Analytical approach:
The analysis was based on a probabilistic Markov model, with a 10-year horizon. The authors stated that the perspective was that of the Italian Health Service (IHS).
Effectiveness data:
The key clinical endpoint was the treatment efficacy, which was the reduction in hospitalisations (due to the worsening of heart failure) plus the improvement in disease symptoms. These data were derived from the Valsartan in Heart Failure Trial (Val-HeFT), which analysed 5,010 patients with New York Heart Association (NYHA) classes II, III, or IV heart failure. There were 2,511 patients (2007 men; mean age 62.4 years) in the valsartan group and 2,499 patients (2000 men; mean age 63 years) in the control group. The length of follow-up was 23 months. A key assumption was that the efficacy of treatment was zero from the 24th month to the end of the simulation. Other inputs, such as the prevalence of heart failure and mortality by NYHA class, were derived from country-specific databases.
Monetary benefit and utility valuations:
The utility values were from quality of life estimates from patients with CHF and low LVEF collected during a large published study, of the general English population, that used the Short Form (SF-36) Health Survey. Regression models were used to convert the SF-36 results into utility weights.
Measure of benefit:
Life-years (LYs) and quality-adjusted life-years (QALYs) were the summary benefit measures and they were discounted at an annual rate of 3.5%.
Cost data:
The economic analysis considered the costs of drugs and heart failure hospitalisations. The drug costs were based on the maximum price to the IHS for a pack of 28 160mg tablets and the mean dose used in the Val-HeFT. The cost of hospitalisations was based on diagnosis-related group reimbursements at a national level. The costs were in Euros (EUR) and were discounted at an annual rate of 3.5%. The price year was 2007.
Analysis of uncertainty:
A probabilistic sensitivity analysis was undertaken on all the inputs, using published confidence intervals or a range of ±10%. Subgroup analyses were carried out to consider various demographic and clinical factors.