肿瘤预防与治疗2025，Vol.38Issue(12)：1060-1069,10.DOI:10.3969/j.issn.1674-0904.2025.12.004
晚期癌症患者姑息需求潜在类别及预测模型构建
Latent Class Analysis of Palliative Care Needs in Advanced Cancer Pa-tients to Build a Prediction Model
摘要
Abstract
Objective:To explore the categories of palliative care needs in patients with advanced cancer and to develop a nomogram for prediction and validation.Methods:A prospective study was conducted involving 414 patients with advanced cancer from the oncology departments of two tertiary hospitals in Zhenjiang,China,between May and October 2023.Data were collected using a general information questionnaire,the Short-Form of the Problems and Needs in Palliative Care questionnaire,the Numerical Rating Scale for pain,the Pittsburgh Sleep Quality Index,the Self Rating Anxiety Scale,the Patient Health Questionnaire-9 and the Barthel Index.Latent profile analysis was employed to categorize the palliative care needs of patients with advanced cancer.The dataset was randomly allocated into a training and a validation set in a 7:3 ratio.We performed LASSO regression and Logistic regression to identify factors associated with different palliative care needs in the training set and constructed a nomogram model.This model was validated in the validation set,with its discrimi-nation and calibration evaluated by the receiver operating characteristic(ROC)curve and the Hosmer-Lemeshow goodness-of-fit test.Results:The palliative care needs of patients with advanced cancer were classified into two latent categories:a high-need group(n=118)and a low-need group(n=296)(Entropy=0.902,Lo-Mendell-Rubin likelihood ratio test<0.001,Bootstrap likelihood ratio test<0.001).The LASSO and logistic regression analyses identified the following as inde-pendent factors for high palliative care needs:gender(χ2=7.382,P=0.007),age(χ2=28.387,P<0.001),marital status(χ2=12.664,P<0.001),literacy(χ2=9.327,P=0.009),family harmony(χ2=29.433,P<0.001),sleep quality(χ2=23.242,P<0.001),and self-care ability(χ2=8.826,P=0.012).Based on the multivariable analysis,we constructed a nomogram that showed strong discriminatory power(training set AUC:0.883,95%CI:0.835～0.931;valida-tion set AUC:0.884,95%CI:0.824～0.945)and good calibration(Hosmer-Lemeshow test P=0.615 and P=0.982 for training and validation sets,respectively).Conclusion:The predictive model enables healthcare professionals to rapidly iden-tify advanced cancer patients with high palliative care needs,thereby facilitating early clinical intervention.关键词
晚期癌症/姑息需求/潜在剖面分析/预测模型Key words
Advanced cancer/Palliative care needs/Latent profile analysis/Predictive model分类
医药卫生引用本文复制引用
Sun Mengdi,Zhang Wei,Zhao Ruobing,Chen Xing,Liu Xiyang,Chen Chen..晚期癌症患者姑息需求潜在类别及预测模型构建[J].肿瘤预防与治疗,2025,38(12):1060-1069,10.基金项目
This study was supported by grants form Jiangsu Education Department(No.2906)and Zhenjiang Science and Technology Bureau(No.RK2022021). 2023年江苏省研究生实践创新计划基金项目(编号:2096) （No.2906）
2022年度镇江市政策引导计划(软科学研究)项目(编号:RK2022021) （软科学研究）
