[Criteria for medical prioritisation: results from a regional survey and methodological reflections].

Aims: In Germany, in contrast to many foreign countries, scientists and medical professionals have been discussing prioritisation in medicine almost without consulting German citizens. We address the question of what questionnaire surveys can contribute to the understanding of citizens' attitudes towards prioritisation - with a focus on some difficulties and challenges of the method.
Method: We conducted a postal survey with a random sample of 3 000 residents of the City of Lübeck (age ≥18). Respondents were asked to appraise different substantial and procedural criteria for prioritisation in medicine. In addition to descriptive statistical analyses, logistical regression models were performed to identify potential explanatory variables for the appraisal of prioritisation criteria.
Results: The response rate was 45.6% (N=1 363). Some prioritisation criteria are accepted by the majority: severity of disease, effectiveness of an intervention and a firm evidence base. Other criteria were appraised controversially: personal life-style, responsibility for family members and general prioritisation of children. A patient's responsibility in society and age as well as an intervention's cost-benefit ratio were generally rejected. The results of logistic regression analyses showed some significant but minor effects of demographic and health-related variables. The citizens in our study want decision-making procedures in health care to be transparent and equally applied to all patients. According to the survey respondents decisions about the catalogue of services of Germany's statutory health insurance should mainly be made by doctors. The statutory health insurance as well as patients and scientists also should take part in the decision-making procedure.
Discussion: Comparing our results to those of a national interview survey reveals some relevant differences: The respondents' assessment of some substantial criteria seems to vary according to the contextualisation and wording of the items. We found less difference - but still some inconsistent results - in the participants' appraisal of potential decision-makers in health care. To our surprise, the logistic regression models including standard demographic and health-related variables account for only a small proportion of the variance of all dependent variables.
Conclusion: Our discussion emphasises some difficulties and challenges of questionnaire surveys on prioritisation criteria - reflecting on the state of the German debate on prioritisation. There has been hardly any public discussion on this issue prior to our survey in autumn 2009. It is thus unlikely that people have been able to state well-informed preferences. Instead they seem to have followed some kind of "social reflexes" depending on the context and wording of each item. Subsequent studies on preferences and priorities should (i) more closely assess the understanding of each item in advance and (ii) adapt the aims of their study and its methodology to the actual stage of the public discourse on the topic in question.
(© Georg Thieme Verlag KG Stuttgart · New York.)