The analysis, though transparent, has methodological shortcomings. The authors seemed to conclude that two-day screening should be adopted because it is more accurate and only slightly more expensive than one-day screening. Thus, their judgement is of an incremental nature but is made arbitrarily. Why is it worth incurring the extra cost of the two-day method but not the three-day one? If current practice is no screening, the decision should be to choose the most effective method that falls within the boundary of acceptable cost-effectiveness. If the option were to move from one- to two-day screening, an incremental analysis should be performed and the decision based on the maximum willingness to pay for a cancer case detected. It is difficult to know what this value is. If the authors had included the costs of cancer treatment and final outcomes, these could be compared with a cost per life year gained or a cost per QALY.
The assumptions that the three-day method is 100% sensitive and that colonoscopy is a perfect diagnostic test could have been subjected to sensitivity analysis or, if final costs and outcomes had been calculated, threshold analysis to see how far these assumptions can be relaxed before the screening programme is no longer cost-effective. The authors adjusted the effectiveness figures based on the assumption of 100% compliance with colonoscopy but state that actual compliance was more relevant than hypothetical compliance. While they claimed that there is no significant difference between these two scenarios, this appears to have been based on the assumption that those who comply are the same as those who do not: it may have been better had the figures not been adjusted.