Interventions:
The authors justified their selection of the comparators. Pregabalin was the newest antiepileptic drug to be approved by the US Food and Drug Administration for use as an add-on therapy for adults with partial onset seizures. The usual care was restricted to first-line drugs without any add-on ones. Thus, the two strategies were appropriately selected.
Effectiveness/benefits:
RCTs were used to derive the bulk of the evidence on treatment efficacy. They are generally regarded as a valid source of evidence given their methodological strengths. Thus, the clinical inputs were likely to be valid. The details of the patients’ characteristics, follow-up, and other methodological issues, such as how the data were pooled, were not reported. Similarly, the details on the source for safety data (i.e. the meta-analysis) were not given. The derivation of the utility valuations was based on a single survey, which might not represent the most appropriate source of evidence, but the authors pointed out that there were few publications on the quality-of-life of patients on add-on therapy. QALYs are a validated benefit measure and are comparable with the benefits of other health care interventions, and this will help to overcome the difficulties associated with a disease-specific measure such as SFDs.
Costs:
The analysis of costs was restricted to drug costs, which were calculated using typical US sources. The authors did not state the perspective adopted, but the use of Medicare costs in the sensitivity analysis suggests that the viewpoint was that of the third-party payer. The unit costs and quantities of resources used were presented separately. The price year was not explicitly reported. The authors stated that considering only drug costs led to conservative estimates for pregabalin, since the favourable impact of a more effective drug on other health care resources was not included in the model.
Analysis and results:
The costs and benefits were appropriately synthesised. The issue of uncertainty was satisfactorily addressed in the probabilistic sensitivity analysis. The results of both the base-case and the sensitivity analyses were clearly presented. The methodology of the decision model was described. The authors noted that a potential limitation of the analysis was the use of a one-year time horizon, which required assumptions on the clinical impact of treatment, given the 12-week follow-up period in the pregabalin RCTs. Another potential drawback was the threshold of 50% reduction in seizure frequency at six months which determined treatment continuation. Such a definition of treatment response might not reflect the real-world pattern of care and might not be generalisable to other health care systems.
Concluding remarks:
The study was well conducted and satisfactorily presented. The authors’ conclusions were valid.