Can observational studies provide a realistic alternative to randomized controlled trials in palliative care?

Evidence-based medicine demands 'gold standard' randomized controlled trials (RCTs). If strict criteria of quality, validity, and size are met, observational studies give the same result. Given the dearth of RCTs in palliative care, our aim was to identify good observational studies using PubMed searches and e-mail letters to experts in palliative care. The prior intention was provide the most comprehensive description possible to date of observational studies in palliative care, rather than to perform any statistical analyses. Three hundred and forty abstracts of study reports were identified, of which 27% (91) included > or = 200 subjects and 8% (27) > or = 1000 subjects. In reports with > or = 200 subjects, 51% included only cancer patients, and 42% included heterogeneous 'palliative care' patients. Prospective and retrospective studies accounted for 38% and 32% of all reports with > or = 200 subjects. In reports with > or = 1000 subjects, 59% were retrospective and 19% prospective. Patients had some input in 26% of studies with > or = 200 subjects, and 15% with > or = 1000 subjects. Only 12 prospective reports had one specific intervention. We found that palliative care is deficient not only in RCTs, but also good quality observational studies. Those that exist are extremely heterogeneous in subject, design, outcome reporting, and intervention.