Anderson Jeffrey P, Parikh Jignesh R, Shenfeld Daniel K, Ivanov Vladimir, Marks Casey, Church Bruce W, Laramie Jason M, Mardekian Jack, Piper Beth Anne, Willke Richard J, Rublee Dale A
GNS Healthcare, Cambridge, MA, USA
GNS Healthcare, Cambridge, MA, USA.
J Diabetes Sci Technol. 2015 Dec 20;10(1):6-18. doi: 10.1177/1932296815620200.
Application of novel machine learning approaches to electronic health record (EHR) data could provide valuable insights into disease processes. We utilized this approach to build predictive models for progression to prediabetes and type 2 diabetes (T2D).
Using a novel analytical platform (Reverse Engineering and Forward Simulation [REFS]), we built prediction model ensembles for progression to prediabetes or T2D from an aggregated EHR data sample. REFS relies on a Bayesian scoring algorithm to explore a wide model space, and outputs a distribution of risk estimates from an ensemble of prediction models. We retrospectively followed 24 331 adults for transitions to prediabetes or T2D, 2007-2012. Accuracy of prediction models was assessed using an area under the curve (AUC) statistic, and validated in an independent data set.
Our primary ensemble of models accurately predicted progression to T2D (AUC = 0.76), and was validated out of sample (AUC = 0.78). Models of progression to T2D consisted primarily of established risk factors (blood glucose, blood pressure, triglycerides, hypertension, lipid disorders, socioeconomic factors), whereas models of progression to prediabetes included novel factors (high-density lipoprotein, alanine aminotransferase, C-reactive protein, body temperature; AUC = 0.70).
We constructed accurate prediction models from EHR data using a hypothesis-free machine learning approach. Identification of established risk factors for T2D serves as proof of concept for this analytical approach, while novel factors selected by REFS represent emerging areas of T2D research. This methodology has potentially valuable downstream applications to personalized medicine and clinical research.
将新型机器学习方法应用于电子健康记录（EHR）数据可为疾病进程提供有价值的见解。我们利用这种方法构建了预测模型，以预测糖尿病前期和2型糖尿病（T2D）的进展情况。
我们使用一个新型分析平台（逆向工程与正向模拟[REFS]），从汇总的EHR数据样本中构建了糖尿病前期或T2D进展的预测模型集成。REFS依靠贝叶斯评分算法来探索广阔的模型空间，并从预测模型集成中输出风险估计分布。我们对24331名成年人进行了回顾性随访，观察他们在2007年至2012年间向糖尿病前期或T2D的转变情况。使用曲线下面积（AUC）统计量评估预测模型的准确性，并在独立数据集中进行验证。
我们的主要模型集成准确预测了T2D的进展（AUC = 0.76），并在样本外得到验证（AUC = 0.78）。T2D进展模型主要由既定风险因素（血糖、血压、甘油三酯、高血压、脂质紊乱、社会经济因素）组成，而糖尿病前期进展模型则包括新因素（高密度脂蛋白、丙氨酸转氨酶、C反应蛋白、体温；AUC = 0.70）。
我们使用无假设机器学习方法从EHR数据构建了准确的预测模型。确定T2D的既定风险因素是这种分析方法的概念验证，而REFS选择的新因素代表了T2D研究的新兴领域。这种方法在个性化医疗和临床研究方面具有潜在的重要下游应用。
