School of Medicine, Radiation Oncology, Johns Hopkins University, Baltimore, MD, 21231, USA.
Epidemiology, University of Chicago, Chicago, IL, 60637, USA.
Med Phys. 2018 Oct;45(10):e863-e869. doi: 10.1002/mp.12817. Epub 2018 Aug 24.
The capture of high-quality treatment data and outcomes is necessary in order to learn from our clinical experiences with big data analytics. In radiotherapy, there are several practical challenges to overcome. Practical aspects of data collection are discussed pointing to a need for a culture change in clinical practice to one that captures structured patient-related data in routine care in a prospective manner. Radiation dosimetry and the contoured anatomy must also be captured routinely to represent the best estimate of delivered radiation. The quality and integrity present in the data are critical which poses opportunities to introduce electronic validity checking to improve them. Similarly, data completeness and methods and technology to improve the efficiency and sufficiency of data capture can be introduced. In the manuscript, the types of clinical data are discussed including patient reports, images, biospecimens, treatments, and symptom management. With a data-driven culture, the realization of a learning health system is possible unlocking the potential of big data and its influence on clinical decision-making and hypothesis generation.
为了从大数据分析的临床经验中学习，有必要获取高质量的治疗数据和结果。在放射治疗中，有几个实际的挑战需要克服。本文讨论了数据收集的实际方面，指出需要在临床实践中进行文化转变，以一种前瞻性的方式在常规护理中采集与患者相关的结构化数据。还必须常规采集辐射剂量学和勾画的解剖结构，以代表所提供辐射的最佳估计。数据的质量和完整性至关重要，这为引入电子有效性检查以提高数据质量提供了机会。同样，可以引入数据的完整性以及提高数据采集效率和充分性的方法和技术。本文讨论了包括患者报告、图像、生物标本、治疗和症状管理在内的临床数据类型。通过数据驱动的文化，可以实现学习型健康系统，释放大数据的潜力及其对临床决策和假设生成的影响。
