Interventions:
Both interventions were well described, including the dosages used. The selection of the interventions was justified and the comparator would appear to represent current practice in the authors’ setting.
Effectiveness/Benefits:
The effectiveness data for first-tier treatments were derived from a published randomised controlled trial, which is considered to be the 'gold' standard study design when comparing health interventions. However, the authors provided no information as to how this study was identified, thus it is unclear whether the authors used the best available evidence. The effectiveness data for second- and third-tier treatments were derived from a physician survey. The authors gave appropriate details of the survey, including the participant composition and sample.
Costs:
The perspective was clear and it would appear that all the relevant costs were considered. The costing was well reported, with clear details of the sources of resource use and unit costs. Discounting was not reported but, as no time horizon was stated, it is not possible to know if this was an issue.
Results and Analysis:
Overall, the analytical approach was well reported, with the model structure described in full detail along with a graphical representation. The authors did not undertake a synthesis, possibly because the intervention was the dominant strategy (i.e. less costly and more effective). The results were presented clearly and were found to be robust to changes in both one-way and probabilistic sensitivity analyses. Probabilistic sensitivity analyses are considered the most thorough way to fully capture parameter uncertainty. However, the use of triangular distributions as a representation of uncertainty, while apparently simple, is also very limited given the characteristics of the distribution. The authors acknowledged and highlighted the main limitations of their analysis.
Concluding remarks:
The quality of the study was satisfactory. Despite some limitations of the clinical data, the authors presented a reasonably transparent analysis and it is likely that the results reflected the available evidence.