25.11.2011: Application of propensity score based matching in oncology
Today Metronomia has delivered the statistical analysis and report for an indirect comparison of data from a single arm phase II trial in oncology (CRC) with data from a large dataset collected by a study group.
Single arm trials are still very common in oncology, particularly in phase II, and there are cases where such trials have been the basis for marketing approval, although randomized, controlled clinical trials are undoubtedly the preferred option for pivotal clinical trials.
Indirect comparisons, i.e. comparing data from one trial with data from another trial, cannot overcome the limitations when no controlled clinical trial is available to compare treatments. However, rather than just discussing summary statistics of clinical trials taken from publications a propensity score based matching of patient records from different trials has many advantages and can promote the acceptance of such indirect comparisons.
Metronomia has implemented in the named project a methodology which allowed a blinded matching of patient records from two different clinical trials based on propensity scores. Finally, very well comparable baseline characteristics of the two treatment were achieved and efficacy data could be compared with statistical methods especially developed for matched data.
Another well-known area of application of indirect comparisons based on propensity scores is outcomes research and analysis for reimbursement (e.g. for AMNOG dossiers).