Meisel A, Jernigan JC, Youngner SJ. Prosecutors and End-of-Life Decision Making. Arch Intern Med. 1999;159(10):1089-1095. doi:10.1001/archinte.159.10.1089
Copyright 1999 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1999
To examine personal beliefs and professional behavior of state criminal prosecutors toward end-of-life decisions.
District attorney offices nationwide.
All prosecuting attorneys who are members of the National District Attorneys Association. A total of 2844 surveys were mailed with 2 follow-up mailings at 6-week intervals; 761 surveys were returned for a response rate of 26.8%. The majority of respondents were white men, Protestant, and served in rural areas.
Main Outcome Measures
On the basis of 4 case scenarios, (1) professional behavior as determined by respondents' willingness to prosecute and what criminal charges they would seek; and (2) personal beliefs as determined by whether prosecutors believed the physicians' actions were morally wrong and whether they would want the same action taken if they were in the patient's condition.
Most respondents would not seek prosecution in 3 of the 4 cases. In the fourth case, involving physician-assisted suicide, only about one third of the respondents said that they definitely would prosecute. Those who would prosecute would most often seek a charge of criminal homicide. A majority of respondents believed that the physicians' actions were morally correct in each of the 4 cases and would want the same action taken if they were in the patient's position. There was a strong correlation between personal beliefs and professional behaviors.
A large majority of responding prosecutors were unwilling to prosecute physicians in cases that clearly fall within currently accepted legal and professional boundaries. In the case of physician-assisted suicide, results reflected a surprisingly large professional unwillingness to prosecute and an even greater personal acceptance of physician-assisted suicide.
IN THE 20 years since the New Jersey Supreme Court's Karen Ann Quinlan decision in 1976, our society has forged an operative consensus about the legality and morality of forgoing life-sustaining medical treatment. This consensus has rejected an absolute vitalism—that life must be preserved as long as possible no matter what the circumstances.1,2 Nonetheless, some treatment limitations, such as forgoing artificial provision of fluids and nutrition or turning off mechanical ventilators remain problematic for many persons from a psychosocial and moral, if not legal, perspective.3 A new and increasingly visible debate involves whether our society should take the next step: allowing physicians to aid their patients in actively hastening death.
The criminal law is often perceived as a barrier to making end-of-life decisions.4 Whether criminal liability is a realistic concern depends heavily on the discretion of participants in the criminal justice system, and perhaps none more so than criminal prosecutors. This study examines the professional and personal attitudes, beliefs, and self-predicted behavior of criminal prosecutors toward a spectrum of decisions at the end of life in an effort to discern how they might exercise their discretion to prosecute or not. We chose to study prosecutors not because they have any particular expertise with end-of-life decisions but because they are a professional group in a social position to affect end-of-life care through their interpretation and enforcement of the criminal law.
A few well-accepted concepts form the bedrock of the moral and legal consensus about treatment limitation at the end of life. This consensus is generally supported by practicing health professionals, professional organizations, courts and legislatures, a Presidential Commission, academic medical ethicists, and the general public. The central point is that it is permissible for competent patients to refuse life-sustaining treatment. Second, competent individuals are entitled to anticipatorily reject life-sustaining treatment through an advance directive in the event of future loss of decision-making capacity. Finally, although somewhat less well accepted, there is still general adherence to the view that a patient's family has the moral and legal authority to forgo life-sustaining treatment for an incompetent patient without an advance directive, based on what they believe the patient would have wanted (the so-called "substituted judgment standard").5
Yet, specific aspects of end-of-life decision making have achieved a less widespread acceptance. For example, refusal of life-sustaining treatment has been easier for many to accept when the patient is terminally ill and less so when the patient suffers from a chronic debilitating and incapacitating disease or injury such as cerebral palsy6 or paralysis.7- 9 For some health care professionals, withholding of life-sustaining treatment is more difficult than withdrawing it.10 For others, specific life-sustaining interventions have powerful symbolic meaning that makes them difficult to withhold and especially to withdraw, such as the provision of nutrition and hydration through feeding tubes.11
The administration of analgesics and sedatives is another controversial area because the doses necessary to relieve pain and anxiety are sometimes large enough to cause the suppression of respiration and then death. To encourage physicians to prescribe adequate analgesics for terminally ill patients, states have begun to pass laws to permit physicians to do so without fear of legal liability if the patient accidentally dies from the medication.12(Table 9-1a) The US Supreme Court, in its two 1997 decisions on physician-assisted suicide, condoned this practice.13,14
As the consensus about forgoing life-sustaining treatment slowly developed, one point remained almost sacrosanct, certainly among judges, legislators, and most medical groups: that although withholding or withdrawing life-sustaining medical treatment ("passively" hastening death) were morally and legally acceptable, "actively" hastening death is morally and legally impermissible. The terminology often used as synonyms for these 2 different forms of hastening death—respectively, "letting die" and "killing"—says it all.
In less than a decade, the bright line distinction between actively and passively hastening death has begun to erode, and a new debate has emerged about whether and how to breach the barrier between the two. Whether a new consensus will form and how long it will take remains unclear, but there are many indications of movement in that direction. Despite the US Supreme Court rulings that state laws criminalizing assisted suicide are constitutional, there are many reasons to believe that the movement for legalization of actively hastening death will continue and achieve at least limited recognition—not just in the United States, but worldwide—by legislation (as in Oregon15 and the Northern Territory of Australia16), by judicial decision (as in Colombia17 and as was unsuccessfully tried in Florida18), or by a more informal legal mechanism (as in the Netherlands19). Regardless of legal developments, physician-assisted suicide is likely to remain an issue very much within the realm of public debate.
A mail survey was sent to all prosecutors (N=2844) who were members of the National District Attorneys Association (NDAA). The NDAA is the only national professional organization for prosecutors in the country. Surveys were sent to prosecutors in every state, except Delaware, Rhode Island, and the District of Columbia in which there were no prosecutors who were members of the NDAA. The project was approved by the University of Pittsburgh (Pa) institutional review board.
Surveys were initially mailed in October 1994, with 2 follow-up mailings at 6-week intervals. All respondents were guaranteed anonymity. Surveys contained respondent identification numbers for tracking purposes only.
The survey consisted of 4 scenarios constructed to study various clinical practices associated with end-of-life decision making. Cases 1, 2, and 3 raise issues that fall within the consensus about end-of-life decision making though there is some lingering controversy about the behavior contemplated in these cases. By contrast, in case 4 the patient's request to a physician to provide a lethal overdose was clearly outside the traditional consensus. However, it is a practice about which there is increasing discussion and possibly support from both the public and physicians.
Turning off respirator for a now incompetent patient based on family reporting of his wishes. Mr B is a 55-year-old-man who suffers from emphysema, chronic respiratory failure, and chronic heart failure secondary to arteriosclerosis (hardening of the arteries). At home he was on oxygen constantly and was very limited in his physical activity. He was admitted to the hospital when his breathing deteriorated even further. A chest x-ray demonstrated a tumor, which turned out to be inoperable and highly malignant. Because of his severe respiratory failure, a tube was inserted through his mouth into his trachea (windpipe) and he was attached to a respirator (breathing machine). Repeated efforts to remove him from the machine and have him breathe on his own were unsuccessful.
After several weeks on the respirator, it became clear that Mr B could never come off it. Although he was conscious, he was no longer mentally lucid. On many occasions he tried to remove the tubes because they were causing him great discomfort, and his hands had to be tied to his bed rails with soft restraints.
His wife and 2 married daughters told his doctor that prior to being hospitalized, Mr B had told each of them on a number of occasions that he was afraid that, because of his emphysema, this kind of thing would happen. He made them swear that if it did, they would not force him to undergo this kind of, what he called, "torture."
If the respirator is turned off, Mr B will experience no distress and can be expected to die in no more than 20 minutes. His family and doctor, though not wishing to see him die, have accepted the inevitability of his death and would like to comply with his wishes. However, the hospital administration is concerned about possible criminal liability if the respirator is turned off.
Administering morphine at the request of a terminally ill, competent cancer patient to alleviate pain. The patient is a 55-year-old married woman hospitalized with pancreatic cancer that has spread throughout her body. She has only days or possibly a month left to live. The patient is awake and mentally competent but is suffering great physical pain from her illness. Her doctor informs her that while he can give her more pain medication (in this case, intravenous morphine), the dose necessary to stop the pain will likely suppress her breathing, thereby bringing her life to an end. The patient and her devoted husband and children say they prefer that the pain be relieved. However, the hospital administration is concerned about possible criminal liability if the morphine is administered.
Terminating tube-feeding for a patient in a persistent vegetative state based on oral statements of his wishes. Following a car accident in which he sustained a severe head injury, a 55-year-old married man remains unconscious after 8 months. His doctor and 2 neurologist consultants agree that the patient is in a persistent vegetative state, ie, while he can breath on his own and his vital signs are stable, he has lost all functions of consciousness and cognition. He has no awareness of himself or his environment. He is unable to experience pain or pleasure. The doctors agree that the condition is irreversible—there is no chance he will ever regain consciousness. However, he could remain alive in this state for another 10 to 20 years. He is being fed through a tube that has been inserted into his stomach.
The patient's family, who have been entirely devoted to his care, informed the patient's doctor and the nursing home that they wish the tube-feedings to be discontinued, allowing the patient to die. If the tube-feedings were discontinued, he would die in about a week and, because of his persistent vegetative state, experience no discomfort or pain. While he had no written living will or durable power of attorney, his family notes that the patient said on many occasions that he would not want to be kept alive in an unconscious state, and had specifically mentioned that he would not want to be maintained by "tube-feeding." His long-time doctor confirms that the patient had said the same thing to him before the accident. However, the hospital administration is concerned about possible criminal liability if the tube-feeding is discontinued.
Providing morphine tablets to a terminally ill brain tumor patient for her to take to end her life. The patient is a fully competent 55-year-old married woman without children. She is dying from an untreatable brain tumor. She has been to specialists in several nationally known medical clinics, all of whom agree that there can be no benefit from surgery, drugs, or radiation therapy. Since learning of her diagnosis 6 months ago, she has been suffering from increasingly severe symptoms including headaches, nausea, and now excruciating pain which persists despite large doses of morphine. There is no need for her to be hospitalized. She spends much of her time confined to bed at home and when she is not moaning and crying out in pain, she is spending increasing amounts of time sleeping.
Beginning 3 weeks ago, she began telling her husband and doctor that she wants to die and asking her doctor to give her extra morphine tablets which she will save and then take all at once to end her life. She is being treated by her family physician who has been her doctor for almost 20 years and who knows her and her husband well. Her family and doctor, though not wishing to see her die, have also accepted the inevitability of her death and would like to comply with her wishes in order to spare her additional pain and suffering that will inevitably occur before her death. However, her doctor is concerned about possible criminal liability if she dies from an excess of morphine.
We intentionally held certain factors constant: age, either irreversible loss of consciousness or a terminal illness, a previously expressed wish not to be kept alive in his or her current state, and a devoted family who accepts the inevitability of the patient's situation and wants to comply with the patient's expressed wishes.
We attempted to assess the respondents' professional reactions to the cases by asking whether they would take measures leading to prosecution if the specific end-of-life decision were actually made, the patient died, and the case came to their attention through a formal referral from the police, medical examiner, or some other governmental entity. Responses were open ended to questions concerning what criminal charges prosecutors would seek, if any. This allowed respondents to provide more complete answers and ensured that more relevant information was obtained. Open-ended questions were coded for analysis.
Prosecutors were also asked to answer 2 questions about their personal beliefs and opinions for each case, namely, whether they believed the physician's actions were morally wrong and whether they would want the same action taken if they were in the patient's situation. Respondents were asked whether they would favor or oppose legislation allowing physicians to assist terminally ill patients in committing suicide as well as legislation allowing physicians to actively end the lives of terminally ill patients and whether they think these actions are morally justified in some circumstances regardless of legality. Finally, prosecutors were asked whether life should be maintained for as long as possible, by whatever means possible, in all circumstances.
Of the 2844 surveys mailed, 761 were returned, for a response rate of 26.8%. Selected demographic characteristics of the respondents are shown in Table 1. At least 1 survey was returned from each of 48 states. Most of the survey responders were white men with a mean±SD age of 43.7 ±7.8 years. A majority (61.0%) were Protestant, and respondents were fairly evenly divided concerning the influence of faith on their beliefs regarding end-of-life decision making. Most respondents (72.2%) served rural areas, 16.5% served urban areas, and 11.3% served suburbs.
Only 12.1% of respondents had ever been formally involved in an end-of-life case, but 30.4% stated that they had been contacted about such issues by a physician, hospital or nursing home administrator, or lawyer for a hospital or nursing home. Of those, most (63.6%) had been contacted 2 to 5 times, and 23.4% had been contacted once.
Table 2 shows that in the less controversial cases (cases 1, 2, and 3), no more than one fifth of respondents would take measures that might lead to prosecution, such as seeking an indictment. However, approximately another one fifth were undecided about what they would do. In the potential assisted-suicide case, almost twice as many respondents would take formal action if death ensued. Table 3 reports that respondents who would take formal measures would most often seek a charge of some form of criminal homicide against the physician in each case. There was no significant difference between respondents' willingness to prosecute in each of the 4 cases based on respondents' age, sex, race, or religion.
Table 4 examines respondents' personal beliefs, as distinguished from their professional views. Most respondents, if they were in the position of the patients in the cases, would want treatment terminated, or as in case 4, would want their request for pills to commit suicide to be honored. An overwhelming majority of respondents (ranging from 86%-94%) believed that these actions were morally correct in cases 1, 2, and 3, and two thirds held this belief in case 4.
Table 5 shows the relationship between respondents' beliefs about the morality of the physician's actions and willingness to prosecute. In all cases, a majority of respondents who thought the physician's action was morally wrong would prosecute whereas a majority of those who did not think the action was morally wrong would not prosecute. Those unsure of the morality of the physician's action were also unsure about whether they would prosecute if each of the cases came to their attention through a formal mechanism.
Table 6 shows a similarly high correlation between willingness to prosecute and prosecutors' wishes if they were in the hypothetical patient's situation. In all cases, a majority of prosecutors who would want the same action taken would not prosecute while a majority of those who would not want the same thing done if they were in the patient's condition would prosecute. Results for Tables 5 and 6 were significant at the .000 level, indicating a strong correlation between personal beliefs and professional behavior.
We also asked prosecutors general questions to elicit their opinions about end-of-life decision making. Table 7 shows that more than 90% of respondents are not "vitalists"; that is, they do not believe that life should be maintained as long as possible regardless of the circumstances. Three fifths of respondents believed that physician-assisted suicide may be morally justified in some circumstances, and more than two fifths believed the same about lethal injection. To the question "Would you favor or oppose legislation that would allow physicians to assist terminally ill patients in committing suicide in some circumstances?" 349 (45.9%) were in favor, 228 (30.0%) were opposed, and 184 (24.2%) were not sure. In response to the question "Would you favor or oppose legislation that would allow physicians to actively end the lives of terminally ill patients in some circumstances?" 254 (33.4%) were in favor, 281 (36.9%) were opposed, and 226 (29.7%) were not sure.
The cases we posed to prosecutors track the evolution of the consensus about end-of-life decision making. None of the cases are clear-cut. Each focuses on aspects of end-of-life decision making that have engendered some degree of controversy. However, the practices in cases 1, 2, and 3 have all achieved a relatively high degree of acceptance and inclusion in the consensus; only case 4, involving physician-assisted suicide, has been and remains clearly outside the consensus.
The responses to our questionnaire present classic half-empty-glass, half-full-glass situations. For example, should we be heartened by the finding that between 80% and 90% of responding prosecutors would not find it problematic that physicians honored a very specific oral advance directive in cases 1 and 3? Or, should we be very concerned that between 10% and 20% of responding prosecutors said they would prosecute in a case involving well-accepted medical practice and that another 20% or so are undecided? In actual practice, when prosecutors closely examine a case, they are very unlikely to prosecute, as evidenced by the fact that there is only one prosecution of physicians for forgoing life-sustaining treatment reported in official legal case reports,20 and at that time (1981-1983) the particular practice—termination of tube-feeding from a patient in a persistent vegetative state—was not nearly as well accepted in law or medical practice as today. This is consistent with our finding that only 12% of respondents had ever been involved in any way in an end-of-life case.
The responses to case 4 demonstrate a considerable willingness by respondent prosecutors to tolerate physician-assisted suicide. According to the formal law, it is a crime in most, if not all, states for anyone—physician or other—to aid another person in actively taking his or her life, regardless of the means or motive. So it is not surprising that significantly more respondents (about twice as many as would prosecute in the next most frequent case) said that they would prosecute a physician for providing the patient in case 4 with extra morphine tablets knowing she would save them and use them to end her life.
On the other hand, it is very surprising that three fifths of respondents would not prosecute conduct that almost certainly was, formally speaking, a crime. This finding is probably explained in large part by the respondents' personal views about whether they believe physician-assisted suicide is ever morally justified, about whether it should be legalized, and about what they would want done for them if they were in the situation of the patient in case 4 (see Tables 4, 7, and 8). In fact, in all 4 cases, respondents' personal beliefs were more liberal than their perceived duty under the law.
It is difficult to believe that 40% of respondents really would prosecute in case 4. Although there are no accurate statistics on the number of prosecutions for assisted suicide and mercy killing, there are only a handful or 2 of newspaper accounts in the United States reporting prosecutions of physicians and laypeople each year for these practices. This fact, taken together with the fact that physician-assisted suicide is a phenomenon that has long existed, makes clear that actual true rate is probably infinitesimal, and certainly lower than 40%. There are other possible explanations for a low prosecution rate, most notably that because of the secretive nature of this behavior, it rarely comes to the attention of prosecutors and thus could not be prosecuted even if the prosecutor would be inclined to do so. Equally plausible explanations are that prosecutors are far more concerned with more traditional forms of crime, that physician-assisted suicide, at least if truly voluntary on the part of the patient, might be viewed as a "victimless crime," and the possibility, as previously mentioned, that once a prosecutor becomes more intimately acquainted with the details of a particular case, his or her theoretical inclination to prosecute melts in the face of the facts.
It is possible that respondents' predicted responses to our 4 cases result from their inadequate knowledge of the law. It was impossible to study that knowledge because of the numerous variations in state law and because the law is vague and uncertain in at least half the states and subject to a broad range of interpretations in the other states. Instead, we used respondents' involvement in end-of-life cases as a proxy for their knowledge of the law. There is very little difference in the action that a respondent would take between those who have been involved in an end-of-life case and those who have not. Thus, if experience with end-of-life cases is a valid proxy for knowledge of the law, knowledge of the law is not a determinative factor in the decision to prosecute.
The only significant difference between the respondents with experience in end-of-life cases and those without occurred with physician-assisted suicide. In this instance, more than half of the prosecutors with experience were willing to prosecute while only 37% of those with no experience would. These findings suggest that experience does not necessarily result in better knowledge of the law.
The findings with the most significant long-run implications concern the personal beliefs of responding prosecutors. In all 4 cases, personal beliefs predicted whether prosecutors would prosecute. The influence of personal belief or preference has also been shown to influence psychiatrists' behavior toward patients requesting physician-assisted suicide. One study revealed that psychiatrists who would choose the option of physician-assisted suicide for themselves are more likely to support its availability for others.21 Public opinion polls demonstrate significant public support for legalization of the practice.22 Despite the failure of the US Supreme Court to invalidate laws criminalizing physician-assisted suicide, continuing public debate on this subject might give further impetus to prosecutors not to bring criminal charges against physicians accused of actively hastening the death of terminally ill patients. As time passes, more people, including prosecutors, will have firsthand experience among immediate family and close friends with contemporary medicalized death and with the grievous pain, suffering, and cost that it can impose on patients and their loved ones, experience that could signal growing tolerance of and increased support for physician-assisted suicide.
Thus, we believe that the forgoing considerations, along with the proportion of respondent prosecutors who do not find physician-assisted suicide morally objectionable and would want it for themselves under some circumstances, and the correlation between personal beliefs and willingness to prosecute, suggest that a likely avenue for the legitimation of physician-assisted suicide would be an informal route similar to the one taken in the Netherlands.
The movement toward physician-assisted suicide, like the 20-year evolution of the consensus about forgoing life-sustaining treatment, is unlikely to follow a straight line. Acceptance of the practice by individual physicians, lawmakers, and the public will occur at different rates. Some will never accept it at all or in particular kinds of cases, just the way some physicians will not participate in the forgoing of life-sustaining treatment in general or the forgoing of tube-feeding in particular. But unless something rather dramatic and unexpected occurs, the acceptance of physician-assisted suicide in some form now seems certain.
Our findings must be considered in the context of the low response rate (26.8%), which prevents us from being able to generalize study results to prosecutors throughout the country. However, the number of prosecutors who completed the survey was substantial (n=761) and we received completed surveys from prosecutors in 48 of 51 jurisdictions, which provided widespread geographic representation. Also, with the exception of these 3 jurisdictions, the distribution of our survey responses is roughly equal to the distribution of NDAA members in each state. Another limitation involves the representativeness of the respondents. Those prosecutors who answered our survey are likely to be more motivated and more interested in these issues. Consequently, respondents might have answered the questions differently from the way in which nonrespondents would have. Nonetheless, our findings provide some insight into how some prosecutors view a spectrum of end-of-life situations and thus provides a starting point for thinking about how to ensure that practices that are generally accepted in the medical profession are shielded from legal consequences in end-of-life situations.
Because our study asked prosecutors to respond to hypothetical situations, the responses to our survey do not necessarily represent the decisions prosecutors would make in real cases in which they might have more motivation and time for legal research, consultation with colleagues, and reflection. Finally, the absence of any previous studies with which to compare our results makes it impossible to discern a trend in the thinking of prosecutors and whether this trend corresponds with or diverges from the trend among physicians or the public generally.
Those who are inclined to believe that criminal prosecutors pose an omnipresent threat to the conscientious medical care of the dying will find support for their view in the finding that about one fifth of the respondents say that they would prosecute for withholding or withdrawing life-sustaining medical treatment, behavior that clearly falls within currently accepted legal and professional parameters. Many will be heartened by the finding that four fifths would not prosecute, and if historical precedent is a guide, those who said they would prosecute would likely be deterred upon further investigation and reflection. Nonetheless, the large number of respondents who were undecided about prosecution of cases that many in the medical profession believe to be within the standard of practice may pose a great concern to already worried physicians.
When it comes to physician-assisted suicide, a behavior that falls outside a clear legal and social consensus, the responses of prosecutors reflects a surprisingly large professional unwillingness to prosecute and an even greater personal acceptance of physician-assisted suicide, even among those entrusted with enforcing the letter of the law. This finding suggests that despite the Supreme Court decisions that there is no constitutional right to physician-assisted suicide, the issue will continue to be debated at the state level, where it might find growing acceptance among voters, courts, and legislatures.
Accepted for publication August 18, 1998.
This work was supported in part by a grant from the University of Pittsburgh Central Development Fund.
We are grateful to Robert Arnold, MD, Rebecca Dresser, JD, and Gary Fischer, MD, for comments on an earlier draft of this article, and to Kathleen Reilly Boehm and Jan Wenzel for their research assistance.
Reprints: Alan Meisel, JD, Dickie, McCamey & Chilcote Professor of Bioethics and Professor of Law, School of Law, University of Pittsburgh, 3900 Forbes Ave, Pittsburgh, PA 15260 (e-mail: firstname.lastname@example.org).