Seo Ju-Hyun, Kim Hyun-Ji, Lee Jea-Young
Department of Statistics, Yeungnam University, Gyeongsan, Korea.
J Appl Stat. 2019 Sep 4;47(5):914-926. doi: 10.1080/02664763.2019.1660760. eCollection 2020.
Dyslipidemia is a chronic disease requiring continuous management and is a well-known risk factor for cardiovascular diseases as well as hypertension and diabetes. However, no studies have so far visualized and predicted the probability of dyslipidemia. Hence, this study proposes a nomogram based on a logistic regression model that can visualize its risk factors and predict the probability of developing dyslipidemia. Twelve risk factors for dyslipidemia are identified through a chi-squared test. We then conduct a logistic regression analysis with two interaction variables to obtain a model and build a nomogram for dyslipidemia. Finally, we verify the constructed nomogram using a receiver operation characteristic curve and calibration plot.
血脂异常是一种需要持续管理的慢性疾病，是心血管疾病以及高血压和糖尿病的众所周知的危险因素。然而，迄今为止尚无研究对血脂异常的可能性进行可视化和预测。因此，本研究提出了一种基于逻辑回归模型的列线图，该列线图可以可视化其危险因素并预测发生血脂异常的可能性。通过卡方检验确定了十二个血脂异常危险因素。然后，我们对两个交互变量进行逻辑回归分析以获得一个模型，并构建血脂异常的列线图。最后，我们使用受试者工作特征曲线和校准图验证构建的列线图。
