Analytical approach:
A Markov model was constructed to estimate the cost-effectiveness of the treatment options over 10 years. The authors stated that the analysis was conducted from the perspective of the UK National Health Service and Personal Social Services.
Effectiveness data:
The key effectiveness data was the progression free survival and overall survival. A systematic review was undertaken to identify the effectiveness evidence. One randomised controlled trial (RCT) was identified for cetuximab and panitumumab each for the KRAS population. For cetuximab plus irinotecan the one RCT identified did not have KRAS status as an inclusion criteria, and two observational studies were used to adjust the estimates.
The relative treatment effects were obtained from an indirect comparison of the pairwise effect estimates. Survival curves were used to determine the number of patients in progression free survival or overall survival at any time, and the time in the progressed disease state was calculated as overall survival-progression free survival. The mean overall survival for BSC in the panitumumab versus BSC RCT was reduced to allow for the substantial crossover of patients from BSC to panitumumab.
Monetary benefit and utility valuations:
Treatment specific utility values were assigned to the three model health states: progression free survival, progressed disease state and death. Most values were obtained directly from Merck Serono (the manufacturer of cetuximab), using mostly RCT data as the sources. Adjustments were made by the authors to bring Merck Serono's estimates in line with general population values, maintain consistency or comply with the authors' assumptions.
Measure of benefit:
The health benefit was measured in quality-adjusted life-years (QALYs). Future benefits were discounted at a rate of 3.5% per annum.
Cost data:
Direct costs were included in the analysis. The cost items included drug acquisition and delivery, KRAS testing, out-patient visits, scans, BSC in progressed disease state and adverse event treatment. Costs were estimated using NHS Reference Costs 2008-9, Merck Serono values, British National Formulary (BNF) 63 and authors' assumptions. Treatment duration was estimated using RCT data, personal communication and authors' assumptions. In the calculation of the cost of KRAS testing, it was assumed that 54% of patient were KRAS wild-type (informed by Merck Serono). Costs were reported in 2011-2012 GBP (£) and were inflated where necessary. Future costs were discounted at an annual rate of 3.5%.
Analysis of uncertainty:
One way sensitivity analysis and probabilistic sensitivity analysis were conducted. A cost-effectiveness acceptability curve (CEAC) was produced, which shows the probability that each alternative was cost-effective at different willingness to pay (per additional QALY) thresholds.