Interventions:
The interventions were poorly described. The authors listed the options, but gave no details of the procedure or the follow-up schedule, so it is unclear what each one entailed. The options appear to have been relevant to the study setting and generalisable to other settings.
Effectiveness/benefits:
The reporting of the effectiveness data was basic. The sources searched to identify relevant studies were reported, and the search was described as systematic, but the methods, search terms, and inclusion and exclusion criteria were not reported. The methods used to derive the clinical estimates were not given. The sources for the evidence of effectiveness were not referenced. The systematic review was in publication elsewhere, but no reference was given. It is not possible to assess whether the best available evidence was used. Some key outcomes, such as the potential for false positive results and their implications, the stage of cancer detected by screening, which will influence outcome, and the potential for injury from the tests, which varied considerably, were omitted from the effectiveness data.
Costs:
The reporting of the cost data was poor. The perspective was not stated, but appears to have been those of the public or private payers. Only the costs of the screening methods were considered. The resource items were listed, but their sources and prices were not reported nor referenced. The formula used to estimate the costs was given. The price year was not reported, and the time horizon was 20 years, but no discounting and no cost adjustments were reported. These factors reduce the study transparency. Some very important costs, such as for the treatment of incorrect diagnoses, the treatment for the stage of cancer detected, and the treatment for harms from screening, were omitted from the cost data.
Analysis and results:
It was appropriate to use modelling to synthesise the data from the published literature, but no incremental analysis of the screening options was conducted to consider the relative costs-effectiveness of each method. The analysis did not consider all the key costs and effectiveness for the different options. The reporting of methods and results was poor and there was no consideration of the uncertainty around the inputs for the model. The results are unlikely to be generalisable to other settings, and their validity for the study setting cannot be assessed. The authors did not discuss the limitations of their study.
Concluding remarks:
The methods and results were poorly reported. The authors’ conclusions are hard to assess and should be considered with caution.