Introduction

Physicians involved in the care of elderly patients with a limited life expectancy are often confronted with the dilemma of whether treatment should be given or should be considered ineffective or undesirable. Decision making is particularly complicated with regard to incompetent patients, for example in the case of resuscitation and hospitalisation. This problem is paramount when considering the administration of antibiotics to demented nursing home patients with pneumonia.¹ Curative treatment, which is aimed at achieving cure of the pneumonia, does not necessarily result in an improvement in physical or mental health status. However, refraining from curative treatment and providing only palliative treatment, solely aimed at alleviating symptoms to improve the patient's wellbeing, could imply that death is hastened. However, in addition to the medical considerations mentioned above, ethical and legal issues are involved as well, which further complicates the decision making process.

In general, guidelines can be a support by systematically addressing the various aspects involved in the decision making process. Medical guidelines, especially when evidence-based, can be quite directive. However, when juridical and especially ethical factors are to be considered as well, the “good” decision is not always as clear cut. The physician has to balance the relative importance of many, possibly conflicting, considerations to come to a decision. Therefore, the requirements for this type of guideline are quite different from the requirements for the traditional medical guidelines.

This paper describes a method of constructing guidelines for decision making concerning demented nursing home patients. This method was used to develop a guideline for do-not-treat decisions in nursing home patients with pneumonia. Pneumonia was chosen because of its high incidence in nursing home patients, and the associated high morbidity and mortality rates.²⁻⁵ An outline of the “checklist of considerations” is presented in the annex, and a slightly adapted front page in table 1.

Methods

step 1: study of literature

The Medline database was used to search the medical literature from 1987 to 1994, for papers dealing with prognostic factors and other possible factors that can play a role in the decision making process about whether to treat pneumonia or other infections curatively. Only those studies were selected in which at least some of the participants were psychogeriatric nursing home patients. The bibliographies of all the papers identified were reviewed, as well as the indices (1987-1993) of relevant journals that were not included in the Medline database. The literature search was repeated twice during the whole process in order to identify more recent literature.

step 2: information from clinical practice

The aim of this step was: a) to examine the relative importance in clinical practice of the medical, juridical and ethical factors identified in the literature search, and b) to identify additional factors which could be considered important. For this purpose, a questionnaire was designed and sent to a jury of 26 nursing home physicians (NHPs) who had experience in nursing home medical practice and were associated with the university. The questionnaire was anonymous. Table 2 shows the topics that were addressed.

step 3: consensus meeting of medical professionals

A meeting of nine experienced NHPs was held to investigate the need for this specific guideline and to reach consensus about the factors (medical, ethical, juridical and also procedural aspects) to be included in such a guideline, based on the results of the literature search and the questionnaire, as well as their expert opinion. In addition, how certain they were about the decision being the right one was discussed.

step 4: design of a concept guideline

As information and consensus on the medical factors determining the usefulness of curative treatment for pneumonia were lacking, it was decided to develop a “checklist of considerations”, based mainly on ethical and juridical issues. Medical issues were addressed in general terms only.

step 5: consultation of (inter)national experts

To ensure widespread support for the “checklist of considerations”, five national and three international experts in the fields of nursing home medicine, ethics and law were consulted. They were sent the conceptual, that is to say, the initial “checklist of considerations”, and a questionnaire containing questions about their agreement or disagreement with each separate section of the list and the usefulness of the checklist as a whole. The experts were asked to state the reason for any disagreements and to suggest alternatives.

step 6: pilot study

Subsequently, a pilot study was performed to detect potential misinterpretation of the text and to evaluate the practical applicability of the checklist. Five nursing home physicians were asked to use the checklist either for two consecutive patients for whom they made a decision as to whether to treat pneumonia curatively, or else to complete it for two previous cases.

step 7: approval and authorisation

In order to obtain a high level of professional acceptance, the “checklist of considerations” was submitted to the international experts of step 5 for approval. Partial approval or approval with specific individual comments, was also an option. Finally, in order to obtain the status of a national guideline, the Dutch Society for Nursing Home Physicians (NVVA) was asked to authorise the “checklist of considerations”.

Results

The literature review identified reports and a discussion paper,⁶ as well as many opinionative articles potentially related to withholding curative treatment in psychogeriatric nursing home patients. Besides these, Dutch reports from leading societies on the acceptability of terminating life in incompetent patients were identified'^7–10 along with the national Dutch Law on the Medical Treatment Agreement (1995). Factors concerning health status and prognosis that could play a role in the decision making process were reported on in numerous studies of community-acquired pneumonia in adult hospitalised patients; these were reviewed in a meta-analysis in 1996 by Fine et al.¹¹ Prognostic information concerning treatment of patients with fever^12–14 or lower respiratory tract infections¹⁵ in demented (nursing home) patients was available, but not specifically for pneumonia patients. Specific guidelines for seriously demented patients appeared to be available neither in the Netherlands, nor elsewhere. However, the need to implement standards for decision making in the care of severely demented patients was confirmed by several authors.^16–18 Repetition of the literature search did not deliver any specific guideline on whether to treat pneumonia curatively in elderly patients. Recently, a report was published on the treatment of elderly pneumonia patients concerning, however, mainly hospitalised patients.¹⁹ Two recent studies on pneumonia mortality in nursing home patients treated with antibiotics did reveal some prognostic factors^{20, 21} which are directly relevant to the “checklist of considerations”.

The results of the questionnaire on clinical practice (step 2) are presented in table 3. Twenty NHPs (77%) responded. Almost all respondents (95%) rated the “overall” physical condition of the patient “important” or “very important” when making the treatment decision. The estimated life expectancy (75%), active or passive communication skills (60%), and the severity of dementia (60%) were considered either very important or important factors by most of the responding NHPs as well. When asked about important somatic factors that were not mentioned in the questionnaire, the NHPs gave a wide variety of answers. Most frequently mentioned (n = 6) was the patient's wellbeing, also described as “mood” or “will to live”, before the onset of pneumonia. Sixteen additional factors were mentioned, most of them related to the patient's prognosis or suffering. According to 55% of the NHPs, contact with the family of the patients (topic 3 in table 2) should always take place the moment pneumonia occurs. Different reasons were given for this: reconstructing the wish of the patient, asking for the family's opinion on the best treatment, or merely informing them about the treatment decision. According to the remaining 45% of the NHPs, contact with the family was not always necessary, mainly when the patient's wish was already known, or when the family was not involved with the patient. All NHPs, however, were of the opinion that contact with the nursing staff should always take place. For this purpose, the same reasons as with “contact with the family” were mentioned. Less significance, however, was given to the role of the nursing staff in reconstructing the patient's wish (15% for nursing staff, versus 35% for the family). As far as decisive factors are concerned, the highest scores for the degree of importance (median score 9 on a 10-point scale) were given to the items “the patient's health prior to the onset of pneumonia”, “medical sense/futility to treat (non-)curatively”, and “the previously expressed wish of the patient to be treated curatively”. The “expected physical health status of the patient in case of cure (prognosis)” ranked fourth. Items with median and mode scores less than nine included the wish of the family and of the nursing staff to treat the patient curatively, the severity of dementia and success achieved by curative treatment of pneumonia in the past. Least important, according to most of the NHPs (mode 0), were comorbidity and the burden on the patient due to the treatment, or the burden on the patient due to the pneumonia itself.

The participating experienced professionals consulted during the consensus meeting (step 3) reported that the results of the questionnaire were on the whole in accordance with their concept of clinical practice. The professionals agreed that if no explicit wish had been expressed, the patient's physical condition before the onset of pneumonia was the most important criterion for determining whether or not to treat the patient curatively. There was no consensus, however, on which specific aspects of the patient's physical condition should be considered (for example, ongoing weight-loss, insufficient fluid intake, etc).

In general, the (reconstructed) wish of the patient was considered to be very important, though not always decisive. The wish of the family (ie not the reconstructed wish of the patient) and the opinion of the nursing staff could be considered in decision making, but should never be decisive. Nevertheless, the more a physician was in doubt, the more the opinion of family and nursing staff was taken into account. Predicting the outcome of a specific treatment for an individual patient was considered to be very difficult, although some professionals were reasonably confident that their estimates were usually accurate when prognostic factors such as the degree of dyspnoea, the presence of tachypnoea and the intake of fluids were taken into account. The necessity for a guideline concerning non-treatment decisions to do with pneumonia in demented nursing home patients was fully confirmed by six of the nine professionals. Of the remaining three, only one did not feel the need for any support by means of a guideline.

For the next step (step 4), the project group decided to construct the guideline in the form of a “checklist of considerations”. A guideline in the form of a traditional protocol was considered to be too directive. The checklist was based on the results of the first three steps. A front-page (table 1) was added, explaining the origin, purpose, structure and definitions.

In the next step (step 5), all the invited experts agreed to take part in the expert panel and gave their comments on the first concept. The only major adaptation was made for the section “medical aspects”. This section was cleared of ethical aspects, which were moved to a new “best interest” section. Possible harmful medical effects, ie the “burden of treatment”, were separated from other medical aspects. Some questions were reformulated to prevent biasing the physician.

As a result of the pilot study (step 6), further adaptations were made. The main adaptation concerned the summary, which until then determined the routing of the checklist. This proved to be impractical and confusing. The summary was made optional, for example one could give a short overview of a complex case.

In step 7, all the experts who were consulted in step 5 now approved of the “checklist of considerations” as a whole, though some of the experts still required minor (mainly textual) adaptations. Subsequently, the Dutch Society for Nursing Home Physicians (NVVA) granted authorisation for use of the checklist in an evaluation study.

Discussion

With the method described above, it proved to be possible to develop a nationally and internationally approved “checklist of considerations” for end-of-life decisions in psychogeriatric patients, ie the decision whether or not to treat pneumonia curatively. Though time-consuming—it took three years to develop the complete checklist—the method guarantees input from the literature and the nursing home profession, as well as input from experts in other disciplines. Widespread support was obtained from renowned experts, as well as from the Dutch Society for Nursing Home Physicians, which will greatly enhance the acceptance of the guideline.

It has been widely accepted that physicians should be guided by four ethical principles: respect for autonomy, doing good, doing no harm, and justice. In daily practice, medical decision making is mainly guided by medical considerations. This is confirmed by the results of the questionnaire and the meeting of professionals, and is reflected in the “checklist of considerations”, as medical effectiveness of the treatment is required and is considered first (doing good and doing no harm). However, information from the literature on prognostic factors was scarce and of poor quality, and, moreover, no consensus could be reached on these medical prognostic factors. Therefore, this aspect had to be presented in general terms, so that a physician could make his/her own judgment. The second main point of the “checklist of considerations”is the patient's wish. Most NHPs and professionals thought the wish of the patient expressed when the patient was still competent was important, though medical considerations prevailed, according to the questionnaire and the meeting of professionals. Of course, the patient's wish should be respected (autonomy). Most psychogeriatric patients, however, can no longer grasp the consequences of their decision, or are incapable of communicating their wishes. When direct information about the wish of the patient is not available, the family or representative should be asked to reconstruct the patient's wish. Though the wish of the family itself was not regarded as decisive, most of the NHPs and experts agreed that the family should be involved in the decision making process. The position of the nursing staff was also thought to be important. The fourth principle, justice, is implicitly reflected in the checklist: only costs and benefits for the individual, and not for the community, are at stake. In order to make a decision based on ethical principles, information about the patient characteristics that determine the course and consequences of pneumonia is also necessary. As these characteristics are still largely unknown, both the physician and the family are compelled to give their personal opinion of what is in the patient's best interest when the patient's wish is unknown. This constitutes the third main point on the “checklist of considerations”.

Although the “checklist of considerations” has been developed for use by NHPs in the Netherlands, and for the specific decision about whether to treat pneumonia curatively in psychogeriatric patients, the problem addressed here is universal. Thus, the method may very well be applicable for developing guidelines to cover the broader scope of non-treatment decisions in incompetent patients. It took a long time to complete the development process, mainly due to the many refinements that had to be made. As these pitfalls can now be avoided, we are confident that developing a new guideline according to this method will be much less time-consuming. According to the nature and extent of the underlying evidence and international variations in law, the process of developing guidelines might need to be adapted for each end-of-life decision. If the literature does not provide sufficient basis for guideline formulation, other sources of information can be used (expert opinion, extrapolation of the results of studies with a different population, analysis of routinely collected data).²² The information provided by the ultimate users of the “checklist of considerations”, ie NHPs, is especially important in terms of the applicability and acceptability of the guideline.

The relative weight of each prognostic factor will be evaluated in a subsequent prospective study, designed to produce data on prognosis (chance of cure, quality of life) of both curatively and palliatively treated patients. The resulting indicators for the prognosis will be included in the “checklist of considerations”. Rigorous evaluation is needed to investigate the actual effect of guidelines on practice.²³ Meanwhile, we have investigated the acceptance of the contents, and the applicability of the “checklist of considerations” in practice. A further paper, evaluating the “checklist of considerations” is currently submitted for publication.