Cost-effectiveness of newer antidepressants compared with tricyclic antidepressants in managed care settings

Our aim was to determine the cost-effectiveness of newer antidepressants compared with tricyclic antidepressants in managed care organization settings.

We employed cost-utility analysis based on a clinical decision analysis model derived from published medical literature and physician judgment. The model, which represents ideal primary care practice, compares treatment with nefazodone to treatment with either imipramine or fluoxetine or to a step approach involving initial treatment with imipramine followed by nefazodone for treatment failures. The outcome measures were lifetime medical costs, quality-adjusted life years (QALYs), and costs per QALY gained.

The base case analysis found that nefazodone treatment had $16,669 in medical costs, compared with $15,348 for imipramine, $16,061 for the imipramine step approach, and $16,998 for fluoxetine. QALYs were greatest for nefazodone (14.64), compared with 14.32 for imipramine, 14.40 for the step approach, and 14.58 for fluoxetine. The cost-effectiveness ratio comparing nefazodone with imipramine was $4065 per QALY gained. The cost-effectiveness ratio comparing nefazodone with the step approach was $2555 per QALY gained. There were only minor differences in costs and outcomes between nefazodone and fluoxetine, with nefazodone resulting in $329 fewer costs and 0.06 more QALYs. The cost-effectiveness ratios comparing fluoxetine with imipramine and with the step approach were $6346 per QALY gained and $5206 per QALY gained, respectively. In the sensitivity analyses, the cost-effectiveness ratios comparing nefazodone and imipramine ranged from $2572 to $5841 per QALY gained. The model was most sensitive to assumptions about treatment compliance rates.

The findings suggest that nefazodone is a cost- effective treatment compared with imipramine or fluoxetine treatment for major depression. Fluoxetine is cost-effective compared with imipramine treatment, but is estimated to have slightly more medical costs and less effectiveness compared with nefazodone. The basic findings and conclusions do not change even after modifying key model parameters.