arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2501.09156stat.APstat.ML
收藏

使用贝叶斯机器学习对青少年和成年早期大麻使用障碍的绝对风险预测

Absolute Risk Prediction for Cannabis Use Disorder in Adolescence and Early Adulthood Using Bayesian Machine Learning

Tingfang Wang, Joseph M. Boden, Swati Biswas, Pankaj K. Choudhary

AI总结:

本研究开发了一种基于贝叶斯机器学习的大麻使用障碍绝对风险预测模型,通过5个风险因素实现个性化预测,在训练和验证数据集中表现出良好的区分度与校准性能,可辅助临床医生进行早期干预。

摘要:

引言:物质使用障碍(SUDs)已成为美国紧迫的公共卫生问题,青少年时期的物质使用往往会导致成年期的SUDs。需要有效的策略来遏制这一进展。为帮助满足这一需求,我们为使用大麻的青少年或年轻成年人开发了一种新型的大麻使用障碍(CUD)绝对风险预测模型。方法:我们训练了一个贝叶斯机器学习模型,利用来自National Longitudinal Study of Adolescent to Adult Health的数据,为使用大麻的青少年或年轻成年人提供个性化的CUD绝对风险。使用5折交叉验证(CV)评估模型性能,指标包括曲线下面积(AUC)和预期病例数与观察病例数之比(E/O)。使用两个数据集对最终模型进行了独立验证。结果:所提出的模型有五个风险因素:生物学性别、违规行为,以及尽责性、神经质和开放性人格特质得分。对于预测首次使用大麻后五年内的CUD风险,训练数据集和两个验证数据集的AUC值分别为0.68、0.64和0.75,E/O值分别为0.95、0.98和1。这表明模型具有良好的区分度和校准性能。讨论与结论:所提出的模型可帮助临床医生评估使用大麻的青少年和年轻成年人患CUD的风险,从而实现临床适当的干预。

英文摘要:

Introduction: Substance use disorders (SUDs) have emerged as a pressing public health concern in the United States, with adolescent substance use often leading to SUDs in adulthood. Effective strategies are needed to stem this progression. To help fulfill this need, we developed a novel absolute risk prediction model for cannabis use disorder (CUD) for adolescents or young adults who use cannabis. Methods: We trained a Bayesian machine learning model that provides a personalized CUD absolute risk for adolescents or young adults who use cannabis with data from the National Longitudinal Study of Adolescent to Adult Health. Model performance was assessed using 5-fold cross-validation (CV) with area under the curve (AUC) and ratio of the expected to observed number of cases (E/O). Independent validation of the final model was conducted using two datasets. Results: The proposed model has five risk factors: biological sex, delinquency, and scores on personality traits of conscientiousness, neuroticism, and openness. For predicting CUD risk within five years of first cannabis use, AUC values for the training dataset and two validation datasets were 0.68, 0.64, and 0.75, respectively, and E/O values were 0.95, 0.98, and 1, respectively. This indicates good discrimination and calibration performance of the model. Discussion and Conclusion: The proposed model can aid clinicians in assessing the risk of developing CUD among adolescents and young adults who use cannabis, enabling clinically appropriate interventions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑