Physician-Assisted Suicide and the Perils of Empirical Ethical Research | End of Life | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Invited Commentary
Health Policy
August 9, 2019

Physician-Assisted Suicide and the Perils of Empirical Ethical Research

Author Affiliations
  • 1The Kennedy Institute of Ethics, Departments of Medicine and Philosophy, Georgetown University, Washington, DC
JAMA Netw Open. 2019;2(8):e198628. doi:10.1001/jamanetworkopen.2019.8628

Al Rabadi et al1 compare statistics on physician-assisted suicide (PAS) available from public databases for the states of Washington and Oregon and find similar profiles and trends, which is unsurprising given the similarity of the laws and demographic characteristics of these states. Among the unanswered questions are what such a study can contribute to medical ethics (about PAS or any other ethical controversy) and what the limits are of such work.


First, it should be noted that the medical literature is, in general, favorably disposed toward the empirical and the new. Although this predilection is often advantageous for scientific progress, it introduces a problematic bias when applied to ethical questions. The appeal of the study by Al Rabadi et al1 is that it is empirical, and by comparing data from 2 states for the first time, it can be considered novel. Because there are new reports each year and the practice of PAS is legal in only a few states, descriptive reports about PAS are published frequently. This means, however, that articles defending the ethical status quo (ie, against PAS) tend to be shut out of the medical literature because they are not reporting anything new and, therefore, cannot have any data. The result is an impression of growing acceptance of PAS, but it really represents an artifact of a scientific bias.

The frequent publication of statistics on PAS and euthanasia also imparts another implicit ethical bias. Although conducting health services research about an ethically disputed question appears ethically neutral, one already presumes without argument that the service to be delivered is ethically good. Therefore, the standards for assessing good and bad within the framework assumed by the research are limited to questions of access, efficiency, effectiveness, implementation, safety, and regulation. This takes the central ethical question of whether the service ought to be delivered in the first place off the table, suppressing critical ethical debate. Frequent publication of studies like that by Al Rabadi et al1 have the effect of “normalizing” the practice, inuring readers to the practice and concealing ethical concerns.

Moreover, data cannot tell a society or a profession what ought to be done—a consequence of what philosophers call the fact-value distinction. Empirical reports alone cannot answer normative ethical questions. Whether just 1 person or 100 000 persons legally avail themselves of lethal prescriptions cannot tell us whether the practice is right or wrong.

Likewise, the language used by empirical investigators can obscure the ethical issues at stake. Language can shape people’s attitudes and approaches toward controversial questions. There is no scientific basis for deciding on terminology, except, perhaps, marketing science. The language around PAS includes such terms as “death with dignity,” “assisted dying,” “assisted death,” and “hastened death.” Al Rabadi et al1 use the phrase “medical aid in dying.” These terms are imprecise. Giving a dying person an aspirin might count as aid in dying every bit as much as 5 g of secobarbital. These terms obscure distinctions of both ethical and empirical importance but have been shown to improve public opinion polls in favor of the practice. Serious proponents of PAS, such as the philosopher Peter Singer,2 prefer to describe it as a form of rational suicide and do not shy away from the most accurate description.

The public databases on which the study of Al Rabadi et al1 relies are not designed for research. They rely on self-reported data supplied by clinicians, who are generally not present at the time of ingestion of the fatal dose of drug by the patient and have an incentive to make their own roles look good. These data have never been validated, to my knowledge. The information collected is purposefully thin to protect prescribers’ confidentiality. There are no data on unreported cases. The health department is not permitted to investigate any allegations of abuse.3 Although these laws do suggest referral to psychiatrists if there are questions about the patient’s decisional capacity, and it is known that large numbers of terminally ill patients, including those seeking PAS, are depressed, since 2003, only 0% to 4% of such patients have been sent for psychiatric referral.3 There are also data suggesting, for instance, that large numbers of patients who ingest these drugs are at least transiently awake and suffering, leading to calls for anesthesiologists to administer euthanasia using advanced medical monitoring techniques as the only way to ensure the sought-after peaceful death.4 Contrary to the characterization of Al Rabadi et al,1 Sinmyee et al4 are advocates, not opponents, of PAS and euthanasia. These concerns make the conclusion of Al Rabadi et al1 that PAS in Oregon and Washington is safe, reliable, and effective an invalid inference.

What We Know

At present, the number of reported cases, following the law, remains small but increasing steadily. Those who make use of the law tend to be white, wealthy, and educated.1,3 Despite public arguments that PAS is needed to avoid excruciating pain and other symptoms, the reasons attributed to patients who seek PAS are not uncontrolled symptoms but lost autonomy, independence, and control.1,3 About one-third of patients die without taking the drugs,3 which may suggest that patients only wanted the security of having a way out, but it could equally indicate that they died before using the drugs or changed their minds about using them. A small (but growing) number of physicians write the prescriptions.3 For example, one of the authors of the study by Al Rabadi et al1 wrote 15% of all PAS prescriptions in Oregon in 2018.5

There is also much that we do not know. We have few direct, valid studies of the patients and practitioners. We do not know how many cases are unreported to the databanks. We do not know how many patients are pressured into obtaining prescriptions. We do not know how many patients engage in “doctor shopping,” finding someone who will agree to their request if turned down by a given physician. We do not really know how often the process goes awry. We know little about the after-effects on practitioners and family, although there are reports of posttraumatic stress disorder.6 We also need more data on suicide contagion, because preliminary reports7 suggest increased rates of suicide in the general population of states that have legalized PAS.

Slippery Slopes?

Safeguards built into the law are coming to be seen as barriers. As already noted, few patients are ever referred to psychiatrists. A new law in Oregon (Oregon Senate Bill 579) now allows a patient to bypass the waiting period and take the pills within 2 days, and legislation has been passed by the Oregon House (Oregon House Bill 2217) to allow injection of lethal drugs, a hair’s breadth away from euthanasia.5 There are increasing calls for permitting patients with dementia to be able to authorize their deaths through advance directives. In Belgium and the Netherlands, 5% of all deaths are by euthanasia and the indications have expanded to include psychiatric illness and life completion.8,9 Euthanized patients are now regular sources for organ donation.10 Although there is no empirical proof that the United States will follow these trends if PAS is more widely adopted, the logic that justifies PAS inexorably points in this direction. Studying these trends empirically will not prevent them from occurring. Are we willing to entertain a serious ethical debate, based on reasoned argument, or will we be content merely to file empirical reports on whatever fate befalls us?

Back to top
Article Information

Published: August 9, 2019. doi:10.1001/jamanetworkopen.2019.8628

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Sulmasy DP. JAMA Network Open.

Corresponding Author: Daniel P. Sulmasy, MD, PhD, Kennedy Institute of Ethics, Healy 419, Georgetown University, 3700 O St NW, Washington, DC 20057 (

Conflict of Interest Disclosures: None reported.

Al Rabadi  L, LeBlanc  M, Bucy  T,  et al.  Trends in medical aid in dying in Oregon and Washington.  JAMA Netw Open. 2019;2(8):e198648. doi:10.1001/jamanetworkopen.2019.8648Google Scholar
Singer  P.  Practical Ethics. 3rd ed. New York, NY: Cambridge University Press; 2011:156. doi:10.1017/CBO9780511975950
Oregon Health Authority. Death with Dignity Act annual reports. Updated February 15, 2019. Accessed July 13, 2019.
Sinmyee  S, Pandit  VJ, Pascual  JM,  et al.  Legal and ethical implications of defining an optimum means of achieving unconsciousness in assisted dying.  Anaesthesia. 2019;74(5):630-637. doi:10.1111/anae.14532PubMedGoogle ScholarCrossref
Foden-Vencil  K. Oregon considers expanding its death with dignity law. Published May 10, 2019. Accessed July 10, 2019.PubMed
Wagner  B, Müller  J, Maercker  A.  Death by request in Switzerland: posttraumatic stress disorder and complicated grief after witnessing assisted suicide.  Eur Psychiatry. 2012;27(7):542-546. doi:10.1016/j.eurpsy.2010.12.003PubMedGoogle ScholarCrossref
Jones  DA, Paton  D.  How does legalization of physician-assisted suicide affect rates of suicide?  South Med J. 2015;108(10):599-604. doi:10.14423/SMJ.0000000000000349PubMedGoogle Scholar
Chambaere  K, Vander Stichele  R, Mortier  F, Cohen  J, Deliens  L.  Recent trends in euthanasia and other end-of-life practices in Belgium.  N Engl J Med. 2015;372(12):1179-1181. doi:10.1056/NEJMc1414527PubMedGoogle ScholarCrossref
Kouwenhoven  PSC, van Thiel  GJMW, van der Heide  A, Rietjens  JAC, van Delden  JJM.  Developments in euthanasia practice in the Netherlands: balancing professional responsibility and the patient's autonomy.  Eur J Gen Pract. 2019;25(1):44-48. doi:10.1080/13814788.2018.1517154PubMedGoogle ScholarCrossref
Gilbo  N, Jochmans  I, Jacobs-Tulleneers-Thevissen  D,  et al.  Survival of patients with liver transplants donated after euthanasia, circulatory death, or brain death at a single center in Belgium.  JAMA. 2019;322(1):78-80. doi:10.1001/jama.2019.6553PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    3 Comments for this article
    Fact Can Inform Value
    Dena Davis, JD, PhD | Lehigh University
    Of course Sulmasy is correct that "empirical reports alone cannot answer normative ethical questions," but they can play an important role. In the case of physician-assisted suicide (or physician aid-in-dying [PAD]), there was a great deal of concern about justice when Washington and other states began to pass these laws. Ethicists were concerned that people who were already in danger of being marginalized--the poor, the underinsured, racial and ethic minorities, eg. --would be the primary users of the law. Feminist scholars worried that women, who are more likely to be poor and to outlive their spouses, would outnumber men. Much was made of the fact that Dr. Kevorkian's clients were primarily women.

    We now know that the typical PAD consumer is male, with health insurance and some college education. 96% of users are nonHispanic whites. Clearly, this is not a case of an overloaded healthcare system encouraging "problem patients" to end their lives so we don't have to take care of them.

    These facts don't make PAD right or wrong, but they do remove what had been a powerful normative concern about justice. Because of these empirical studies, there are certain arguments that ought to be off the table. This is an important advance in the ethical debate..
    Evidence-based Ethics and Medical Aid in Dying (MAID)
    William Southworth, M.D. | Practicing internal medicine physician, Oregon. Affiliated with End of Life Choices--Oregon
    As an Oregon physician who goes to the bedside of terminally ill individuals to consult and occasionally prescribe death-hastening medications according to and in full compliance with Oregon’s Death With Dignity ACT, I take exception to several points in Dr. Sulmasy’s commentary. He criticizes Rabadi (Trends in Medical Aid in Dying in Oregon and Washington) charging the authors with “normalizing” the practice while overlooking or “concealing” ethical concerns. Sulmasy reveals his own bias against the established legal practice of medical aid in dying (MAID) in these two states and elsewhere. I agree that data alone do not answer the question of whether MAID is right or wrong from an ethical perspective. Ribadi make no attempt to address ethical considerations but simply analyze and interpret empirical data. Such data are a valuable resource for gaining knowledge and improving medical practice. I would ask Dr. Sulmasy what kind of evidence would be useful for ethical research if empirical evidence is inadequate for this purpose?

    Ethical attitudes toward MAID continue to be divided based on cultural factors and beliefs, not on evidence. But it is settled legal and medical practice in now 10 states and the District of Columbia. Other jurisdictions will eventually legalize MAID as more people demand that this choice be open to them when beset with a fatal illness.

    Dr. Sulmasy is apparently an ethical scholar in a Jesuit university, which makes his bias understandable. It might have been appropriate for him to mention this when asked to disclose conflicts of interest before publishing his commentary.
    In reply
    Charles Blanke, MD | SWOG Cancer Research Network
    We noted with interest Daniel Sulmasy’s invited commentary “Physician-assisted suicide and the perils of empirical ethical research”. His opinions opposing medical-aid-in-dying (MAID) are consistent with his role as an appointee to the Pontifical Academy for Life, an organization dedicated to promoting the Catholic Church’s “consistent life ethic”. Sulmasy criticized our manuscript concerning MAID trends in the Pacific Northwest, saying it does not contribute to medical ethics. He also accused the authors of “concealing ethical concerns.” Our article never intended to address the ethics of MAID, as its stated objective was “to evaluate and compare patterns of MAID use between the states with the longest running US death with dignity programs”.

    Sulmasy suggests the term "medical-aid-in-dying" is imprecise. That term describes exactly what MAID offers. He himself unfortunately uses the outdated phrase “physician-assisted suicide.” In Oregon, a MAID death certificate diagnosis is the underlying terminal illness. Comparing MAID to, or labeling it suicide does patients and public understanding a disservice. He also tangentially decries euthanasia, currently illegal and not the procedure described in our report.

    Sulmasy questions the manuscript’s underlying public database, stating physicians supplying its data wish to make their own roles look good. Statute mandates the collection of demographic information including race, age, and diagnosis; these data are simple facts not conceivably subject to manipulation. Completeness is also enhanced by the prescribing physicians being approached for information within ten days of the patient’s death from any cause.

    Sulmasy suggests large numbers of those seeking MAID are depressed, though he presents no data that depression affects decisional capacity (the possession of which is required by law) or that its treatment would affect a terminally ill patient’s decision to follow-through with MAID. We ourselves called for further study of this important topic.

    Sulmasy alludes there exist data suggesting that large numbers ingesting MAID agents are transiently awake or suffering. His reference suggests re-emergence from coma occurs in up to 2% of cases, but large-volume Oregon data put the actual percentage at 0.58%. He also debates whether the authors of that reference are pro- or anti-MAID, objecting to our conclusion that MAID in the Pacific Northwest is “safe, reliable, and effective.” We cited different, completely independent data sets to support out conclusion.

    He also offers a traditional slippery slope argument, stating a new law in Oregon allows a patient to take the pills two days after prescription. This is incomplete and misleading. Recently signed Oregon law allows an exception to the 15-day waiting period only for those expected to die within that two-week+ period.

    Finally, Sulmasy suggests more research is needed. We could not agree more. He believes our paper is not sufficient to support “serious ethical debate”. We concur. However, MAID is legal in 11 US jurisdictions, with many more to come. Debate about the existence of this already-legal procedure seems far less important than investigating further patient protections, and Sulmasy and we clearly both believe future, prospective studies, as difficult as they may be to conduct, are warranted.