Background: In 2001, Good and Stafford published a study attempting to show that inpatient palliative medicine is evidence-based.
Aim: To critically examine the papers cited in support of the claim that palliative medicine is evidence-based.
Method: All the papers cited were re-examined by at least two reviewers. The papers were graded using a tool used in the production of 22 evidence-based national clinical guidelines over the past three years. Differences of opinion about quality were resolved by discussion or independent arbitration.
Results: The 41 papers cited by Good and Stafford included 25 randomized controlled trials (RCT), six other controlled studies and 10 uncontrolled studies. Sample size. Mean sample size was 160 (range 9-1404). The median sample size was 56; 19/41(46%) had <50 subjects; 30/41 (73%) had <100. Power calculations. 16/25 (64%) of RCTs, had no power calculations, making it difficult to draw conclusions about true differences between groups. Quality rating and grade of evidence: Of the RCTs 15/25 (60%) had a high risk of bias; of the other controlled studies 5/6 (83%) had a high risk of bias. Applicability. The studies were generally applicable to a population with advanced disease: 30/41 (73%) related with populations with advanced disease.
Conclusions: This study demonstrates that studies cited in support of palliative medicine as an evidence-based specialty are of variable, and at times poor quality. This study seems to show that the claim that inpatient palliative medicine is evidence-based is at best tenuous and at worst misleading. Palliative Medicine 2007; 21 : 623-627.