Selection bias in patients' recruitment for palliative care research
Borgsteede, SD, Institute for research in extramural medicine, Dept. Social Medicine, Amsterdam, Nethlerlands, Deliens, L, Institute for research in extramural medicine, Dept. Social Medicine, Amsterdam, Nethlerlands, Francke, AL, Institute for research in extramural medicine, Dept. Social Medicine, Amsterdam, Nethlerlands, Stalman, WAB, Institute for research in extramural medicine, Dept. Social Medicine, Amsterdam, Nethlerlands, Van der Wal, G., Institute for research in extramural medicine, Dept. Social Medicine, Amsterdam, Nethlerlands

Object: Participation of patients in palliative care research is important, but may be problematic because these patients are both physically and mentally vulnerable. The research question of this paper is how GPs select their end of life patients for participation to a palliative care study.
Methods: 20 GPs approached 34 patients with a life expectancy of less than 6 months to be interviewed and registered the patients they did not approach. In a qualitative research design, we interviewed these GPs and asked for reasons not approaching some of their patients. Data were analysed inductively and separately by the first two authors.
Results: 13 patients who did meet the selection criteria were not approached by their GP due to: short life expectancy (n=2), cognitively/emotionally/physically not able to be interviewed (n=3), or other reasons (n=8). GPs acknowledged the need for researching vulnerable patients, but most GPs did not approach their vulnerable patient for an interview. The following themes emerged for not approaching these patients: denial of terminal illness, instable phase of disease and communication problems between patient and GP.
Conclusions: Vulnerable patients who need special attention with respect to physical and psychosocial care are less frequently by GPs for participation to a palliative care study. Hence, the most vulnerable patients are underrepresented in palliative care research, whereas the need to study these patients is obvious. New inclusion methods and new ways to approach patients are needed. The consequences of these findings in qualitative research should be explored to quantitative studies, e.g. randomized controlled trials.