Test studies of new cancer drugs often insufficient
Clinical trials designed to determine the potential of new cancer drugs are often poorly planned and sometimes improperly executed, researchers at the Memorial Sloan-Kettering Cancer Center have complained. The scientists retrospectively analyzed published phase II studies from two years ago. There were often deficiencies in the formulation of an appropriate study objective and in the statistical methodology.
The researchers led by Andrew Vickers wanted to find out why a striking number of new drugs, which were initially classified as potentially effective in the second phase of the usual clinical evaluation procedure, fail in the subsequent complex third test phase. According to the authors, the reason for this may often be an incorrect assessment of efficacy due to poor design of phase II clinical trials.
In this second test phase, new drugs are tested on a small group of sick patients. In order to evaluate the effectiveness of new drugs, a medical goal must be formulated in the tests - such as reducing the size of tumors in patients by a statistically measurable value within a specific time frame. The doctors can then tell whether this goal has been achieved by comparing the average effectiveness of the existing drugs with the treatment successes determined in the study in a group of patients who are given the new drug candidates in addition to the standard therapy.
Vickers and colleagues now complain that in 70 studies out of 134 published studies from the years 2003 to 2005 that were examined more closely, there was only insufficient documentation of how the basic effectiveness of the already existing standard therapy, which was decisive for the assessment, was determined. For example, it was not possible to determine whether the effect determined in earlier studies had been obtained in a patient group that was not at all comparable to the group examined in the current test. Even in the 9 of the 70 studies that provided sufficient source and method information for the compared standard cohorts, statistical methods that can compensate for differences in the composition of the previous and current test groups were missing.
The deficiencies could lead to the efficacy of a tested substance being overestimated in phase II tests, according to the authors. It is particularly striking that studies in which references to the values used as standard effectiveness of a therapy are completely missing, come to the conclusion that the tested new drug justifies a subsequent phase III test. Vickers and colleagues believe that new guidelines for defining phase II targets and for conducting the studies are needed in order to avoid unnecessary follow-up tests in the future, in which an overly optimistic drug is then revealed as ineffective.(yo)
