基于实验室指标评估COVID-19患者病情严重程度并构建预测模型

Evaluating the Severity of COVID-19 Patients Based on Laboratory Indicators and Constructing a Prediction Model

摘要: 目的：探讨实验室指标及联合指标对新型冠状病毒感染(COVID-19)患者病情严重程度的预测价值并构建临床预测模型。方法：回顾性收集2022年12月~2024年12月就诊于某医院呼吸与危重症医学科，依据《新型冠状病毒感染诊疗方案(试行第十版)》诊断为COVID-19的患者共107例，将轻型和中型归为非重症组(60例)，重型和危重型归为重症组(47例)。收集患者的一般资料、实验室数据，分别通过单因素二元Logistic回归及多因素二元Logistic回归分析筛选出可影响COVID-19患者病情严重程度的指标，应用R4.3.3软件将在多因素分析中筛选出的指标构建风险评分列线图并采用受试者工作曲线、校准曲线和决策曲线评估模型的校准度和临床净获益。结果：重症组HALP指数低于非重症组、重症组CRP、BUN水平高于非重症组(P < 0.05)，基于三者构建的列线图预测COVID-19患者发生重症的ROC曲线下面积AUC值为0.886 (95% CI: 0.8226~0.949)，校准曲线及决策曲线进一步证实该模型具有较好的拟合度和临床价值。结论：依据HALP指数、CRP、BUN建立的预测模型效能良好，能较好地预测COVID-19患者病情的严重程度。

Abstract: Objective: To explore the predictive value of laboratory indicators and combined indicators for assessing the severity of COVID-19 patients and to construct a clinical predictive model. Methods: A retrospective study was conducted on 107 patients diagnosed with COVID-19 according to the Diagnosis and Treatment Protocol for Novel Coronavirus Infection (Trial Version 10). These patients were admitted to the Department of Respiratory and Critical Care Medicine at a hospital between December 2022 and December 2024. The patients were categorized into a non-severe group (60 cases, including mild and moderate cases) and a severe group (47 cases, including severe and critical cases). General patient data and laboratory results were collected. Univariate and multivariate binary logistic regression analyses were performed to identify indicators influencing disease severity. The selected indicators from the multivariate analysis were used to construct a risk assessment nomogram using R4.3.3 software. The model’s calibration and clinical utility were evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis. Results: The severe group exhibited a lower HALP (Hemoglobin, Albumin, Lymphocyte, Platelet) index and higher CRP (C-reactive protein) and BUN (blood urea nitrogen) levels compared to the non-severe group (P < 0.05). The nomogram constructed using these three indicators demonstrated an AUC (area under the ROC curve) of 0.886 (95% CI: 0.8226~0.949) for predicting severe COVID-19. Calibration and decision curves further confirmed the model's good fit and clinical value. Conclusion: The predictive model based on the HALP index, CRP, and BUN shows strong performance and can effectively predict the severity of COVID-19 patients.