Whereas Health Technology Assessments (HTAs) by the National Institute for Health and Care Excellence (NICE) rely heavily on cost utility analysis, HTAs by the German Institute for Quality and Efficiency in Health Care (IQWiG) and the Federal Joint Committee (GBA) focus on an assessment of comparative effectiveness, rejecting a cost per quality-adjusted life year benchmark. The present study aimed to explore the differential impact of methodological choices by NICE and IQWiG/GBA on HTA outcomes.

We extracted data from all GBA decisions between January 2011 (when early benefit assessments were implemented) and April 2015 (cut-off date for the present study), as well as all single technology appraisals (STAs) by NICE published during the same period. We compared early benefit assessment results by IQWiG/GBA and by NICE overall, and by additional criteria including therapeutic area, clinical and incremental cost effectiveness, and patient-relevant endpoints.

During the study period, NICE issued guidance for 88 technologies (with 125 subgroups). GBA completed 105 appraisals (with 226 subgroups). We identified thirty-seven matched condition-intervention pairs; of these, twenty-four were evaluated differently by NICE and GBA. NICE recommended twenty-nine of thirty-seven interventions (78 percent), whereas GBA confirmed additional benefit for 21/37 only (57 percent; p< .05, two-tailed chi-square test). By therapeutic area, NICE was more likely to evaluate interventions for metabolic and cardiovascular disorders favorably, whereas IQWiG/GBA appraisals were more favorable for treatments of hematological and oncological diseases. Results including all HTAs were consistent with those for matched pairs.

Our results suggest that, overall, NICE tends to evaluate new interventions more favorably than IQWiG/GBA. However, our analysis revealed conspicuous differences by therapeutic area. The results are consistent with the hypothesis that different methodological choices may lead to systematic differences in decision making. It seems plausible that the observed differences reflect, at least in part, differences in underlying value judgments.