METHODS A Markov model was developed to evaluate the cost-effectiveness of levodopa/carbidopa/entacapone (LCE;Stalevo), in the treatment of patients with Parkinson's disease (PD) and end-of-dose motor fluctuations (wearing-off). LCE, with or without other antiparkinsonian medications, was compared to UK standard care, comprising traditional levodopa/ dopa-decarboxylase inhibitor (DDCI) with other antiparkinsonian medications (e.g. selegiline or dopamine agonists) added as needed. The costs and outcomes of both treatments were projected over a period of 10 years from the perspective (a) of society as a whole and (b) of the UK National Health Service (NHS). Sensitivity analyses, including second-order Monte Carlo simulations, were performed to assess the confidence level of the primary results. RESULTS Treatment with LCE produced an average gain of +1.04 quality-adjusted life-years (QALYs) per patient (2.57 vs. 1.53) in the base-case analysis (discount rate 3.5%). This gain was accompanied by a reduction in the total 10-year direct cost of care to society of 10198 pounds per patient ( approximately E14800). From the societal perspective, therefore, LCE was dominant, producing better clinical outcomes with lower costs. This dominance was reiterated in all sensitivity analyses of society-focused analysis, including a shortening of the time-frame to 5 years. Although treatment with LCE resulted in an increase in direct costs per patient of 3239 pounds (25756 pounds versus 22517 pounds) to the NHS over the 10-year period analysed, the incremental cost-effectiveness ratio (ICER) of LCE was only 3105 pounds per QALY gained (approximately E4500). All ICERs to the NHS remained below 3800 pounds per QALY gained in univariate sensitivity analyses applying different discount rates. When a shorter, 5-year, time-horizon was analysed, the NHS-related ICER for LCE was 6526 pounds per QALY gained. All these ICERs are within the range usually considered to indicate acceptable or highly acceptable cost effectiveness (defined as < 30000 pounds per QALY gained). The results of the Monte Carlo simulations indicated that the likelihood of LCE being either 'dominant' or more effective at an 'acceptable cost' from either the societal or the NHS perspective was high, exceeding 96% in the base-case sensitivity analysis, and was 93% even when all the uncertainties associated with the model were taken into consideration simultaneously. In particular, compared to standard care, the probability that LCE would provide better outcomes at a lower cost to society as a whole was 77% in the base-case sensitivity analysis and 72% in the scenario involving the highest degree of uncertainty. CONCLUSIONS In the UK the use of LCE to treat PD patients with wearing-off is beneficial to individual patients and likely to offer money savings to society as a whole, compared with UK standard therapy. The added cost of the medication itself is exceeded by the savings made in other direct costs of PD, mainly those relating to social care or PD-related private expenditures.