Background: Duloxetine is a serotonin and norepinephrine reuptake inhibitor and may be useful for treating women with stress urinary incontinence (SUI) in general practice.

Objective: The objective of this study was to examine the cost-effectiveness of 2 duloxetine strategies (duloxetine alone and duloxetine after inadequate response to pelvic floor muscle training [PFMT]) compared with PFMT or no treatment for women aged>or=50 years with SUI.

Methods: A Markov model with a 3-month cycle length was developed, with a time horizon of 5 years. Incontinence severity was based on incontinence episode frequency per week (IEF/week). Four SUI health states were distinguished in the model: no SUI (0 incontinence episode [IE] per week), mild SUI (19 IEs/week), moderate SUI (10-25 IEs/week), and severe SUI (>or=26 IEs/week). Transition probabilities were calculated, based on published evidence, expert opinion, and demographic data. Outcomes were expected total societal costs and expected IEs. The analysis was performed from the societal perspective of The Netherlands, and all costs were reported in year-2002 euros. One-way sensitivity and probabilistic sensitivity analyses were performed.

Results: In the model, providing PFMT cost euro0.03/IE avoided, compared with no treatment. Duloxetine after inadequate PFMT cost euro3.81/IE avoided, compared with PFMT One-way sensitivity analyses indicated that these results were robust regarding variation in age, IEF/week, and discount rate. Below the ceiling ratio of euro3.65/IE avoided, PFMT had the highest probability of being cost-effective. With higher ceiling ratios, duloxetine after inadequate PFMT had the highest cost-effectiveness probability.

Conclusions: Treating patients with duloxetine after inadequate PFMT response yielded additional health effects in the model, but would require society in The Netherlands to pay euro3.81/IE avoided for women aged>or=50 years with SUI being treated in general practice. It is up to policy-makers to determine whether this ratio would be acceptable.