Confounding factors in clinical studies of gliomas.

Clinical investigation of gliomas is difficult because the investigator must compensate for factors that confound data analysis. The investigation of glioma therapy is affected by a number of factors that influence the outcome of the patients' disease: the tumor grade and location, the extent of surgical resection, the dose and portal for radiation therapy, and the patient's functional status at presentation. Either an observational or an experimental (clinical trial) strategy may be used to compensate for confounding factors. Matching for the characteristics known to influence disease history strengthens the conclusions from an observational study, but cannot compensate for unknown influences. Randomization in a clinical trial will minimize unknown influences, but clinical trials are not always feasible. To be conclusive, glioma studies should either match patient populations for known confounding factors in an observational study or randomize patients in a clinical trial.