Interventions:
The treatment strategies appeared appropriate; the authors did not discuss any other relevant alternative strategies. The most appropriate comparator (usual care) was included as an alternative, although it was unclear what the 'usual care' strategy actually entailed.
Effectiveness/benefits:
Details of the systematic search used to identify relevant trials were clearly reported. Details of trial follow-up, life-years and probability of hospitalisation for each strategy arm for each of the included trials and age subgroups were reported. An overview of the estimate values used as inputs in the model was not provided. A brief description of the methods used to derive utility was reported, but no actual values were reported. The meta-analysis of five trials was not described in detail. Risk ratios were produced from the meta-analysis for binary outcomes, but did not infer the risk of hospitalisation. It was unclear what method was used to synthesize hazard ratios.
Costs:
The costs and the sources used to derive them were clearly reported. Costs were appropriately discounted and appropriate inflation techniques were applied. Individual costs for each of the resource use items for each of the three strategies were clearly reported. Costs were specific to the UK context.
Analysis and results:
Only brief details of the model were supplied. There was no model diagram or description of the model structure, so the appropriateness of the model could not be assessed. The model was an update of a previous model used in NICE guidance, and the publication details were given (see Other Publications of Related Interest) for more information on the model.
Appropriate distributions were assigned to parameters in the probabilistic sensitivity analysis. Probabilistic sensitivity analysis was the most appropriate method to adopt when assessing the effect of parameter uncertainty on the results of a non-linear model.
The authors stated that, although UK-specific costs were used in the analysis, the results may be broadly generalisable to European countries as the most conservative estimates and assumptions were used (against the natriuretic peptide strategy), with relatively low incremental cost-effectiveness ratio results which corresponded well with all of the international trial analyses.
Concluding remarks:
Although the reporting was limited in a few places, the methods seemed appropriate. Overall, the authors' conclusions appear to be appropriate.