams from an AI boom.  Healthcare always seem to pay the maximum retail price for any innovation whether it is a pharmaceutical, a medical device, or information technology.  Overhead is already huge for most organizations and there will be no great deals from a cash starved AI industry.   

At the time I am writing this – the bottom line for me is that AI systems are unregulated and potentially dangerous. Even if they turn out not to be – any medical application will requires extensive vetting to prevent errors that seem predictable and obvious at the machine level.  Beyond that it is important to know who owns and runs that machine.  We have had four decades of business management of medicine that has made the US healthcare system the most expensive, least efficient, and one with the least access.  We are on the verge of losing more access and facilities by the Trump administration's defunding of Medicaid. Putting a tool like AI in the hands of these organizations will greatly amplify their power over physicians and patients.  Getting someone into a hospital may look something like HAL9000 locking Dave Bowman out of his spacecraft.   

 

George Dawson, MD, DFAPA

 

Supplementary 1:

As I think about how AI has the potential to seriously disrupt life there are obvious concerns.  At the top of the list is hacking for the purpose of theft or destroying infrastructure.  Several countries have routinely hacked into government and business servers with the goal of stealing secrets and intellectual property.  If you do any research into how the systems important in our day-to-day life you will not get any satisfactory answers.

A recent example was the hacking of municipal water supplies in Minnesota.  In 4 cities the control of the water supply was disrupted and manual control needed to be used.  No contamination was noted but a cybersecurity incident was noted.  President Trump famously blamed it on the Governor of Minnesota at least until it happened in several other states.  From the NYTimes piece it is not a stretch to think that AI agents are much more likely to create successful hacks. A rogue state may not care how much damage occurs anywhere if they have access to the technology.

From a utility standpoint, the electric grid would be much more of a disruption if successfully attacked.  In the upper Midwest, gas pipelines are another obvious target.  Any major attack on either could potentially affect tens of thousands of people.  The only way to insulate these systems from cyber attacks is air-gapping.  Air-gapping just means no direct or wireless connect to the Internet.  Some of these systems claim that the most sensitive parts of their network are air gapped.  Is that enough if an AI agent has access to part of the network?  That answer is unknown at this time.

A scenario that has been written about and depicted in a television series is the hacking of the financial system. I have asked about security measures used by banks and other financial firms and typically get assurances that they have security departments that monitor cyber security threats and make compensatory changes.  Every year Russian hackers steal billions of dollars from the US financial system using conventional techniques of ransomware attacks, stealing unreleased financial reports of publicly held companies and trading stock based on that information, and malware bank fraud and wire theft. Russian hackers are given cultural hero status and protected from prosecution and extradition. What happens of they have access to more invasive AI agents to work on these same systems?

I don’t think there is any way to feel good about the current scenario given the lack of transparency surrounding AI, the steep incentives to invent more invasive forms, the lack of ethics involved up to and including overt criminal behavior, and governments that are either unwilling or unable to address these problems. 

Supplementary 2:  A Brief Note About Reification-

Reification: Means to treat an invisible or theoretical/abstract concept as if it were a physical object or a hard fact.  Computer science alters the meaning to turning something that is hidden, implicit, or abstract in programming into an object or an explicit data model.  

If I consider the sentence:  "The AI has decided that you should be discharged from the hospital."  This is a reification because the AI is not making a decision like humans do.  There is no moral or ethical decision making or empathy.  There is no intent.  Any decision making is the product of probabilistic algorithms. You can probably say the same thing about managed care except there is a clear intention to make money by the denial of care. 

References:

1:  Emanuel EJ, Baker-Butler A, Khosla N, Khosla V. Will Autonomous AI Exceed AI-Aided Physicians as the Best Medical Care? JAMA. 2026 Aug 17. doi: 10.1001/jama.2026.15380. Epub ahead of print. PMID: 42606838.

2:  Carter E.  If You Weren’t Worried About A.I., You Should Be After the Past Few Weeks.  New York Times August 13, 2026  https://www.nytimes.com/2026/08/13/opinion/ai-danger-openai-anthropic-models.html

3:  Migliarini M, Pizzini JP, Moresca L, Santini V, Spinelli I, Galasso F. Quantifying Self-Preservation Bias in Large Language Models. arXiv preprint arXiv:2604.02174. 2026 Apr 2.

4:  Pacing the Frontier:  A statement from 1,378 employees of frontier AI companies.  July 2026. https://www.pacingthefrontier.com/

Floridi, L., & Sanders, J. H. (2004). On the morality of artificial agents. Minds and Machines, 14(3), 349–379. https://doi.org/10.1023/B:MIND.0000035461.63578.9d

5:  Gabriel, I. (2020). Artificial intelligence, values and alignment. Minds and Machines, 30(3), 411–437. https://doi.org/10.1007/s11023-020-09539-2

Cited by: 2091

6:  Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 389–399. https://doi.org/10.1038/s42256-019-0088-2

7:  Lumbreras, S. (2017). The limits of machine ethics. Religions, 8(5), 100. https://doi.org/10.3390/rel8050100

8:  Moor, J. H. (2006). The nature, importance, and difficulty of machine ethics. IEEE Intelligent Systems, 21(4), 18–21. https://doi.org/10.1109/MIS.2006.80

 

AI related Executive Orders:

9:  Biden Administration:  Executive Order 14110—Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.  October 30, 2023. (63 pages) https://www.presidency.ucsb.edu/documents/executive-order-14110-safe-secure-and-trustworthy-development-and-use-artificial

10:  Trump Administration:  Removing Barriers to American Leadership In Artificial Intelligence. January 23, 2025. (3 pages) https://www.whitehouse.gov/presidential-actions/2025/01/removing-barriers-to-american-leadership-in-artificial-intelligence/

11:  Trump Administration:  Executive Order 14365- Ensuring a National Policy Framework for Artificial Intelligence.  December 11, 2025. (5 pages)  https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/

 

Photo Credit:

Thanks to Eduardo Colon, MD my friend and colleague. 

Gunfire In the Streets

 

True crime television is very big in the United States right now.  It is well produced and, in most cases, more engaging than typical fictional crime dramas.  Much of it shows real time police work with body cam and camera crew footage.  It gives a generally good presentation of police officers and detectives including demands of the job and their disclosure of the emotional toll and empathy for crime victims.  A cross section of police departments are sampled and they represent themselves very well.  I have been consistently impressed with their degree of professionalism and approaches to solving crimes. 

The striking part of these videos is the level of violence on the streets of the US and what the police have to contend with.  There seems to be a pattern of reactivity that many people have that leads them to try to evade the police or physically resist their efforts to detain them at all costs.  These aggressive confrontations often occur for trivial reasons.  A common example is a routine traffic stop that turns into a high-speed chase and multiple collisions or a single high impact crash.  The person apprehended has suddenly gone from a misdemeanor crime or traffic ticket to multiple felonies and jail time. 

Nowhere is the escalation more evident than firearm related crimes. I have seen many of these situations where the first action taken by the offender is to open fire on the police.  Open the car door and start shooting.  Pull off the covers in bed and come out shooting.  Wrestle with two officers in close proximity to your girlfriend, get loose, grab a gun, and start shooting. Hole up in your house or apartment and when the police show up start shooting through the windows and walls.  I saw one episode where 96 rounds from an assault rifle were fired though the apartment walls at police officers on a stairway (they did not fire a single shot).  And all of that gunfire took less than a couple of minutes.

Any reader of this blog knows that I support reasonable gun laws, abhor the current trend to maximize the number of guns on the street, and see easy access to guns as a major public health problem.  I also dispel the conventional political wisdom on guns as only being problems in the hands of the mentally ill or people taking antidepressants.  Gun violence in this country is directly proportional to high gun density and lax gun laws.  True crime TV quickly reveals that the primary equation for gunfire on the streets is an angry young man with access to firearms.  I have seen about a hundred of them interviewed on these shows. Most have no clear explanation for what happened.  They uniformly regret that it did happen – mostly because the consequences they are facing. And nothing would have happened if there was no easy access to handguns.  What would have been a fistfight became a gunfight with somebody dead and somebody in prison for the rest of their lives.

Although, the sampling for the shows is undoubtedly biased toward the sensational – I decided to take a look at what has been researched and written about in terms of gunfire on American streets.  For context I have posted the tree stump from outside the Spotsylvania Court house.  At least it was in that spot when it was sawed off by gunfire during a battle in the Civil War in 1864.  It was subsequently sent to the Smithsonian where it is displayed today.  For scale, historians estimate that during the three-day Battle of Gettysburg – 6 - 7 million rounds were fired from small arms and 50,000 artillery shells were fired from 630 cannons.  The combined weight of the bullets, shells, and gunpowder was about 569 tons.  That is a lot of gunfire considering the standard military rifle of the day was a single shot musket that fired a Minié ball. 

One of the main questions I thought about was how much gunfire exists on American streets today with much more liberal gun laws and firearms with high-capacity magazines.  How would a typical day on American streets compare with the Civil War figure?  Globally we know that Americans purchase about 15.5 -22.7M firearms (rifles, shotguns, handguns) per year over the past 5 years. According to a 2017 estimate from Small Arms Survey, US citizens held 393.3M legal and illicit firearms or about 120.5 firearms for every 100 citizens.  The next closest country was Yemen with 52.8 per 100 citizens.  No other high-income countries had more than 40 firearms per 100 citizens.  

Incredibly the government is prohibited by law from collecting data and statistics on ammunition purchases so it is typically estimated from financial reports.  According to those estimates, Americans purchase about 10-12B bullets per year. That is 1600 times the number of rounds fired in the bloodiest battle of the Civil War.  For a weight comparison, the metal projectiles from 10B 9mm rounds would weigh 82 Tons, the full cartridges 132 Tons.  Of course, within that number there are people who may fire thousands of rounds at a gun range and some people who may not fire a single round every year.  Some rounds will be fired from illegal firearms. Some rounds will be fired during hunting.  Is there any way to estimate how many gunshots may be a direct danger to the public?

There are several sources of information.  Public health statistics offer indirect data on the end results.   There are about 132 firearm related deaths per day.  40% are homicides and 57% suicides.  The remaining gun deaths are from accidental deaths, police shootings, and undetermined gun deaths.  Beyond that another 200 people are struck by gunfire each day. A total of 320 people per day are struck by gunfire in the US and killed or wounded.  For comparison during Vietnam War (1965–1973), an average of approximately 20 American soldiers were killed and 53 were wounded every day.

It is difficult to impossible to find valid estimate of all gunfire in the US on any given day.  Several sources extrapolate from 30M recreational shooters and 10M hunters hunting and shooting at target ranges every day.  That us typically in the range of millions of rounds fired each day.  Using the most conservative figure for ammunition purchased, that figure might be as high as 27M rounds per day across all settings (gun ranges, hunting, fatal and non-fatal shooting).  Acoustic detectors of gunfire are deployed in only 100 cities in the US. Some files of gunfire detection are publicly available but as difficult to read CSV files.  There is some controversy about whether these systems lead to more traffic stops in minority areas and I cannot locate any national estimates from these devices.  Two studies have documented that about 88% of acoustic detected gunshots were unreported to the 911 systems.   Some authors have also included celebratory gunfire as another source of gunfire on the streets.

As a physician interested in public health and reducing gun deaths – the only way any of this makes sense is political interference.  Gun extremism is a problem in this country.  It has been a progressive political force for the past 40 years.  It is the most likely reason for a lack of tangible data on total rounds of ammunition sold per year in this country or even more basic information about total gunfire occurring at all crime scenes in the US every year.  At every crime scene, all of the expended shell casings are marked, numbered and photographed by location.  Collecting and reporting that data to a central location would be an easy job in this era of information, even if it would be an underestimate (not all firearms eject expended cartridges). 

There is only access to the end result – gun deaths and injuries.  The politics is such that we have become desensitized to it – even though the rate of gun violence in the US is higher than the casualty rate in the Vietnam War.  That was a war that was protested as unnecessary, but nobody is protesting the unnecessary gun violence in this country.  It is being sold at the highest levels as the price we all must pay for the hallowed Second Amendment.  

If you watch true crime TV like I do – pay attention to the senseless gun violence. Pay attention to how much gunfire occurs and the setting.  Pay attention to what the veteran police officers say about it.  Most of those shows do a good job of showing the impact it has on families and the police officers. 

And ask yourself why we can’t do better.   

 

George Dawson, MD, DFAPA

 Image Credit:

This is an image of a tree stump that resulted when the tree was sawed off by small arms fire during a Civil War battle.  The details of that battle are from where it resides in the Smithsonian Institute.  

"Until May 12, 1864, this shattered stump was a large oak tree in a rolling meadow just outside Spotsylvania Court House, Virginia. That morning, 1,200 entrenched Confederates, the front line of General Robert E. Lee's Army of Northern Virginia, awaited the assault of 5,000 Union troops from the Second Corps of the Army of the Potomac. Twenty hours later, the once-peaceful meadow had acquired a new name, the Bloody Angle. The same fury of rifle bullets that cut down 2,000 combatants tore away all but twenty-two inches of the tree's trunk. Several of the conical minie balls (bullets) are still deeply embedded in the wood. Unusual objects of war, such as this tree stump, come to symbolize the horror and heroism of a great battle. Originally presented to the U.S. Army's Ordnance Museum by Brevet Major General Nelson A. Miles, the stump was transferred to the Smithsonian in 1888."

 

References:

1:  Carr J, Doleac JL. The geography, incidence, and underreporting of gun violence: new evidence using ShotSpotter data. Incidence, and Underreporting of Gun Violence: New Evidence Using Shotspotter Data (April 26, 2016). 2016 Apr 26.

2:  Hansen JH, Boril H. Gunshot detection systems: Methods, challenges, and can they be trusted?. InAudio Engineering Society Convention 151 2021 Oct 13. Audio Engineering Society.

3:  Kaufman EJ, Wiebe DJ, Xiong RA, Morrison CN, Seamon MJ, Delgado MK. Epidemiologic Trends in Fatal and Nonfatal Firearm Injuries in the US, 2009-2017. JAMA Intern Med. 2021;181(2):237–244. doi:10.1001/jamainternmed.2020.6696

2009–2017: ~120,232 firearm injuries/year, or 329 every day — comprising 34,538 deaths (~95/day) and 85,694 ED visits for nonfatal injury (~235/day)

4:  Kaufman EJ, Song J, Xiong R, Seamon MJ, Delgado MK. Fatal and Nonfatal Firearm Injury Rates by Race and Ethnicity in the United States, 2019 to 2020. Ann Intern Med. 2024 Sep;177(9):1157-1169. doi: 10.7326/M23-2251. Epub 2024 Jul 30. PMID: 39074371.

2019–2020: the burden rose to an average of 1 firearm injury every 4 minutes and 1 firearm death every 12 minutes — ~252,376 total injuries with 84,908 deaths over the two years (~346 injuries and ~116 deaths per day).

A key public-health caveat is that these figures capture people who are shot and wounded or killed, not all rounds fired — there is no comprehensive national surveillance system that counts total gunfire or nonfatal, uninjured shooting incidents; nonfatal shootings and firearm uses "far outweigh" homicides and suicide deaths

5:  Koper CS, Baas G, Taylor BG, Liu W, Sheridan-Johnson J. Validating open-source data on fatal police shootings against self-reports from a national sample of police agencies. Inj Epidemiol. 2025 Oct 21;12(1):68. doi: 10.1186/s40621-025-00625-6. PMID: 41121442; PMCID: PMC12538985.

A Few More Lost Pieces of the Antidepressant Discussion

 

There was a lengthy article in the New York Times 2 days ago on antidepressants (1).  The main premise of the article is that millions of children and young adults were prescribed antidepressants.  The antidepressants may have been effective for the crisis but now that they continue on them, they have side effects and want to discontinue them because they feel emotionally blunted and disconnected. The chemical imbalance trope was invoked several times.  The author proceeds to extend the argument to suggest that there are millions of people in this predicament, physicians are unwilling or unable to help, and HHS Secretary Robert F Kennedy (RFK) is going to solve the problem largely by reforming psychiatry purportedly by moving into the “rift between patient and doctor” and promising to “free the mental health of Americans from dependence on pills”.

I have addressed this issue in many places on this blog.  That includes my response to RFKs arguments, RFKs comparison of SSRI withdrawal to heroin withdrawal, and strategies to prevent the problem including exposing only those people to antidepressants who need them, using medications with lower withdrawal risk, and generally being aware of the problem. I have also written about how long the problem has existed, when the first review of the problem was written, and that it is a fundamental skill that all physicians should have. 

In the RFK response I was also able to find out that the detailed antidepressant tapering and transition instructions have been in the premier online medical resource UpToDate for at least 18 years. Any practicing physician with access has those details. The facts as they exist show that RFK does not have a rational or even reasonable approach to the problem.  His idea that antidepressants are overprescribed and impossible to stop obfuscates the real problems of undertreatment and both active removal of treatment resources while promoting more high-risk treatment by the Trump administration.  In brief, the RFK approach to mental health has taken a solvable problem of not enough resources and blown it up into another problem.

To be clear, I am not commenting on the anecdote in the New York Times piece.  I have no personal knowledge of the patients or families mentioned.  I think a presentation of any similar situations in medicine should raise the questions: Why is a treatment being continued if it is not at least partially effective for the symptoms?  Why is a treatment not discontinued if the patient or family would prefer, they not take it?  What other treatments were tried before a medication was tried and what other treatments were tried concomitantly?  Invariably that all comes down to the expertise of the physician and knowing that the primary role to give the best advice to the patient and not make up their mind for them.  It is their job whether to take it or not. Almost everybody I encountered in practice had a preference to not take any medications. They either changed their mind about that after a lengthy informed consent discussion or not.  Either answer was fine with me. 

A confounding factor in any longitudinal analysis of medication effects is how people adapt to medications over time.  The first time I discontinued a maintenance antidepressant was in 1986.  That patient was a blue-collar worker who had been on doxepin for many years for migraine headaches and depression.  When I started seeing him his main problems were fatigue and hypersomnia.  I tapered and discontinued the doxepin.  He did well and eventually told me that he did not realize it at the time but he felt like he had the flu while he was on doxepin and that resolved after it was stopped.

Just as people can be unaware of long-term side effects after years on any medication – they can also get habituated to the therapeutic effects. That is the main reason people need close follow up when medications have been discontinued. Those effects go far beyond treating the primary disorder. People will report not feeling as well, not thinking as clearly, and not being as mentally flexible as when they were taking the medication.  In some cases they will experience a recurrence of a secondary disorder that was treated by the antidepressant like migraine headaches or back pain.  None of the symptoms described are due to withdrawal or a clear recurrence of the primary disorder but they would prefer to stay on the medication.

My main concern with the antidepressant controversy is that it is essentially a polarizing political argument at this point.  As far as psychiatry goes, nobody is talking about psychiatry the way I practiced it, how my colleagues practiced it, or how I trained residents to practice it.  The idea that any medication is good or evil is absurd.  The idea that the prescribers of medication are forcing it on clueless unsuspecting patients for the benefit of the pharmaceutical industry is equally absurd.  The most absurd arguments is that RFK or antipsychiatrists are going to save everyone on an antidepressant or psychiatric medication.  How will that happen when all the evidence is ignored and you have the expertise of a political podcaster?

To end this post, I want to include an important part of the antidepressant argument that I have ignored until now and that is how many people stop taking them.  There is a constant drumbeat of how many people take them and inaccurate suggestions that too many take them, so why would I think people stop taking them?  For many years, I worked for a healthcare company and I was on their Pharmacy and Therapeutics (P&T) Committee.  As such we were gatekeepers for medications that would be listed on the health plan formulary and were available to subscribers. We assessed the scientific data on efficacy but also the cost. At some point we also discussed compliance (now referred to as adherence) to the prescribed medications.  We found that a significant number of people never refilled their second prescription for antidepressants. At that point we began sending out reminder letters about the importance of adhering to the medication until there was agreement with their doctor that they should stop.

This occurred in the time frame of about 1995-2005. It was a health plan wide initiative meaning that most of the antidepressants were prescribed by primary care doctors. Stigma and the stigma of psychiatric medications was still hotly debated. There was still active misinformation about psychiatry and psychiatric medications.  There is always a personal bias to not take medications unless they are absolutely necessary. In terms of pure antidepressant prescribing I am sure that in many cases they were prescribed unnecessarily for self-limited crisis situations.  The important information here is that large number of patients stop taking these medications and that is never discussed.  It would be in opposition to the usual political argument of excessive prescribing and an inability to stop.       

What would that look like at a national and international level?  A commonly cited statistic is that 50% or persons prescribed antidepressants discontinue them in the first 6 months although many of the studies have lower estimates (see table below).  There are also several studies that estimate population wide use of antidepressants as both a percentage of the population (2-4) and absolute numbers (5).  The 2023 National Health Interview Survey found 11.4% of all adults ≥18 took antidepressant medication. The total population is 269.8M people so that is about 31M people.  Based on the available discontinuation percentages of 22 to 42.9% that means in any given year – 6.82 to 13.2 million people stop taking antidepressants.  Considering that as many as half of the antidepressants starts are for conditions other than depression it is likely that as many people are starting antidepressants as stopping them.   At least some of the survey data indicates that some of the reasons for stopping has to do with negative press and misinformation (fear of dependence)  – a known factor in the nocebo effect.

 

The above estimate is just that.  Four of the 5 studies are surveys.  There is no patient level data on a national scale that looks specifically at the antidepressant discontinuation issue.

There is data that looks at what happens at the treatment level.  That comes from a Danish register study of 66,540 older adults aged 65 or greater treated for depression who received a first-time antidepressant prescription between 2006 and 2016 (7).  Register studies are observational but they have the advantage of detailed information at the individual patient level about interventions when the databases are linked to clinical and pharmacy information for nationwide health plans. In this study, 33.7% of patient stopped antidepressants within 6 months, 26.5% gradually stopped over a period of 2 years, and 39.8% were on antidepressants for the entire 3-year period.  They studied the social determinants of this antidepressant use and also compared use to the recommended guidelines for antidepressant treatment in Denmark. Those guidelines suggest at least 6 months of maintenance treatment after initial remission and 2 years of maintenance for recurrent depression.  One of the social determinants was living in a non-urban area and the authors suggest this may be related to access to mental health specialty care.

The interesting aspect of the information in the Danish study was the detailed information across three general trajectories of antidepressant use.  Those trajectories are commonly seen in psychiatric practice and are far more realistic than what is typically portrayed in the media.  That includes the NYTimes article.       

When you read an article like that one or one of the many I have noted in the past – ask yourself what you really learned.  Like all medications antidepressants have side effects.  In fact, I routinely advised patients was that one person in six or seven would not tolerate them at all either due to initial side effects or withdrawal effects.  In clinical practice or real life - I never met a patient who told me they liked taking medications of any type.  Neither of those factors was a deterrent to trying a medication for most people.  The reason is that they were seeing me was for a severe, life changing problem and they had tried many other interventions. The Danish study (7) showed that only about 3% of that sample was ever hospitalized for severe depression.  Those were the patients I was treating.  

The psychiatric treatment of people over time is a dynamic process. It generally involves more than just medication with close attention to psychological factors and necessary lifestyle interventions. It requires a close collaborative relationship between the patient and the psychiatrist that includes a focus on optimizing therapy, minimizing or eliminating side effects and paying close attention to patient preferences. With that general approach, nobody should regret taking a medication longer than they should.  Nobody should put up with significant side effects.  And nobody should take a medication that is not working.  All of that is open for discussion.

Where I come from there is no rift between the patient and the doctor for RFK to fill.  And if there was – he is the wrong man to fill it.  

      

 George Dawson, MD, DFAPA

 

References:

 

1:  Bromley C.  A Generation on Antidepressants Searches for the Exit.  New York Times.  August 7, 2026.

2: Brody DJ, Gu Q. Antidepressant Use Among Adults: United States, 2015-2018. NCHS Data Brief. 2020 Sep;(377):1-8. PMID: 33054926.

3:  Mojtabai R, Olfson M. National trends in long-term use of antidepressant medications: results from the U.S. National Health and Nutrition Examination Survey. J Clin Psychiatry. 2014 Feb;75(2):169-77. doi: 10.4088/JCP.13m08443. PMID: 24345349.

4: Fu G, Li M, Lang X, Luo M, Chen S. Trends in depression and antidepressants use by social determinants of health among adults in the United States: Data from NHANES 2005-2018. J Affect Disord. 2026 Feb 1;394(Pt B):120662. doi: 10.1016/j.jad.2025.120662. Epub 2025 Nov 10. PMID: 41224008.

5:  Chai G, Xu J, Goyal S, et al. Trends in Incident Prescriptions for Behavioral Health Medications in the US, 2018-2022. JAMA Psychiatry. 2024;81(4):396–405. doi:10.1001/jamapsychiatry.2023.5045

6:  Elgaddal N, Weeks JD, Mykyta L. Characteristics of adults age 18 and older who took prescription medication for depression: United States, 2023. NCHS Data Brief. 2025 Apr;(528):1-9. DOI: https://dx.doi.org/10.15620/cdc/174589.

7:  Ishtiak-Ahmed, K., Rohde, C., Köhler-Forsberg, O., Christensen, K.S. and Gasse, C. (2024), Depression Treatment Trajectories and Associated Social Determinants: A Three-Year Follow-Up Study in 66,540 Older Adults Undergoing First-Time Depression Treatment in Denmark. Int J Geriatr Psychiatry, 39: e70006. https://doi.org/10.1002/gps.70006.

 

Supplementary 1: The decision about medications is common in any country with the availability of advanced therapeutics.  I personally take three medications every day that I would prefer not to take. I have to self-monitor for side effects including blood pressure readings every day.  If a doctor tries to give me a temporary antibiotic prescription – I personally do a drug interaction check and let them know if that medication is compatible.  I decide to take the medication not because it makes me feel better every day but because I know the cumulative effects of not taking it are potentially very bad and therefore, I decide to take it.  I have experienced side effects and complications that I had to figure out myself and that doctors missed.  This is all part of what it means to take a prescription medication. 

All of the steps I take to protect myself are the same steps I took with any medication I prescribed for patients.