In the past decade, the cancer research community has made considerable progress in characterizing the genomic features of human tumors. Knowledge of molecular drivers of cancer has therefore increased greatly. Although the oncology community hoped that this would lead to more effective therapies, our ability to translate laboratory cancer research into clinical success has been remarkably low ? only 5% of the agents demonstrated to have anticancer activity in laboratory studies went on to achieve success in phase III clinical trials. Many factors are responsible for this disturbingly high percentage of failures, including misuse or misunderstanding of statistical design and analysis concepts, interpretation and reporting methods. Although correctable, these issues are widely prevalent in the cancer biomedical research enterprise. Another concern is that statistics training is not a required component of graduate programs in biomedical sciences. Recognizing the seriousness of these issues, the National Institutes of Health has called for the full engagement of the entire biomedical research enterprise to implement the resources needed to improve and sustain the statistical rigor and successful translation of preclinical cancer research. To this end, we propose to develop a statistics curriculum for early-career preclinical cancer researchers (i.e., postdoctoral researchers) conducting laboratory research at Memorial Sloan Kettering Cancer Center as well as the broader research community. Our proposed curriculum will be delivered over a 10-week period with one 90-minute class per week in a flipped classroom format, where lecture notes will be disseminated one week prior to each classroom contact and classroom time will be dedicated to hands-on activities with experimental designs, data analysis and participatory dialogues. We will develop a plan for evaluating our proposed curriculum and make all the resources available to the broader statistics and cancer research communities.
Statistically rigorous laboratory experiments are essential for developing successful treatments for human cancers. However, statistics training is not a required component of graduate programs in biomedical sciences. We are developing a statistics curriculum to equip early-career cancer researchers with the knowledge and tools required to execute robust preclinical experiments by providing training using hands-on activities with experimental designs, data analysis and participatory dialogues.
