- © Copyright??
- © .
In the March issue of Annals of Oncology, Fossati et al. presented what they called indirect evidence for the existence of a wish bias in oncology trials that may lead to overestimation of treatment results when a drug is being considered an innovation. The authors observed a decrease in the overall response rate and partial response rate to treatments of metastatic breast cancer, including doxorubicin, over time. They explained this finding with the conscious or unconscious overestimation of treatment effects by non-blinded investigators when efficacy is assessed using end points exposed to some subjectivity. The authors did not, however, consider the, at least theoretical, possibility of confounding through a shift over time towards patients with worse prognosis being included in the trials. The first trials could have concentrated on highly selected patient groups with a better prognosis. With widespread use of chemotherapy, trials conducted later would have progressively opened to less selected patients, thus including also women with a worse prognosis. If this had happened, one would expect to observe a progressive decrease in the overall response rates not only in the doxorubicin arm, but also in the group treated with the alternative drugs, whichever these were.