Benci Joseph L, Vachani Carolyn C, Hampshire Margaret K, Bach Christina, Arnold-Korzeniowski Karen, Metz James M, Hill-Kayser Christine E
Department of Radiation Oncology, Perelman Center for Advanced Medicine, University of Pennsylvania, Philadelphia, PA, United States.
Front Oncol. 2020 Jan 31;9:1577. doi: 10.3389/fonc.2019.01577. eCollection 2019.
Nearly half of all Americans will develop cancer at least once in their lifetime. Through improved screening and treatments, the number of cancer survivors is reaching all-time highs. However, survivorship care plans (SCPs) are inconsistently used, denying many survivors access to critical information. This study used 46,408 SCPs generated from 2007 to 2016 and applied machine learning to identify predictors of SCP creation, including cancer type, type of physician, and healthcare center where they received care, as well as regional variations in care plan creation. Identifying these disparities in SCP use is a critical first step in efforts toward expanding access to survivorship care planning. Using a convenience sample of survivors, it is possible to model the factors that predict generation of SCPs either by the survivor or by a healthcare provider. This study identifies several important disparities both survivor intrinsic such as cancer type, as well as treatment associated and geographic differences in SCP generation. Identifying these disparities at the national level across cancer types will allow for more targeted recommendations to improve SCP creation and dissemination in underserved groups.
近一半的美国人一生中至少会患一次癌症。通过改进筛查和治疗,癌症幸存者的数量达到了历史最高水平。然而,生存护理计划(SCPs)的使用并不一致,导致许多幸存者无法获得关键信息。本研究使用了2007年至2016年生成的46408份SCP,并应用机器学习来识别SCP创建的预测因素,包括癌症类型、医生类型以及他们接受治疗的医疗中心,以及护理计划创建的地区差异。识别SCP使用中的这些差异是扩大生存护理计划获取机会的关键第一步。使用幸存者的便利样本,可以对预测SCP由幸存者或医疗服务提供者生成的因素进行建模。本研究确定了几个重要的差异,包括幸存者内在因素如癌症类型,以及SCP生成中的治疗相关和地理差异。在全国范围内识别不同癌症类型的这些差异,将有助于提出更有针对性的建议,以改善SCP在服务不足群体中的创建和传播。