ABSTRACT
Outcome measures in oncology are notoriously challenging. Meaningful endpoints often take many years to acquire and are often confounded by disease progression and other effects. Contrast this with other areas of health care where patient safety and quality improvement efforts have shown an impact on outcomes. Examples include central line acquired bloodstream infection (CLABSI) 1 and postoperative morbidity and mortality. 2 In these contexts relatively acute endpoints are readily available. Nevertheless, there are data in the realm of radiation oncology that tie the parameters of treatment to quality and safety. Perhaps the best examples are dose–response data for toxicities of normal tissues. Many studies are available and summaries appear from the quantitative analysis of normal tissue effects in the clinic (QUANTEC) study, 3 the American Association of Physicists in Medicine (AAPM) Working Group on Biological Effects of Hypofractionated Radiotherapy/Stereotactic Body Radiation Therapy (SBRT), 4 and other studies. Although these results guide practice, they would not be considered as big data by most investigators. Typically only one dimension is considered (e.g., dose) and only a very few data points are evaluated per patient. Also problematic in these studies is the high variability in the quality of the radiation treatment that is actually delivered. Recent studies suggest that normal tissue doses that are achieved are highly variable between institutions and even within a single institution. 5,6
