Since Yale Kamisar’s seminal paper from 1958, one of the most common arguments of opponents of the legalization of euthanasia has always been that permitting the ending of a patient’s life at his explicit request will necessarily lead to permitting it without request.Footnote ¹ The basic claim is a logical one: having permitted one practice, a society would be committed to permitting the other. Usually, this is explained as follows: if euthanasia is justifiable at all, its basic justification can only be that it ends a patient’s unbearable suffering by ending his life. But a patient can suffer to the same extent, whether or not he is able to formulate a voluntary and well-considered, and hence valid, request for his life being ended. The justification therefore applies to both cases. Now, if permitting one practice commits a society to permitting the other, we can predict the legal introduction of the first practice causing the spreading and eventually the legalization of the other. The logical connection will be reflected by a law-like causal relation.

This argument seemed to many to receive empirical confirmation long ago, in 1991, when the results of the first national survey of medical decisions at the end of life were published in the Netherlands.Footnote ² Termination of life without request occurred in 0.8 percent of all deaths: the so-called Remmelink 1000. It was pointed out, however, that the mere occurrence of such cases did not prove that the practice could be causally linked to the recognition of the legal permissibility of euthanasia by the Dutch Supreme Court a few years before (1984). (‘Euthanasia’ is understood in the Dutch context to be the ending of a person’s life at his explicit request.)Footnote ³

Data that became available after 1991 did not seem to corroborate the hypothesis of a causal connection.Footnote ⁴ For some other countries that had not legalized euthanasia under any conditions, comparable percentages were found (e.g., 0.67 percent in Denmark in 2000–2001; 0.6 percent in France in 2010);Footnote ⁵ the percentage of cases in the Netherland in 5 succeeding surveys declined to 0.3 percent in 2015;Footnote ⁶ and in the only case for which we have pre- and post-legalization data, Flanders (the Dutch-speaking part of Belgium), the percentage decreased from 3.2 percent in 1998 to 1.7 percent in 2013.Footnote ⁷

In a recent contribution to a book on Belgian euthanasia, David Albert Jones sets out to revitalize the argument, focusing on the Flemish case.Footnote ⁸ Actually, his more modest claim is only that in this particular case, the legalization of euthanasia has led to an increase in the number of cases of the ending of life without request. Because the claim of a causal link between the two practices is usually presented as a general one, I will, however, also consider the extent to which it is confirmed by recent Dutch data.Footnote ⁹

Using data from 2007, Jones points out that termination of life without request occurred in more than the 1.8 percent of deaths recorded for that year. In a substantial percentage (17 percent) of the cases (14.5 percent of all deaths) in which doctors reported to have deeply sedated the patient until his death, they indicated to have done so with the intention or co-intention to shorten the patient’s life, but only in a minority of sedation cases did they act on the request (10 percent) or at least with the consent (20 percent) of the patient.Footnote ¹⁰

According to the most recent Dutch survey, continuous deep sedation until death was decided upon in the Netherlands in 2015 in 18 percent of all death cases. In 2 percent of those cases (circa 500 cases), hastening of the end of life was asserted to be the purpose of the action. The questionnaire did not ask whether this was the ‘explicit purpose’ or only one of the purposes, in addition to the alleviation of suffering. No data about request or consent have been provided either.Footnote ¹¹ Nevertheless, according to Jones’ reasoning, chances are that by taking this information into account, we would find a much higher percentage of cases of terminating life without request than the presently reported 0.3 percent.

Because of the long-standing prominence of this particular argument in the controversy about the legalization of euthanasia, it is important to scrutinize Jones’ claim carefully. I will present four comments.

1. 1) Since 1990, questionnaires in research of this kind are sent to a large sample of physicians, and are always focused on one particular case of death. As in all other cases, the questionnaire used in Flanders in 2007 was designed to find answers to the following questions:

   1. a) What did the doctor do? (Administer drugs, withhold or withdraw treatment?)

   2. b) Did she explicitly intend or co-intend to shorten the patient’s life?

   3. c) Did she act on an explicit request by the patient?Footnote ¹²

   If a case was classified as a case of the ending of life with or without the patient’s request, this was done on the basis of the answers to this first set of questions. The same questionnaire also included a set of questions about continuous deep sedation until the patient’s death.Footnote ¹³ Note that both sets of questions concern the same death. If, replying to the second set of questions (about sedation), the doctor indicated to have sedated the patient with the explicit intention or the co-intention of hastening the patient’s death, we should therefore assume that this case has already been classified, on the basis of the answers to the first set of questions, as either a case of symptom-relief with a co-intention to hasten death, or as the ending of a life with or without request. It is true that some of the answers to the two sets of questions may be inconsistent with each other, but we have no idea whether such inconsistencies occurred, and if so, how frequently. In any case, we cannot add up the outcomes of the second set of questions to the outcomes of the first set. In doing this, Jones makes the elementary mistake of double-counting.

2. 2) Contrary to the practice of the researchers, and of most authors who have argued for the existence of this particular slippery slope, including John Keown and Neil Gorsuch, Jones does not only classify the cases in which the doctor reported to have had the ‘explicit intention’ of shortening life as cases of termination of life, with or without request. He also reckons the cases to that category in which the doctor declared to have had that intention in addition to the intention to alleviate suffering. But even if we distinguish between the ending of life and the intensification of symptom relief by reference to the intention of the doctor (see below, sub-section 4), this information is insufficient to classify at least the second class of cases as instances of killing. Authoritative statements of the Doctrine of Double Effect hold that, in order to know whether an effect is really intended and not merely foreseen as a side-effect, we should ask whether the agent would have changed his plan on learning that it would not have this particular effect.Footnote ¹⁴ (This has been called the ‘Failure Test’.) We do not know the answer to this question in the cases of reported ‘co-intentions.’The doctors involved may not even have known the answer to that question themselves. If you consider an action with two welcome effects, you usually need not ask yourself whether you would act in the same way if the action would have had only one of those effects, in order to decide to act in that way. Sometimes you can reconstruct the answer afterwards from your past actions and your general way of thinking. But that need not be the case.In these particular cases, there is, however, a special reason why many doctors would be able to sort out their intentions to some extent.For if their action could not be justified by a palliative aim, it would always be an improper medical action and, if a life-shortening effect would really be plausible, illegal.Footnote ¹⁵ It is therefore probable that many doctors would not act with the aim of hastening death if that action could not also be justified by the aim of alleviating suffering. This may even be true of some doctors who had the shortening of the patients’ life as their primary aim, and therefore reported to have acted with ‘the explicit intention to hasten death.’ If forms of palliative treatment are medically indicated, it is plausible that in the large majority of these cases, including many ‘explicit intention’ cases, doctors would also provide them if they did not believe them to have any life-shortening effects. To that extent, they would not satisfy the Failure test. However, an unknown percentage of them would probably use lower dosages of opiates or benzodiazepines, either within or beyond the normal range.Footnote ¹⁶An additional problem is that we do not know whether all or even most doctors are clearly aware of the distinction between wishes and intentions as that distinction has been elucidated by philosophers since Elisabeth Anscombe.Footnote ¹⁷ Perhaps they merely hoped that the patient’s life, and thereby his suffering, would end as a result of their intervention. Self-ascriptions of intention are notoriously unstable.Footnote ¹⁸ Jones is aware of this fact because he suggests that, because of it, the lower incidence of reported termination of life without request in the Netherlands may only be an effect of different descriptions by doctors of their intentions in similar cases. Maybe, but how does he know that the original description would have been the correct one?Footnote ¹⁹

3. 3) Benzodiazepines are the proper drugs to be used for the purpose of sedation, as all guidelines for ‘palliative sedation’ indicate. If only opiates have been used (31 percent of these Flemish ‘sedation’ cases in 2007), these cases should be classified as cases of symptom management rather than of deep sedation until death. However that may be, the current belief is that both morphine and proper sedatives have no life-shortening effects at all, if used in proper dosages or even somewhat larger ones.Footnote ²⁰ But as regards these sedation cases we have no direct information about the dosages used, neither about the total class nor about the subclass characterized by an explicit intention to hasten death. We know that doctors interpret the requirement of proportionality from the Dutch and Flemish Guidelines for Palliative Sedation in different ways, some scrupulously titrating to symptoms, others immediately starting with a higher dosage in order to make sure that from that moment on the patient will not suffer at all.Footnote ²¹ But neither practice by itself implies a foreseeable life-shortening effect.What we do know, however, is that if a doctor reports to have intended, or co-intended, to shorten the patient’s life, that does not mean that he has used larger dosages of drugs than were necessary for relieving the patient’s symptoms. For example, about the reported Flemish cases of the termination of life without request (1.8 percent of all deaths in 2007) we know that in 93.8 percent of those cases, opioids (with or without benzodiazepines) had been administered. In 22.7 percent of these cases, the dosages of opioids used were reported to be higher than necessary for symptom control. But in 43.9 percent of the cases (29 of 66) the dosages had not been higher, and only low-dose benzodiazepines or none at all had been used.Footnote ²² This means that at least these 43.9 percent of cases in the sample have been misclassified: whatever the intention of the doctor, it is highly improbable that the drugs he administered actually caused the death of the patient. And if an action does not cause a death, it cannot be a killing.As for the Netherlands, in 2010, only in 1 case of the 13 classified as cases of the ending of life without request (in a sample of 6861 deaths studied) had regular euthanatics (muscle relaxants) been used; in 42 percent of the cases, an overdosage of morphine had been administered.Footnote ²³ In 2015, only in 2 of 18 cases in the sample had regular euthanatics been used, and only in 3.2 percent of the cases had an overdosage of morphine been administered.Footnote ²⁴ We cannot conclude that in the Netherlands, only about 60 cases a year of the ending of life without request still occur, for a sample of 18 cases is too small to allow extrapolation of any figures to population level.Footnote ²⁵ The actual number may be higher than that, or lower. But the conclusion is warranted that the estimation of the number of cases of termination of life without request has been much too high since 2010, and quite possibly from 1991 on, even though we do not know exactly by how much it is too high.It would be extremely odd if doctors who reported to have deeply sedated a patient with the explicit intention or co-intention of hastening death, used large overdosages of midazolam or morphine more frequently, particularly if in those cases they did not also report, in reply to the second set of questions, that they had administered a drug to the patient with a life-shortening intention. In any case, we have no data confirming that hypothesis.

4. 4) Both Dutch and Belgian criminal law have a doctrine of the criminal mind that is incompatible with the Doctrine of Double Effect. The Dutch doctrine of ‘opzet’ (dolus, intent), for example, stipulates two conditions, one regarding belief, and one regarding volition.Footnote ²⁶ Regarding belief, the agent has to be aware in advance that the action he plans will or may have a harmful effect or effects. No special mental act of recognition of this prospect is required, only awareness in the dispositional sense. As regards volition, it is enough that, being aware of this significant probability, the agent decides to execute his plan. He doesn’t take the harmful effect as a reason to renounce the action, and in this sense willingly allows the effect to occur. He may intend it as his end or as a means to his end, or he may only take it into the bargain; that doesn’t matter. Nevertheless, in Dutch and Belgian law, a medical action that foreseeably hastens death, is legally permitted when hastening death is only “a subsidiary effect of a treatment that is necessary for the alleviation of suffering and adjusted to that end.”Footnote ²⁷ What justifies the action in that case is the availability of a proper medical justification. Whether that justification actually played any role in the deliberations of the doctor is legally irrelevant.This means that in cases in which the doctor acted with the explicit intention or co-intention of hastening death, but a medical indication was available for the use of midazolam or morphine in the dosage she actually used, her action would legally count as leading to a ‘natural death,’ even if, contrary to present beliefs, a life-shortening effect would be probable. Her action should be classified as the medically-indicated and proportional use of palliative medicine, irrespective of her intentions.Footnote ²⁸Of course, Jones and other opponents of the legalization of euthanasia under any conditions may also be critical of the understanding of ‘intent’ in Dutch and Belgian criminal law (and in many other civil law jurisdictions). As a matter of fact, such opponents have occasionally applauded the Dutch and Flemish researchers for using a classification scheme of end-of-life decisions that (seemingly) enabled them to appraise their results in terms of the doctrine of double effect.Footnote ²⁹ That doctrine is, however, very much disputed in ethics, in particular as regards cases in which both the intended and the merely foreseen effect apply to the same person. Why not aggregate those effects in such cases?Footnote ³⁰ Obviously, this is too large a topic to be discussed here.Footnote ³¹ Since 2002, however, the surveys are all meant to evaluate the existing euthanasia laws. It is therefore problematic that they use a classificatory scheme that is foreign to the law.