The Lack of Adequate Scientific Evidence Regarding Physician-assisted Death for People with Psychiatric Disorders Is a Danger to Patients

In this edition of the Canadian Journal of Psychiatry, van Veen and colleagues provide a thorough scoping review of the literature regarding physician-assisted death (PAD) in patients with a psychiatric disorder (PPD). ¹ In view of recent legal challenges and efforts in Canada to include “irremediable” mental illness as eligible for PAD (called medical assistance in dying/MAID in Canada), reading this review was exceedingly demoralizing. The decision of whether to allow PAD for mental illness alone is highly consequential with important implications for psychiatric practice and suicide prevention. It may also be an irrevocable one since, once a right is granted, it can become politically difficult or impossible to take away. As with any newly introduced delivery of a medical intervention, it is the obligation of the scientific community and, in particular, those who advocate for it to demonstrate rigorously that this change benefits patients and does not cause harms that outweigh that benefit. Yet, it is clear that governments are moving forward based mainly on selective expert conjecture in a vacuum of scientific evidence and that should raise serious concerns regardless of one’s personal and/or professional beliefs about this complex and weighty issue.

The scoping review mainly identified conceptual studies such as commentaries and essays. Of these 32 papers, 20 had a main finding which described concerns about PAD in PPD and/or suggested that it be banned outright in such cases. ¹ The review identified 8 empirical studies, of which only 5 examined actual patients seeking or receiving PAD (total n = 310). ¹ None examined impact on psychiatric care in general or on suicide rates, 5 included evidence indicating expert disagreement about whether the mental health conditions were irremediable, and 2 included evidence that patients were refusing potentially effective treatments in favor of PAD. ¹

The problem of the limited and low-quality scientific evidence available to date is compounded by imprecision and spurious reasoning within the conceptual articles which really should not have been called “studies.” There is a crucial difference between an incurable and an irremediable condition. Chronic diseases such as diabetes are often incurable but that has no bearing on whether or not they are remediable. To be eligible for PAD in Canada, a person must “experience unbearable physical or mental suffering” from his or her illness. ² If a person’s suffering can be made bearable, they become ineligible for PAD and, therefore, that is the logical way to define the threshold for irremediability.

van Veen et al. mention three essays that “cite clinical studies that show that some psychiatric patients will never recover.” ¹ Yet, lack of recovery is not evidence that a condition is unbearable. Two of these articles emphasize results of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study as evidence of irremediability of mental disorders given only a two thirds remission rate at trial endpoint. However, that outcome is actually a good one when one understands that STAR*D was essentially a study randomizing depressed patients to sequential, low-intensity outpatient pharmacological treatment with limited use of psychotherapy (only 3.5% of subjects received cognitive therapy). Lack of full remission is also very different than symptoms being unbearable. It is an unjustifiable and irresponsible leap to infer based on this data that major depressive disorder is irremediable. STAR*D does not comment on whether the suffering of nonremitters was intolerable nor whether it could have been made tolerable had they been offered additional treatments that are commonly used for depression, including more targeted and individualized medication trials, evidence-based psychotherapies, light therapy, brain stimulation techniques such as electroconvulsive therapy, complementary and alternative medicines, exercise, and dietary changes. ³ These same concerns apply to each of the studies cited which all examine a narrow range of treatments that fail to achieve 100% recovery. Whether mental illnesses are ever truly irremediable or simply very challenging to remediate in some cases is a worthy yet unanswered question that is scientifically testable and reliance on a handful of essays that overstate current scientific knowledge is an unacceptable way to inform public policy.

A proper analysis of the utility of a proposed intervention must also involve careful attention to both potential benefits and potential harms. Even if we bypass the crucial question of whether mental illnesses are ever irremediable and assume that, in rare instances, they can be, do the benefits of providing PAD to such patients outweigh harms to others and to society? For example, would PAD lead to premature death in some people whose illnesses could actually have been remediated? This is really a question of the ratio between true positives and type I errors (i.e., incorrectly identifying cases as irremediable). While a study design of this kind would likely be very challenging because of the aforementioned issues related to demonstrating irremediability, estimates could be made and qualified with caveats and assumptions. We would then need to address questions such as whether we would be prepared to tolerate one “correct” application of PAD for every false-positive premature death it caused? What if the ratio was 4:1 or 1:4? The fact that no one seems to be asking these questions is an indicator of concern.

As mood disorders experts who spend a substantial proportion of our time helping patients understand that life is worth living when their illnesses often relentlessly suggest the opposite, we are concerned that legalizing PAD for PPD will fundamentally harm psychiatric care. Suicide is the most common cause of death due to illness in the prime of life, ⁴ and it arises through a series of cognitive distortions leading to the conclusion that death is the best available way of handling a painful life. The reasonably foreseeable death criterion has historically functioned as a bulwark acting to conceptually separate PAD applied to prevent a painful death from suicide which occurs in response to a painful life. Removing that criterion effectively eliminates this distinction. The overwhelming majority of suicidal patients seeking our care endorse painful lives and subjectively unbearable suffering (if it was bearable, they wouldn’t be in our offices or hospitals). Endorsing death as a sometimes desirable strategy for treating a painful life and lowering the barrier to accessing it are exactly counter to evidence-based suicide prevention efforts. ⁵

Nevertheless, there have been no studies rigorously testing the impact of legalizing PAD for PPD on routine psychiatric care or suicidal patients. One U.S. study examining suicide rates after legalization of PAD for terminal illness found that rates of nonassisted suicide were 1.1% higher in states that legalized PAD although this difference was not statistically significant. ⁶

Furthermore, the Canadian Psychiatric Association (CPA) guidelines for responsible media reporting ask that journalists “[avoid] portraying suicide as achieving results and solving problems” ⁷ This is because of what is now a robust literature on suicide contagion showing that messages of death result in more suicides across a population, while stories of resilience and mastery of suicidal crises can have the opposite impact. ⁸ Whether legalization of PAD for PPD might lead to a similar social contagion phenomenon also remains unexplored. A study we previously conducted found that media articles on PAD were associated with an increased probability of higher subsequent suicides that approached but did not reach statistical significance (OR = 1.13; 95%CI, 0.99 to 1.28). ⁹ Whether legalizing PAD for PPD and associated media exposures in at-risk patients might result in a suicide contagion phenomenon is once more a testable yet unanswered question.

The two of us, like the various essayists, do not share exactly the same personal views on PAD. But our own views, like theirs, are largely irrelevant. PAD for PPD is a potential intervention as part of medical care, and as such, it should not be exempt from usual scientific evidentiary standards in favor of reliance on the strongly held personal opinions and philosophical arguments of interested parties. The CPA recently released a position statement which is agnostic to whether PAD should be allowed in PPD. ¹⁰ This was most unfortunate. The CPA is not obliged to take a moral stand, but it ought to advocate for adequate scientific evidence to inform public policy. Its position statement is an abdication of that responsibility.

We cannot quantify, according to any of the usual scientific standards, whether PAD for PPD will benefit or harm patients and to what degree. The van Veen review makes it clear that no one else can claim to do so either. We believe that the only rational approach is to insist on further study before expanding legislation. That perspective is informed by our experience in the medical field with other rushes to bypass rigorous science and to rely on conjecture and dubious “studies.” The opioid epidemic is the best recent example of a situation where assumptions made about the safety of an intervention and a hurry to apply it to help patients in the absence of proper testing led to tragic and irreversible consequences. Meaningful changes to medical care always require careful study and high-quality evidence. Earnest people may disagree philosophically in this area, but personal opinions cannot bypass the usual scientific process and, in this regard, those proposing expansion of the use of PAD have fallen woefully short. The fact that they have not been held to the usual requirements of medical science is absurd and even shameful. Sadly, no one in power seems to have noticed.