Vanderbilt University Medical Center, Nashville, TN, USA.
Florida State University, Tallahassee, FL, USA.
J Child Psychol Psychiatry. 2018 Dec;59(12):1261-1270. doi: 10.1111/jcpp.12916. Epub 2018 Apr 30.
Adolescents have high rates of nonfatal suicide attempts, but clinically practical risk prediction remains a challenge. Screening can be time consuming to implement at scale, if it is done at all. Computational algorithms may predict suicide risk using only routinely collected clinical data. We used a machine learning approach validated on longitudinal clinical data in adults to address this challenge in adolescents.
This is a retrospective, longitudinal cohort study. Data were collected from the Vanderbilt Synthetic Derivative from January 1998 to December 2015 and included 974 adolescents with nonfatal suicide attempts and multiple control comparisons: 496 adolescents with other self-injury (OSI), 7,059 adolescents with depressive symptoms, and 25,081 adolescent general hospital controls. Candidate predictors included diagnostic, demographic, medication, and socioeconomic factors. Outcome was determined by multiexpert review of electronic health records. Random forests were validated with optimism adjustment at multiple time points (from 1 week to 2 years). Recalibration was done via isotonic regression. Evaluation metrics included discrimination (AUC, sensitivity/specificity, precision/recall) and calibration (calibration plots, slope/intercept, Brier score).
Computational models performed well and did not require face-to-face screening. Performance improved as suicide attempts became more imminent. Discrimination was good in comparison with OSI controls (AUC = 0.83 [0.82-0.84] at 720 days; AUC = 0.85 [0.84-0.87] at 7 days) and depressed controls (AUC = 0.87 [95% CI 0.85-0.90] at 720 days; 0.90 [0.85-0.94] at 7 days) and best in comparison with general hospital controls (AUC 0.94 [0.92-0.96] at 720 days; 0.97 [0.95-0.98] at 7 days). Random forests significantly outperformed logistic regression in every comparison. Recalibration improved performance as much as ninefold - clinical recommendations with poorly calibrated predictions can lead to decision errors.
Machine learning on longitudinal clinical data may provide a scalable approach to broaden screening for risk of nonfatal suicide attempts in adolescents.
青少年有很高的非致命性自杀企图率，但临床上实用的风险预测仍然是一个挑战。如果要进行筛查，那么在规模上实施起来可能会很耗时。计算算法可以仅使用常规收集的临床数据来预测自杀风险。我们使用了一种在成人纵向临床数据上经过验证的机器学习方法来解决青少年中的这一挑战。
这是一项回顾性的、纵向的队列研究。数据来自 1998 年 1 月至 2015 年 12 月的范德比尔特综合衍生数据，包括 974 名有非致命性自杀企图的青少年和多个对照比较：496 名有其他自伤（OSI）的青少年，7059 名有抑郁症状的青少年和 25081 名青少年普通医院对照。候选预测因子包括诊断、人口统计学、药物和社会经济因素。结果由电子病历的多专家审查确定。随机森林在多个时间点（从 1 周到 2 年）进行了乐观调整验证。通过等渗回归进行了重新校准。评估指标包括判别（AUC、灵敏度/特异性、精度/召回率）和校准（校准图、斜率/截距、Brier 分数）。
计算模型表现良好，不需要面对面的筛查。随着自杀企图变得越来越迫在眉睫，性能得到了提高。与 OSI 对照组（720 天 AUC=0.83 [0.82-0.84]；7 天 AUC=0.85 [0.84-0.87]）和抑郁对照组（720 天 AUC=0.87 [95%CI 0.85-0.90]；7 天 AUC=0.90 [0.85-0.94]）相比，表现良好，与普通医院对照组（720 天 AUC 为 0.94 [0.92-0.96]；7 天 AUC 为 0.97 [0.95-0.98]）相比表现最佳。随机森林在所有比较中均显著优于逻辑回归。重新校准将性能提高了近九倍——预测校准不良的临床建议可能导致决策错误。
基于纵向临床数据的机器学习可能为扩大青少年非致命性自杀企图风险筛查提供一种可扩展的方法。
