Abstract:

Objective To construct a nutritional risk prediction model for elderly inpatients with mental disorders,and to provide a basis for clinical nutritional status assessment. Methods Using the method of cluster sampling,584 elderly patients with mental disorders who were hospitalized in Beijing Huilongguan Hospital from June 2020 to June 2021 were divided into a good nutritional status group（152 cases） and a malnutrition risk group（129 cases） and a malnutrition group（303 cases）. The general condition questionnaire,frailty screening scale （FP）,mini-nutrition assessment scale（MNA）,simplified physical performance scale（SPPB）,the Barthel Index（BI） of daily life and Family Care Index（APGAR） for ability evaluation were used for investigation. Logistic regression analysis and Graphpad Prism software were used to draw receiver operating characteristic curve to test the model fit and prediction effect. Results The incidence of malnutrition in elderly hospitalized patients with mental disorders was 51.88%,and advanced age（OR=1.049）,frailty（OR=2.578）,and BMI（OR=0.629） were independent risk factors for malnutrition risks（P<0.534）. 0.05）,in which advanced age was the greatest risk factor for malnutrition. Advanced age（OR=1.031）,diagnosis（OR=2.150）,frailty（OR=2.485）,and physical dysfunction（OR=2.123） were independent risks of malnutrition（P<0.05）,among which frailty is the greatest risk factor for malnutrition risk to be transformed into malnutrition. The area under the receiver operating characteristic curve（AUC） of the malnutrition risk is 0.732;the Youden index is 0.388;the optimal critical value is 0.473;the sensitivity of the prediction model is 0.736;the specificity is 0.651. Regarding the ROC curve of malnutrition,AUC is 0.808;Youden index is 0.531;the best critical value is 0.581;the sensitivity of the prediction model is 0.865;the specificity is 0.667. Those indicate that the 2 prediction models are accurate and reliable. Conclusion The risk prediction model of nutritional status of hospitalized elderly patients with mental disorders constructed in this study has high accuracy,and provides a basis for risk assessment and prevention of nutritional status of elderly patients with mental disorders in hospital.

Key words: Mental Disorders, Nutritional Status, Predictive Model, Aged, Inpatients, Nursing Care

摘要：

目的 构建老年住院精神障碍患者营养不良风险预测模型,为临床营养状况评估提供依据。 方法 采用便利抽样法,将2020年6月—2021年6月在北京市某精神疾病医院住院的584例老年精神障碍患者分为无营养不良组（152例）、有营养不良风险组（129例）、营养不良组（303例）3组,采用一般资料调查表、衰弱筛查量表、微型营养评估量表、简易体能状况量表、Barthel指数评定量表、家庭关怀度指数测评量表对患者进行调查,采用Logistic回归分析及Graphpad Prism软件绘制受试者操作特征曲线,检验模型拟合度及预测效果。 结果 老年住院精神障碍患者营养不良发生率为51.88%,高龄（OR=1.049）、衰弱（OR=2.578）、BMI（OR=0.629）是有营养不良风险的独立危险因素（P<0.05）;高龄（OR=1.031）、诊断（OR=2.150）、衰弱（OR=2.485）、肌体功能障碍（OR=2.123）是营养不良的独立危险因素（P<0.05）,衰弱是有营养不良风险转化为营养不良的最大危险因素;有营养不良风险的受试者操作特征曲线下面积为0.732,约登指数为0.388,最佳临界值为0.473,预测模型的灵敏度为0.736,特异度为0.651;营养不良的受试者操作特征曲线下面积为0.808,约登指数为0.531,最佳临界值为0.581,预测模型的灵敏度为0.865,特异度为0.667,显示两个预测模型准确可靠。 结论 该研究构建的住院老年精神障碍患者营养风险预测模型具有较高的准确性,可为老年住院精神障碍患者营养状况风险评估及防治提供依据。

关键词: 精神障碍, 营养状况, 预测模型, 老年人, 住院患者, 护理