Analytical approach:
A Markov model was used to estimate the costs and outcomes of the two interventions. The time horizon was two years and the authors reported that a third-party payer perspective was adopted.
Effectiveness data:
PubMed was searched to identify the relevant studies published up until February 2008. The search strategy and key words used were reported and a manual search of reference lists from the retrieved articles and reviews was performed. The main effectiveness parameter was the probability of disease flare-up, per three months, with and without mesalazine maintenance therapy. This evidence was derived from the weighted average results of five randomised controlled trials comparing mesalazine with placebo for the maintenance of remission in ulcerative colitis.
Monetary benefit and utility valuations:
The utilities were from published studies, which used the time trade-off and standard gamble methods to elicit the utility weights. When the utilities for a certain health state were available from more than one source, a weighted average was used.
Measure of benefit:
The number of flare-ups of disease prevented and quality-adjusted life-years (QALYs) gained were the measures of benefit. As QALYs could be gained over a two-year period, future QALYs were discounted at an annual rate of 3%.
Cost data:
The costs included those relating to out-patient medications and medical care. The costs of medical care were derived from the year 2004 Medicare professional fee schedule Current Procedural Terminology (CPT) codes and the mean reimbursement for all US hospitals for the relevant diagnosis-related groups. These costs included the treatment of: in-patient flare-up; out-patient flare-up; colectomy; small bowel obstruction; abscess or sepsis; dehydration; and pouch excision. The costs of out-patient medications were from average wholesale drug prices, except those for infliximab infusion, which were from a published study. The price year was 2004 and, as the costs were incurred over a two-year period, future costs were discounted at an annual rate of 3%. All costs were reported in US dollars ($).
Analysis of uncertainty:
A series of one-way sensitivity analyses was performed to assess the impact of varying each model parameter. Two-way sensitivity analyses were performed by varying those parameters that were most influential in the one-way sensitivity analyses. A probabilistic sensitivity analysis was also undertaken by fitting distributions around each of the model parameters. Analyses were then run 10,000 times, and the results were presented in a cost-effectiveness acceptability curve.