Interventions:
The selection of the interventions was appropriate as the available strategy for patients with heart valve disease was no surgery.
Effectiveness/benefits:
The analysis was on data derived from a long-term database, which held administrative data on patients who had undergone AVR. The large sample of patients and the long-term follow-up should have ensured the validity of these estimates, but the use of an administrative database is associated with limited internal validity, especially given the long time frame, which mixed different eras of heart surgery. The authors acknowledged that the clinical evidence for the natural history of heart valve disease was derived from old studies with relatively small sample sizes. Medical treatment changes over time and the life expectancy of patients with severe heart valve disease may be greater now than that used in this analysis. These issues may have affected the validity of the clinical comparison. QALYs are a valid benefit measure and appropriate for this disease. The authors noted that several other tools were available for the assessment of quality of life, but these were not appropriate for their historical cohort. NYHA functional class was used as a proxy for quality of life and these two measures were demonstrated in the literature to have some correlation.
Costs:
The authors did not explicitly report the economic perspective adopted, but the categories of costs suggest a payer’s perspective. These costs were presented as macro-categories and were not broken down into individual items. Details on the sources used were not clearly reported. The authors acknowledged that cost estimates were not precise and that some approximation was required for certain categories. Discounting would have been appropriate, but it was not mentioned in the paper. In general, the economic analysis had some limitations.
Analysis and results:
The costs and benefits were clearly reported and synthesised, but the issue of uncertainty was not investigated. The key assumption that the costs for non-surgical patients were zero was biased against the AVR strategy. The authors noted some potential drawbacks of their study, which are reported in the relevant sections above.
Concluding remarks:
The study has some methodological limitations which might affect the validity of the authors’ conclusions.