OBJECTIVES: One of the challenges in obtaining reimbursement for novel technologies for small patient groups arises when establishing clinical effectiveness using conventional statistical analysis. Thus, conducting reliable cost-effectiveness analyses is compromised by significant uncertainty related to outcome measures. This challenge is recognized in academia, by public decision-makers and in the pharmaceutical industry. Different initiatives have been made to overcome the problem like innovative reimbursement agreements. However, solving the statistical challenges enables decision-making based on results from economic evaluations. This study is a proof of concept (POC) for a method for estimating clinical effects of treatments for which the biochemical mechanisms are well understood using small sample sizes, but without lowering the level of evidence. The method is applied to a study on mannose supplementation for phosphomannomutase 2 deficiency (PMM2-CDG). PMM2-CDG is a disease for which the metabolic pathways are well understood, but where the rarity of the condition limits the evidence of possible treatments due to lack of power.

METHODS: Using Bayesian statistics on clinical trial data, treatment effects can be inferred with reduced standard error by incorporating in vitro experiments and knowledge about the physics of the protein catalysts.

RESULTS: The effect of mannose supplementation for PMM2-CDG is a beneficial 84% increase in glycosylation (95% CI: 36-133). After incorporating in vitro experiments, the range of the 95% CI was reduced to 51-134. Estimations are based on 20 patients in the trial data.

CONCLUSIONS: The results of this POC suggest that evidence for clinical treatment effects is attainable for small sample size studies, where lack of power is a major concern. However, researchers should ensure that statistical significance of the in vitro treatment effect does not outweigh that of the clinical trial. Thus, new guidelines are warranted and related adjustments in the current framework for economic evaluations is needed.