Special Article
Physician-Assisted Suicide and Patients with Human Immunodeficiency Virus Disease
N Engl J Med 1997; 336:417-421February 6, 1997
- Abstract
Background
Data are limited on the attitudes and practices of physicians regarding assisting the suicide of patients with human immunodeficiency virus (HIV) disease.
Methods
Between November 1994 and January 1995, we used an anonymous, self-administered questionnaire to survey all 228 physicians in the Community Consortium, an association of providers of health care to patients infected with HIV in the San Francisco Bay area. The responses were compared with those in a 1990 survey of consortium physicians. Physician-assisted suicide was defined as “a physician providing a sufficient dose of narcotics to enable a patient to kill himself.” Respondents were to “assume that the patient is a mentally competent, severely ill individual facing imminent death.”
Results
One hundred eighteen of the questionnaires were evaluated. Respondents reported a mean of 7.9 “direct” and 13.7 “indirect” requests from patients for assistance. In responses based on a case vignette, 48 percent of the physicians said they would be likely or very likely to grant the request of a patient with the acquired immunodeficiency syndrome (AIDS) for assistance in a suicide, as compared with 28 percent of the respondents in 1990. Asked to estimate the number of times they had granted the request of a patient with AIDS for assistance in committing suicide, 53 percent said they had done so at least once (mean number of times, 4.2; median, 1.0; range, 0 to 100). In a multivariate analysis, factors positively associated with a physician's having, in fact, assisted a suicide were having had a higher number of patients with AIDS who had died; having received a higher number of indirect requests from patients for assistance; having a stated gay, lesbian, or bisexual orientation him- or herself; and having a higher “intention to assist” score (as calculated from the physician's responses to the case vignette).
Conclusions
Within a group of physicians caring for patients with HIV disease, the acceptance of assisted suicide increased between 1990 and 1995. A majority of respondents in 1995 said they had granted a request for assisted suicide from a patient with AIDS at least once.
Media in This Article
Figure 1
Distribution of the Numbers of Patients Assisted in Suicide, as Reported by 117 Respondents to the 1995 Survey.Table 1
Characteristics of Respondents to the 1990 and 1995 Surveys.Article Activity
- Article
We investigated factors influencing the attitudes and practices of physicians in the San Francisco Bay area regarding physician-assisted suicide for patients with the acquired immunodeficiency syndrome (AIDS). We evaluated whether, and to what extent, physicians are participating — or would be willing to participate — in physician-assisted suicide. Participation was defined as the prescription of a lethal dose of narcotics to a patient who was infected with the human immunodeficiency virus (HIV) or who had AIDS, who requested such a prescription in order to commit suicide. This definition was specified to distinguish physician-assisted suicide from active euthanasia (administering a lethal injection) and passive euthanasia (the withdrawal of life support in response to a patient's advance directive), both of which were outside the scope of our study.
We also assessed whether there have been changes over time in physicians' attitudes and practices regarding assisted suicide. In 1990, we surveyed a similar group of physicians in the same organization,1 but we did not ask whether or how frequently they had actually participated in assisted suicide. In the present study, we assessed how frequently physicians actually prescribe lethal doses of medications to their patients with AIDS.
Methods
We surveyed all 228 physician members of the Community Consortium, an association of providers of health care to patients infected with HIV in the San Francisco Bay area. The members include physicians in private practice and those who work at academic medical centers, community hospitals, neighborhood clinics, and health maintenance organizations. The number of members in the Community Consortium has increased since the 1990 survey (from 150 to 250), and there has been some turnover. Because the responses to both surveys were anonymous, it was not possible to identify people who responded to both.
On the basis of the questionnaire used in the 1990 survey, we developed a self-administered, anonymous questionnaire that included sections on demographic characteristics, professional and personal experience with AIDS, beliefs and attitudes regarding physician-assisted suicide, and participation in assisted suicide. Respondents were also asked how they would respond to a request from a patient with AIDS for assistance in committing suicide, as described in the following vignette:
Tom is a 30-year-old gay male computer programmer diagnosed with AIDS two years ago. He has severe wasting syndrome and painful oral ulcers, and responded poorly to treatment for his third episode of Pneumocystis carinii pneumonia. There is no evidence of neurological impairment, and it is clear that Tom is mentally competent. His mood is mildly depressed, but the depression is not pronounced given the seriousness of his condition. Tom has been in a primary relationship for eight years and worked until several months ago. As his personal physician since his diagnosis, you consider Tom a thoughtful, intelligent patient who does not appear to have any significant psychopathology. During Tom's biweekly clinic visit, he asks you to prescribe a lethal dose of narcotics for possible use at some future date.
Respondents were asked how likely they would be to prescribe a lethal dose of medication for Tom and, if Tom was adamant about obtaining assistance in committing suicide, what course of action they would take.
We defined physician-assisted suicide as “a physician providing a sufficient dose of narcotics to enable a patient to kill himself.” Physicians were told to assume that the patient was “a mentally competent, severely ill individual facing imminent death.” To determine the frequency with which respondents have participated in assisted suicide, we asked them to estimate the number of times they had granted the request of a patient with AIDS for assistance in committing suicide (e.g., by prescribing a lethal dose of medication). No time frame was specified. We did not obtain data on whether the medication was actually used by the patient to commit suicide. We also asked the physicians to estimate the number of times they had been asked “directly” or “indirectly” by a patient with AIDS to assist him or her in committing suicide. Neither “directly” nor “indirectly” was defined in the survey.
The questionnaire and study procedures were approved by the Committee on Human Research at the University of California, San Francisco. The study was conducted between November 1994 and January 1995. The survey packet included a cover letter, a copy of the questionnaire (including information on informed consent), and a postcard to be returned separately to request a synopsis of the study's findings. Providers who were no longer in clinical practice were asked to indicate this fact on the questionnaire and to return it unanswered.
Statistical Analysis
Fisher's exact test,2 the Mantel–Haenszel chi-square test,3 and the normal approximation to the nonparametric Wilcoxon two-sample test 4 were used to compare the demographic characteristics, attitudes toward assisted suicide, and responses to the case vignette of the 1990 and 1995 survey populations.
A scaled composite variable called “intention to assist” was constructed to aid in the analysis of attitudes toward assisted suicide. An individual score was derived from the responses to two questions about the case vignette. The possible scores ranged from 0 to 4. A response of “very likely” or “likely” to the question as to whether a respondent would prescribe a lethal dose after a patient's initial request contributed 2 points to an individual score; a response of “neither likely nor unlikely” contributed 1 point. To the question regarding an “adamant” request for a lethal dose, granting the request contributed 2 points; referral to the Hemlock Society contributed 1 point.
A forward, stepwise multiple logistic-regression analysis was conducted of the following variables: intention to assist (respondents with a score of 0 were compared with those with a score of 1 to 4), the physician's sexual orientation, the physician's number of patients with AIDS, the physician's number of acutely ill patients with AIDS, the number of the physician's patients with AIDS who had died, the number of direct and indirect requests from a physician's patients for assistance in committing suicide, the physician's estimate of the likelihood that he or she would consider or actually commit suicide if given a diagnosis of AIDS, and the amount of professional literature about assisted suicide the physician had read. The statistical significance of the coefficients assigned to individual variables was calculated with the Wald chi-square statistic with 1 degree of freedom.
Results
Of the 228 physicians surveyed, 137 (60 percent) responded. Of these 137 respondents, 19 indicated they were no longer in clinical practice, so 118 questionnaires were evaluated. In the 1990 survey, 69 of 150 physicians (46 percent) had responded.
As compared with the respondents in 1990, the respondents in 1995 were a more racially diverse group, more likely to be heterosexual, and more likely to have treated a relatively high number of patients with AIDS (Table 1Table 1
Characteristics of Respondents to the 1990 and 1995 Surveys.). Over three quarters of the respondents in 1995 had treated more than 80 patients with AIDS, as compared with 63 percent of the respondents in 1990.Respondents to both surveys reported similar numbers of direct requests for assistance in committing suicide from patients with AIDS (for the 1995 survey [n = 116]: mean, 7.9; median, 2; range, 0 to 200; for the 1990 survey [n = 69]: mean, 3.6; median, 1; range, 0 to 50). However, respondents to the 1995 survey reported more indirect requests (1995 [n = 115]: mean, 13.7; median, 8; range, 0 to 100; 1990 [n = 68]: mean, 6.8; median, 3; range, 0 to 100; P = 0.007 for the comparison of indirect requests in the two surveys, by the Wilcoxon test).
In regard to the case vignette, 48 percent of the respondents in 1995 said they would be likely or very likely to grant Tom's initial request for assistance, as compared with 28 percent of the respondents in 1990 (P = 0.005) (Table 2Table 2
Responses to the Case Vignette in 1990 and 1995.). The physicians were also asked how they would respond if Tom was adamant about getting assistance. Of the respondents in 1995 (n = 110), 51 percent said they would grant Tom's adamant request, as compared with 35 percent of the respondents in 1990 (n = 69) (P = 0.05). The respondents in 1995 were also less likely than those in 1990 to say they would talk Tom out of his request if he were adamant (P = 0.04). If they heard that another physician had offered assistance to Tom in committing suicide, the respondents in 1995 (n = 114) were also less likely to say they would talk to that physician to challenge or condemn his or her action than were the respondents in 1990 (2 percent vs. 14 percent, P<0.001).Asked to estimate the number of times they had “granted an AIDS patient's request for assistance in committing suicide,” 53 percent of the 117 respondents to the 1995 survey who answered this question reported that they had done so at least once (mean number of times, 4.2; median, 1.0; range, 0 to 100) (Figure 1Figure 1
Distribution of the Numbers of Patients Assisted in Suicide, as Reported by 117 Respondents to the 1995 Survey.). Asked how many times they had consulted with colleagues about helping patients with AIDS to commit suicide, 50 percent of the respondents in 1995 and 49 percent of those in 1990 reported that they had done so at least once (1995 [n = 115]: mean number of times, 2.7; median, 1; range 0 to 20; 1990 [n = 69]: mean, 3.8; median, 0; range, 0 to 50).In the multivariate analysis, four factors were found to be associated with whether a physician had ever assisted in a patient's suicide: a higher intention-toassist score, a higher number of the physician's patients with AIDS who had died, a higher number of indirect requests from patients for assistance, and a gay, lesbian, or bisexual orientation on the part of the physician (Table 3Table 3
Factors Associated in the Multivariate Analysis with Physicians' Having Assisted a Patient with AIDS to Commit Suicide.).Discussion
Previous surveys of physician-assisted suicide have reported that 7 to 9 percent of physicians have complied with requests from terminally ill patients for assistance in suicide.5,6 We found that about one half the physicians we surveyed in 1995 reported assisting at least once in the suicide of a patient with AIDS by prescribing a lethal dose of medication. This is a surprisingly large proportion, given the possible legal and ethical repercussions of such an action. It should be noted that California state law does not limit the amount of narcotics that a physician can prescribe at one time, and the quantity required for a lethal dose is not in excess of what might be obtained in a single prescription; a lethal dose of narcotics would thus not necessarily require multiple prescriptions or the hoarding of prescriptions by patients.
There is increasing support in the United States and elsewhere for so-called right-to-die legislation, euthanasia, and physician-assisted suicide.5,7-10 Recent rulings by the Second and the Ninth Circuit Courts of Appeals support the right of terminally ill patients to hasten their death with drugs prescribed by their physicians.11,12 Recent data indicate that, in some communities, rates of euthanasia and physician-assisted suicide among patients with AIDS may be higher than in the general population. In a study of 131 homosexual men with AIDS in Amsterdam, 29 (22 percent) died by euthanasia or physician-assisted suicide,13 as compared with 5 percent of patients with cancer and 2 percent of the general population.14 Several studies also suggest that persons with AIDS have a greater relative risk of committing suicide than the general population.15-17
In our surveys, the change from 28 percent in 1990 to 48 percent in 1995 in the proportion of respondents who reported willingness to assist in a patient's suicide suggests that acceptance by physicians of assisted suicide is increasing over time and that resistance to assisting in suicide appears to be lessening. The high level of reported assistance in 1995 may be due to the number of patients with AIDS who have died while under the care of physicians, as well as the increase in the number of requests for assistance received. Our respondents' estimates of the number of direct and indirect requests from patients for assistance may have been low in both surveys. In an effort to avoid a difficult issue, physicians may not hear their patients' subtle or disguised requests for assistance and, as a result, may be less attuned to their patients' needs. The physicians who reported a greater number of direct and indirect requests may represent a subgroup who are known in the community to prescribe lethal medications sought by patients with AIDS. The burdensome nature of this role is obvious. We did not study the potential for psychological repercussions among physicians who assist in suicides; however, this activity is likely to have psychological reverberations and to carry a social stigma.
A physician's being gay, lesbian, or bisexual was also positively associated with assisting in a patient's suicide; this may reflect identification with the patient. Identification with patients with AIDS may solidify the alliance between doctor and patient because such physicians may have a greater understanding of the patient's experiences and desires. One could speculate, on the other hand, that a gay, lesbian, or bisexual physician may have difficulty objectively assessing a patient's request for assistance. The physician may assume the patient's request is rational and fail to explore fully the psychological and motivational status of the patient. Ultimately, this could lead physicians who strongly identify with their patients with AIDS to prematurely prescribe potentially lethal medications in a situation in which the patient's request is an indicator of psychological distress rather than of a carefully considered desire to die. Although physicians who were not heterosexual appeared to identify more strongly with their patients and assisted patients in committing suicide to a greater extent than their heterosexual colleagues, sexual orientation was only one of the four factors that were associated with assisting a patient with AIDS to commit suicide.
Some physicians may view prescribing medications in response to a patient's request for assistance in committing suicide as a psychological intervention rather than a means of hastening death. The prescription of medications carries symbolic as well as pragmatic value in that it can increase patients' level of comfort by giving them a sense of control over a disease that tends to rob them of control. It is important to emphasize that although we assessed the frequency with which physicians prescribed a lethal dose of medication, we did not ascertain whether the medication was actually used by the patient to commit suicide.
A major methodologic problem in our study is selection bias, which limits the generalizability of our findings. Physicians who had an interest in or experience with assisted suicide may have been more likely than others to respond to the survey. Although selection bias may have influenced the magnitude of the reported findings, the associations we found are still likely to be genuine. An additional limitation is that the study was conducted in the San Francisco Bay area, where attitudes are generally more liberal and where there is a higher incidence of HIV infection than in other regions of the country. Furthermore, the terms “direct” and “indirect” were not defined in the questionnaire, nor was a time frame specified for the estimate of the number of times requests for assistance had been granted. These limitations notwithstanding, we found substantial evidence that some physicians are making it possible for their patients with HIV infection to commit suicide and that acceptance of this practice in a consortium of physicians who care for many such patients appears to be increasing.
Supported in part by a grant from the AIDS Clinical Research Center of the University of California, San Francisco.
We are indebted to Nicolé Mandel for assistance in designing and producing the questionnaire, David Townley for assistance in data management, and David Anderson for preparing the tables; we are also indebted to the many physicians who took the time to complete and return the questionnaires.
Source Information
From private practice in San Francisco (L.R.S., J.M.B.); the Community Consortium, University of California, San Francisco, AIDS Program at San Francisco General Hospital (T.F.M., D.I.A.); and the Department of Epidemiology and Biostatistics, University of California, San Francisco (E.C.).
Address reprint requests to Dr. Slome at 3180 18th St., Suite 201, San Francisco, CA 94110.
- References
References
1
Slome L ,Moulton J ,Huffine C ,Gorter R ,Abrams D . Physicians' attitudes toward assisted suicide in AIDS. J Acquir Immune Defic Syndr 1992;5:712-718
Web of Science | Medline2
Mehta CR ,Patel NR . A network algorithm for performing Fisher's exact test in r × c contingency tables. J Am Stat Assoc 1983;78:427-434
CrossRef | Web of Science3
Mantel N ,Haenszel W . Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719-748
Web of Science | Medline4
Conover WJ. Practical nonparametric statistics. 2nd ed. New York: John Wiley, 1980.
5
Lee MA ,Nelson HD ,Tilden VP ,Ganzini L ,Schmidt TA ,Tolle SW . Legalizing assisted suicide -- views of physicians in Oregon. N Engl J Med 1996;334:310-315
Full Text | Web of Science | Medline6
Fried TR ,Stein MD ,O'Sullivan PS ,Brock DW ,Novack DH . Limits of patient autonomy: physician attitudes and practices regarding life-sustaining treatment and euthanasia. Arch Intern Med 1993;153:722-728
CrossRef | Web of Science | Medline7
Russell S. Poll respondents OK euthanasia for terminally ill. San Francisco Chronicle. March 29, 1995.
8
Bachman JG ,Alcser KH ,Doukas DJ ,Lichtenstein RL ,Corning AD ,Brody H . Attitudes of Michigan physicians and the public toward legalizing physician-assisted suicide and voluntary euthanasia. N Engl J Med 1996;334:303-309
Full Text | Web of Science | Medline9
Solomon A. A death of one's own. New Yorker. May 22, 1995:54-69.
10
Nelson B. Mother's last request. Harper's. March 1996:35-43.
11
Holding R. Appeals court declares “right to die.” San Francisco Chronicle. March 7, 1996:A1.
12
Bruni F. Court overturns ban in New York on aided suicides. New York Times. April 3, 1996:A1.
13
Bindels PJE ,Krol A ,van Ameijden E , et al. Euthanasia and physician-assisted suicide in homosexual men with AIDS. Lancet 1996;347:499-504
CrossRef | Web of Science | Medline14
Van Der Maas PJ ,Van Delden JJM ,Pijnenborg L ,Looman CWN . Euthanasia and other medical decisions concerning the end of life. Lancet 1991;338:669-674
CrossRef | Web of Science | Medline15
Goldblum P, Moulton J. HIV disease and suicide. In: Dilley J, Pies C, Helquist M, eds. Face to face: a guide to AIDS counseling. San Francisco: AIDS Health Project, University of California, 1989:152-64.
16
Kizer KW ,Green M ,Perkins CI ,Doebbert G ,Hughes MJ . AIDS and suicide in California. JAMA 1988;260:1881-1881
CrossRef | Web of Science | Medline17
Marzuk P ,Tierney H ,Tardiff K , et al. Increased risk of suicide in persons with AIDS. JAMA 1988;259:1333-1337
CrossRef | Web of Science | Medline
- Citing Articles (23)
Citing Articles
1
Clive Seale. (2009) Hastening death in end-of-life care: A survey of doctors. Social Science & Medicine 69:11, 1659-1666
CrossRef2
Boudewijn E. Chabot, Arnold Goedhart. (2009) A survey of self-directed dying attended by proxies in the Dutch population. Social Science & Medicine 68:10, 1745-1751
CrossRef3
George E Dickinson, David Clark, Michelle Winslow, Rachael Marples. (2005) US physicians' attitudes concerning euthanasia and physician-assisted death: A systematic literature review. Mortality 10:1, 43-52
CrossRef4
Alexandra E. Shields, David Blumenthal, Kevin B. Weiss, Catherine B. Comstock, Douglas Currivan, Caryn Lerman. (2005) Barriers to Translating Emerging Genetic Research on Smoking into Clinical Practice. Perspectives of Primary Care Physicians. Journal of General Internal Medicine 20:2, 131-138
CrossRef5
M.K. Ben Diane, A. Galinier, H. Pegliasco, J.M. Lapiana, Roger Favre, Patrick Peretti-Watel, Yolande Obadia, Yolande Obadia. (2005) Deux points de vue des médecins sur les soins palliatifs. Santé Publique 17:1, 75
CrossRef6
James L. Werth JR, Howard Wineberg. (2004) A CRITICAL ANALYSIS OF CRITICISMS OF THE OREGON DEATH WITH DIGNITY ACT. Death Studies 29:1, 1-27
CrossRef7
L. Gary Hart, Thomas E. Norris, Denise M. Lishner. (2003) Attitudes of Family Physicians in Washington State Toward Physician-Assisted Suicide. The Journal of Rural Health 19:4, 461-469
CrossRef8
Gary L. Stein, Karen A. Bonuck. (2001) Attitudes on End-of-Life Care and Advance Care Planning in the Lesbian and Gay Community. Journal of Palliative Medicine 4:2, 173-190
CrossRef9
R Andraghetti, S Foran, R Colebunders, D Tomlinson, P Vyras, Cj Borleffs, Y Fleerackers, W Schrooten, M Borchert. (2001) Euthanasia: from the perspective of HIV infected persons in Europe. HIV Medicine 2:1, 3-10
CrossRef10
L. M. Cohen. (2000) Psychiatric Evaluation of Death-Hastening Requests: Lessons From Dialysis Discontinuation. Psychosomatics 41:3, 195-203
CrossRef11
Stephen Jamison. (2000) Factors to consider before participating in a hastened death: Issues for medical professionals.. Psychology, Public Policy, and Law 6:2, 416-433
CrossRef12
Barry Rosenfeld. (2000) Assisted suicide, depression, and the right to die.. Psychology, Public Policy, and Law 6:2, 467-488
CrossRef13
Barry Rosenfeld. (2000) Methodological issues in assisted suicide and euthanasia research.. Psychology, Public Policy, and Law 6:2, 559-574
CrossRef14
Judith C Ahronheim, Samuel B Davol. (1999) Pursuit of Assisted Dying. Journal of Pain and Symptom Management 18:6, 401-405
CrossRef15
K. V. Heath, E. Wood, G. Bally, P. G. A. Cornelisse, R. S. Hogg. (1999) Experience in treating persons with HIV/AIDS and the legalization of assisted suicide: The views of Canadian physicians. AIDS Care 11:5, 501-510
CrossRef16
Ellen L. Csikai. (1999) Euthanasia and Assisted Suicide. Journal of Gerontological Social Work 31:3-4, 49-63
CrossRef17
Colm Leonard, Vibha Mohindra, Stephen Ruoss, Ramona L. Doyle, Thomas A. Raffin. (1999) European Life-Support Questionnaire. Critical Care Medicine 27:8, 1686-1687
CrossRef18
Luigi Grassi, Katia Magnani, Mauro Ercolani. (1999) Attitudes Toward Euthanasia and Physician-Assisted Suicide Among Italian Primary Care Physicians. Journal of Pain and Symptom Management 17:3, 188-196
CrossRef19
Chin, Arthur E., Hedberg, Katrina, Higginson, Grant K., Fleming, David W., . (1999) Legalized Physician-Assisted Suicide in Oregon — The First Year's Experience. New England Journal of Medicine 340:7, 577-583
Full Text20
Meier, Diane E., Emmons, Carol-Ann, Wallenstein, Sylvan, Quill, Timothy, Morrison, R. Sean, Cassel, Christine K., . (1998) A National Survey of Physician-Assisted Suicide and Euthanasia in the United States. New England Journal of Medicine 338:17, 1193-1201
Full Text21
Fabrizio Starace, Lorraine Sherr. (1998) Suicidal behaviours, euthanasia and AIDS. AIDS 12:4, 339-347
CrossRef22
Kathy Faber-Langendoen. (1998) Death by request. Cancer 82:1, 35-41
CrossRef23
McKeogh, Maeve, . (1997) Physician-Assisted Suicide and Patients with HIV Disease. New England Journal of Medicine 337:1, 56-57
Full Text
- Letters
