Cost-effectiveness analysis of solifenacin versus oxybutynin immediate-release in the treatment of patients with overactive bladder in the United Kingdom.

Objective: To carry out a cost-utility analysis comparing initial treatment with solifenacin 5 mg/day vs oxybutynin immediate-release (IR) 15 mg/day for the treatment of patients with overactive bladder (OAB) from the perspective of the U.K. National Health Service (NHS).
Methods: A Markov model with six health states was developed to follow a cohort of OAB patients treated with either solifenacin or oxybutynin during a 1-year period. Costs and utilities were accumulated as patients transited through the health states in the model and a drop-out state. Some of the solifenacin patients were titrated from 5 mg to 10 mg/day at 8 weeks. A proportion of drop-out patients were assumed to continue treatment with tolterodine ER. Utility values were obtained from a Swedish study and pad use was based on a multinational clinical trial. Adherence rates for individual treatments were derived from a U.K. database study. For pad use and utility values, the drop-out state was split between those patients who were no longer receiving treatment and those on second-line therapy. Patients on second-line therapy who drop-out were referred for a specialist visit. Results were expressed in terms of incremental cost-utility ratios.
Results: Total annual costs for solifenacin and oxybutynin were £504.30 and £364.19, respectively. First-line drug use represents 49% and 4% of costs and pad use represent 23% and 40% of costs for solifenacin and oxybutynin, respectively. Differences between cumulative utilities were small but were greater for solifenacin (0.7020 vs. 0.6907). The baseline incremental cost-effectiveness ratio was £12,309/QALY.
Conclusion: Under the baseline assumptions, solifenacin would appear to be cost-effective with an incremental cost-utility of less than £20,000/QALY. However, small differences in utility between the alternatives and the large number of drop-outs means that the results are sensitive to small adjustments in the values of utilities assigned to the drop-out state.