OBJECTIVES: Variations in appraisal outcomes by European Health Technology Assessment (HTA) agencies may be explained by differences in legal and institutional context and evaluation criteria. Here we report results of a comparative study analyzing appraisal outcomes by the German Federal Joint Committee (Gemeinsamer Bundesausschuss, G-BA), the French Haute Autorité de Santé (HAS), and the British National Institute for Health and Care Excellence (NICE).

METHODS: We searched matched intervention pairs from all available G-BA resolutions, HAS reports and NICE single technology appraisal guidance published between January 2011 and June 2018. We compared appraisal outcomes of matched pairs by G-BA, HAS and NICE overall, and separately for non-cancer and cancer drugs. Subsequently, we explored additional attributes related to cancer drugs, such as orphan drug designation in Germany, reimbursement rates in France, and end of life considerations in England. The relationship between attributes and appraisal outcomes was tested for statistical significance using a chi-square test or, if required, Fisher´s exact test.

RESULTS: We observed outcome variations in 64/102 (63%) matched intervention pairs. Overall, NICE (85/102, 83%) evaluated new health technologies more favourably than HAS 59/102 (58%) and G-BA (72/102, 71%). Matched cancer treatment pairs (33/58, 57%) were less divergent compared with non-cancer treatments (31/44, 70%). G-BA cancer drug appraisals (49/58, 85%), however, were more favourably than cancer-related recommendations by NICE (42/58, 72%) and HAS (37/58, 64%). An orphan designation significantly influences G-BA cancer drug appraisals (p<.05, chi-square test). While HAS results indicate that the reimbursement rate correlated with the assessment of added clinical value (p<.05, chi-square test), we found no significant impact when testing NICE recommendations for the consideration of end of life criteria (p>.05, Fisher´s test).

CONCLUSIONS: Findings confirm the frequent occurrence of variation in HTA outcomes, although cancer-related appraisal outcomes seem to diverge less compared to non-cancer results.