Interventions:
The interventions under examination were appropriate comparators. PI represented the most commonly used agent in several countries, including Thailand, while CG had been recently proposed, in a meta-analysis, as an effective and cost-effective alternative to PI. These two antiseptics are also likely to be relevant in other settings.
Effectiveness/benefits:
The approach used to identify primary sources of data was not reported. However, the use of a meta-analysis to obtain relative risk estimates, and of administrative data from a local hospital to obtain baseline risk for the population analyses, appears appropriate given their internal validity (the meta-analysis) and their relevance to the authors’ setting (administrative source). A limitation of the analysis was the fact that little information on these sources was provided. Both benefit measures were derived using modelling and represent intermediate end points of the interventions. They are specific to the disease under examination.
Costs:
The perspective was clearly defined and it would appear that all the relevant costs have been considered. The authors stated that the adoption of a broad perspective, such as that of society, would have further favoured the CG strategy given the reduction in costs related to time lost from work. A breakdown of cost items was not given, and the unit costs were presented separately from resource quantities only for antiseptics. Other costs were reported as macro-categories, which limits the possibility of replicating the analysis in other settings. The price year was reported. Administrative data were used to derive the costs and the authors reported the key calculations made to determine total costs.
Analysis and results:
Given the dominance of CG over PI, the costs and benefits were not combined. Extensive information on the decision model was reported. The issue of uncertainty was addressed in a deterministic sensitivity analysis, which helped identify the most influential model inputs. The results of both the base-case and sensitivity analyses were extensively reported. The authors pointed out that their findings were very robust and, despite the fact that the study was carried out in a single academic centre, they may be generalisable to other settings.
Concluding remarks:
This cost-effectiveness analysis was satisfactorily performed in terms of transparent reporting and calculation of all model inputs and results. However, there were some limitations, namely the narrow perspective and the lack of detail on some data sources. The conclusions reached by the authors are robust and appear appropriate.