Responders and non-responders in a study on medical end-of-life decisions in Denmark, the Netherlands, Sweden and Switzerland.

Objectives: To determine the direction and magnitude of participation bias in end-of-life research.

Methods: Within the framework of a European survey on medical end-of-life decisions, a non-response study was conducted among physicians in Denmark, the Netherlands, Sweden and Switzerland. People were asked about their attitudes and experiences in the area of medial end-of-life decision. The response rates ranged from 12.8% (Netherlands) to 39.4% (Switzerland). The responders (n = 5403) and the non-responders (n = 866) were compared regarding socio-demographic characteristics, experiences with terminal patients and agreement with attitudes towards "end-of-life decisions". The reasons for non-participation to the study were analyzed.

Results: Non-response did not cause socio-demographic distortion, but non-responders had statistically significantly fewer terminal patients than responders. Agreement rates were statistically significantly higher among responders than among non-responders for euthanasia, non-treatment decision and life-preserving statements. Neutral answers were statistically significantly more frequent among non-responders than among responders for life-preserving and euthanasia statements. The most commonly mentioned reason for non-participation was "lack of time".

Conclusion: Non-participation does cause an overestimation of proponents of life-shortening, as well as of life-preserving end-of-life decisions. Non-responders more often have ambiguous attitudes towards end-of-life decisions than responders.