MATERIAL

An article (of the 752) or a chapter (in two books) had to satisfy three different criteria to be included in our review. First, the text should be written in English, French, German, or some Scandinavian language. Second, it should be published between August 1980 and August 2000. Third, the main theses put forward regarding conditions and/or consequences of medical futility should express the opinion of the authors. In this way we identified 43 articles of which 37 related (explicitly or implicitly) to the situation in North America,^{1–6, 9–15,17–40} four in Europe,^41–44 and two worldwide.^45,46

METHOD

We will use a method for conceptual analysis that was first presented by the Swedish philosopher Anders Wedberg.⁴⁷ In his analysis Wedberg used the concept of “property” as an intermediate between the conditions of property (called A) and the consequences of property (called B). He described the different conditions and consequences of property using the following scheme: “If condition A, then property” and “If property, then consequence B”.

In a similar way the word “futility”, as used in the clinical setting, involves two steps: the conditions of futility and the consequences of futility. If the focus is on the consequences, the use of the word “futility” implies that there are specifiable conditions that are not mentioned in the specific context. And if the focus is on the conditions, the use of the word “futility” implies that there are specifiable consequences not mentioned.

We will analyse the concept of “medical futility” by using empirical data on possible conditions and possible consequences based on the different opinions about futility expressed in the 43 articles.

RESULTS

First, we needed a relevant version of the two schemes: “If conditions A, then a particular measure is futile” and “If a particular measure is futile, then the consequences B”. Thus, our task was to identify categories useful to capture the conditions and the consequences described or indicated in the literature.

Different conditions were mentioned or implied in attempts to define or clarify the meaning of “medical futility”. Such conditions are: A1) who decides, A2) about what type of measure, A3) with what goal, and A4) what probability to succeed (box 1). Similarly, different moral consequences of a futile measure were also mentioned or implied. Such moral consequences are: B1) the health care professionals' possible actions, B2) communication about options, and B3) who is responsible for the decision to forgo the measure (box 2).

The second task was to identify the different opinions about conditions and consequences stated or indicated in the 43 articles (summarised in table 1 and table 2, respectively). Four articles included different statements regarding the conditions of futility within the same article (called 13a-b, 19a-b, 22a-b, 37a-d), and four gave more than one statement about moral consequences (called 11a-b, 13a-b, 22a-b, 37a-f).

Of the 43 articles, one focused on the conditions only²³ and one described only consequences of futility.³⁸ For 20 articles it was possible to identity all four conditions of futility and all three consequences of futility. In the remaining 21 articles at least one of the four conditions or the three consequences was not mentioned.

ANALYSIS OF CONDITIONS OF FUTILITY

Who decides about futility? Regarding this condition only four articles gave no suggestion at all, but in 35 articles the authors stated that they thought of physicians as decision makers, in three cases together with patients and in one case together with relatives. The health care team was mentioned in four of the articles. Two of the authors referred to guidelines and “publicly conducted evaluation”, respectively. In two single cases only the patient's or family's opinion was taken into account. In one article¹³ the authors stated more precisely that if no medical benefit was to be expected, the physician should decide, while the patient was to determine whether he or she would accept “a poor quality of life”. Another article¹⁹ mentioned “patient preferences” when the patient was to decide, but did not specify when the physician's opinion decided the issue. There were no systematic differences over the years.

What type of measure may be futile? All but one of the articles stated or at least implied that some kind of treatment was futile, not infrequently cardiopulmonary resuscitation. Two articles mentioned diagnostic measures and one article included nursing measures.

What is the goal of the measure? In 40 articles a goal was mentioned, in six of them physiological goals, and in 11 life-prolonging goals. Four of the articles mentioned patients' or surrogates' goals or preferences and in one the aim was to avoid expensive treatments. In 35 articles “quality of life” was emphasised.

What is the probability of success? Regarding these conditions, four articles stated a numerical probability of success (less than one per cent), and 29 articles described different semiquantitative probabilities. In ten articles there were no such suggestions at all. Four gave more or less qualitative descriptions (“already attempted”, “not validated”, “possible to balance”, and “scientific evidence”).

ANALYSIS OF CONSEQUENCES OF FUTILITY

What may, ought or should health care professionals do, assuming futility? Though the health care professionals have to make up their mind in such situations, 13 articles did not mention any options. One article stated it was not permissible at all to withhold or withdraw treatment, while three articles discussed the difference between “permissible, desired or required”. Again, there was no change over the years.

Involvement in the communication? In nine articles, seven of them published before 1994, the authors stated that no information to the patient or surrogate was needed when a specific futile measure was to be withheld or withdrawn. In four articles one-way information from the health care professionals was said to be enough, but in 27 articles the authors recommended a dialogue with the patient and family before decision making. One article discussed both a “dialogue” and “no information”¹³ and one article²¹ tried to make a distinction between “full information” and “dialogue” in relation to the goals (physiological versus patient goals).

Who is responsible for the decision to withhold or withdraw the measure? Regarding the final decision making, the authors of 19 articles recommended that the physician should decide unilaterally. Four proposed the possibility of second opinion or appeal if the patient or the surrogate disagreed. Advance directives were recommended by three, of which one also wrote that they were not always followed.³² Over time, more and more articles recommended joint, “group process” or compassionate decision making, by physicians and patients together. Differences of attitudes between American and British physicians were also noticed.⁴² In America they are more patient-oriented, implying that the patient should decide. British physicians seem to be more paternalistic. Some articles proposed joint decision making with possibilities of an appeal to an ethics committee.

DISCUSSION

The purpose of this article is to provide a practical approach to handling questions about medical futility. Instead of proposing another formal definition we have tried to make explicit the different conditions and consequences of futility. Our own opinion about this issue is summarised in box 3.

Our model makes it clear that there are two critical questions in the clinical setting, “Who has the final authority . . .” and “Who will be responsible . . .”, and that different persons may be designated for these tasks. That health care professionals are best suited to describe and assess the measure and probability of reaching a certain goal is hardly controversial, but this doesn't justify them making all decisions. Thus, we would especially like to emphasise the need to improve communication with the persons involved and affected. The key issue is how to help health care professionals talk with the patient and the family to resolve as many disagreements as possible and arrive at a consensus.

With this goal in mind we combined a philosophical method of conceptual analysis with empirical data from the literature on medical futility. Usually there was no difficulty in understanding the positions put forward by the authors. In some cases, however, our interpretations are more uncertain. To be more precise, we found it difficult to understand, for instance, the meaning of “clinical judgment”,⁶ “medical enterprise”,²³ “patient's goals”,³⁹ “scientific evidence”,⁴⁰ “obligation to support dignity”,²⁴ “5-step counselling”,²⁶ “joint bodies”,²² and “benefit”, especially this last, which was used by many of the authors).

A possible limitation of this study is therefore that the statements of some authors are rather vague and difficult to understand. The positions specified in the different columns of table 1 and table 2 may therefore not always be correct. The study is also a “retrospective” survey of opinions.

Some authors focused on the conditions, for example,²³ others on the consequences, for example,³⁸ but in the clinical situation the two are always related. If a certain measure is considered futile, this has practical consequences. Conversely, if a treatment is withheld or withdrawn, justification is required.

Sometimes different positions regarding the conditions or consequences were maintained in the same article.^{13,19,21,31,36,40,45} This may reflect the complexity of the situation arising when trying to decide who should decide: the physician or the patient. Most authors maintain that the physician's opinion ought to rule when a measure and its probability of success are to be assessed, however imprecise the prognosis may be. The same does not apply to the goals, since most goals are related to quality of life. In the past physicians usually made such judgments, but now there seems to be general agreement that the decision should also take into account the patient's opinion. We agree with Tomlinson and Brody⁴⁸ that the autonomy of all involved—that is, patient, physician, relative etc, must always be considered when the moral consequences of futility are discussed.

CONCLUSION

Based on this literature review we have tried to construct a model (box 3) which we hope will facilitate decision making for health care professionals, patients and relatives. Of course, the conceptual scheme needs confirmation by forthcoming prospective studies about futility assessment, communication and decision making in the clinical setting. Further, it may not help in all futility situations, but used systematically it may make it easier to gather the necessary information and to make the relevant values visible. What is most important of all perhaps is to separate the two questions: “What may be regarded as a futile measure?” and “What is justified in futility situations?” Thus, our model makes communication about futile measures possible without using words such as “futile”, “meaningless”, “useless” and the like. In our opinion there is no need to define the term “futility” in order to make an acceptable decision in an individual situation; yet this has been a major problem for many authors.