OBJECTIVES: Pharmaceutical options for insomnia that treat both sleep induction and maintenance have only recently been launched in the U.S. The economic impact of treatment patterns with older drugs has not been thoroughly investigated in the literature. We hypothesized that since the older drugs only provided benefit for sleep induction, insomniacs who alternate therapy within one year of initiation would have greater economic burden compared to maintainers. METHODS: Treated insomnia patients were identified from Medstat Marketscan claims database with at least one prescription for existing insomnia agents during the study period (05/01/01 to 11/30/03). Change in therapy was defined as having a recorded prescription for an insomnia treatment other than the index study medication (estazolam, flurazepam, quazepam, temazepam, triazolam, low-dose trazodone, zolpidem tartrate or zaleplon) during the post-index period, which included overlapping augmentation or complete therapy switch. Economic costs included direct medical and pharmacy expenditures. One-year follow-up costs were compared using Wilcoxon tests for bivariate analyses and generalized linear models with gamma functions for multivariate comparisons. RESULTS: Approximately 88% of the sample did not have a change in insomnia treatment. Patients on monotherapy (n=24,540) differed from patients altering therapy (n=3,267) with respect to age, gender and baseline Deyo-Charlson score. Patients altering therapy had a higher frequency of insomnia diagnoses (22.4% vs. 11.4%, p<0.001) and other comorbidities than maintainers. Unadjusted and adjusted direct medical costs for patients with changes in insomnia therapy were approximately 66% and 67% greater than patients who remained on their index therapy ($16,053 vs. $9,798, p<0.001; and $16,580 vs. $9,942, p<0.001), respectively. Similar results were also observed among patients with available productivity data. CONCLUSION: Results point to a significant relationship between changes in therapy and economic burden. Further research would be warranted to find the reasons for therapy change and the potential economic benefit of newer therapies.