Abstract and Introduction
Objectives: Several US states have legalized or decriminalized physician-assisted suicide (PAS) while others are considering permitting PAS. Although it has been suggested that legalization could lead to a reduction in total suicides and to a delay in those suicides that do occur, to date no research has tested whether these effects can be identified in practice. The aim of this study was to fill this gap by examining the association between the legalization of PAS and state-level suicide rates in the United States between 1990 and 2013.
Methods: We used regression analysis to test the change in rates of nonassisted suicides and total suicides (including assisted suicides) before and after the legalization of PAS.
Results: Controlling for various socioeconomic factors, unobservable state and year effects, and state-specific linear trends, we found that legalizing PAS was associated with a 6.3% (95% confidence interval 2.70%–9.9%) increase in total suicides (including assisted suicides). This effect was larger in the individuals older than 65 years (14.5%, CI 6.4%–22.7%). Introduction of PAS was neither associated with a reduction in nonassisted suicide rates nor with an increase in the mean age of nonassisted suicide.
Conclusions: Legalizing PAS has been associated with an increased rate of total suicides relative to other states and no decrease in nonassisted suicides. This suggests either that PAS does not inhibit (nor acts as an alternative to) nonassisted suicide, or that it acts in this way in some individuals but is associated with an increased inclination to suicide in other individuals.
A significant stream of literature has focused on how socioeconomic factors and policy changes may affect suicide rates at the population level. It is well established that adverse economic conditions can lead to significant increases in suicide rates.[1,2] Individual-level attitudes (eg, toward religion) also are known to affect suicides.[3–6] Other authors have found that stricter alcohol regulations can be associated with fewer suicides, whereas research suggests that the legalization of marijuana for medical purposes may have led to a reduction in suicide among boys and men. Reporting of celebrity suicide also seems to have population-level effects on suicide rates.[9,10] A policy area that has received surprisingly little attention is the effect of changes to the legal code addressing suicide itself.
Several US states have moved either to legalize or to decriminalize some forms of assistance with suicide. In 1998, Oregon became the first state to legalize physician-assisted suicide (PAS) for patients with terminal illness. Washington state passed a similar law in 2008 and Vermont followed in 2013. In addition, in 2010 a Montana court decision declared that "physician aid in dying" was not contrary to legal precedent or public policy. In 2013, there were 73 deaths under the assisted dying law in Oregon and 133 in Washington state.
The likely effect of legalizing PAS on suicide rates is not easy to predict a priori. It is necessary to distinguish between those deaths that conform with PAS law versus suicides outside this legal framework (which we term "nonassisted suicides," although in practice these would include some assisted suicides outside the parameters of the law). The rationale of PAS laws is to enable people who would otherwise have died from an underlying illness such as terminal cancer to end their lives at an earlier stage with the assistance of a physician. In the absence of PAS, however, there will be people who are seriously ill who die by suicide.[15,16] A study from Switzerland found that in the 20% of nonassisted suicides that involved physical illness, "the range of physical illnesses reported with suicide is similar to that reported with assisted suicide."
Similarly, in Oregon, approximately 25% of individuals carrying out nonassisted suicides were found to have had physical health problems, whereas in the cohort of men older than 65, 66% had a physical illness (26% with cancer, 25% with chronic pain, and 16% with heart disease). The legalization of PAS could provide an alternative to nonassisted suicide for some people with chronic or terminal illnesses. If so, the direct effect of legalizing PAS would be for the total number of intentional self-inflicted deaths (including assisted suicides) to increase but for deaths by nonassisted suicide to decrease.
There also may be significant indirect consequences of legalizing PAS. Richard Posner has conjectured that legalizing PAS may have the effect of reducing the total number of suicides and postponing those that do occur. The knowledge that PAS is available for people who are physically incapacitated could enable such patients to delay their decision to attempt suicide. Furthermore, some may be contemplating suicide because of an overly pessimistic belief about the progress of their disease and/or about their ability to cope with their declining condition. If people delayed their attempt at suicide they might then come to see that they had been mistaken. As a result, "if physician-assisted suicide in cases of physical incapacity is permitted, the number of suicides will be reduced. Moreover, in the fraction of cases in which suicide does occur, it will occur later than if physician-assisted suicide were prevented." An implication of Posner's conjecture about delays to suicide is that there would be an increase in the average age of suicide.
Posner's conjectures have come to renewed prominence in the context of debates about the legalization of assisted dying on both sides of the Atlantic Ocean. In 2014 "evidence of premature death" resulting from the lack of access to PAS was presented before the Supreme Court of Canada. In February 2015 the court concluded that "the prohibition deprives some individuals of life." In the UK House of Lords in July 2014 it was argued that "many people… are dying earlier" because of the prohibition of PAS and the some "might have chosen to live" had PAS been legal. On the same basis, the Swiss organization EXIT claims that the "option of physician-assisted suicide is actually an effective form of suicide prevention."
Systematic empirical analysis of the Posner hypothesis is limited. Although Posner examined state data on suicides to illustrate his hypothesis, his data predated Oregon's legalization of PAS. To date there have been no formal tests on the impact of the state-level regulation of PAS on suicide rates. Furthermore, no research has examined the association between PAS and the age of suicides. In this article we aim to help fill these gaps in knowledge by exploiting the "natural experiments" that have occurred in various states legalizing or decriminalizing PAS at different times.
South Med J. 2015;108(10):599-604. © 2015 Lippincott Williams & Wilkins