To the Editor––Tunis et al. [1] present the results from a randomized, open-label, 1-year trial that evaluates the cost-effectiveness of olanzapine as first-line treatment versus a generic fail-first regimen for schizophrenia. The authors must be congratulated for a well-designed study and for a rigorous analysis of the results. There are many interesting and important results that are revealed by this analysis. Here, I would like to raise an issue that is tied to the economic evaluations of a fail-first design, which the authors do not account for in their results and fail to discuss in their conclusions.

The underlying economic rationale for the advantage of an algorithm with generic first-line treatment is an implicit individualization mechanism that enables patients, who benefit from generic drugs and for whom branded drugs do not offer any additional benefits, to stay on the generic drug. As expected, more patients switch from their assigned therapy in the CON group than in the olanzapine group [1]. Nevertheless, patients in the conventional antipsychotics group who do not switch and continue to benefit from the conventional antipsychotics would represent a recurring cost-saving
for as long as they receive them. For example, in Tunis et al.’s study [1], 31.4% of patients remain on conventional antipsychotics till the end of the study while 56.1% remain on olanzapine (figure 1 in Tunis et al. [1]). If we assume that out of the 31.4% of patients who presumably benefited from conventional drugs, at least 56.1% would have similarly benefited from olanzapine, then without a fail-first regimen the additional drug costs per patient can be estimated conservatively to be $367 (=0.314 × 0.561 × [3756 − 1652]). This translates to a 5-year cost-saving of $1731 per patient at 3% discount rate associated with a fail-first regimen, without any significant offset to the effectiveness side. With generic second-generation antipsychotics, these savings may be even more with a fail-first regimen. These savings may be enough to substantially influence the results of cost-effectiveness presented by Tunis et al. [1].

These issues suggest that the cost-effectiveness of a generic fail-first regimen evaluated over short duration should be viewed and interpreted cautiously.––Anirban Basu, PhD, Section of General Internal Medicine, Department of Medicine, Center for Health and Social Sciences, University of Chicago, and Decision and Information Sciences Division, Argonne National Laboratories, Chicago, IL, USA.

Reference
1 Tunis SL. Faries DE, Nyhuis AW, et al. Cost-effectiveness of olanzapine as first-line treatment for schizophrenia: results from a randomized, open-label, 1-year trial. Value Health 2006;9:77–89.