髋部骨折患者术后健康相关生命质量下降的预测模型构建

Construction of a predictive model for postoperative health-related quality of life decline in hip fracture patients

目的评估髋部骨折患者术后健康相关生命质量(HRQoL)并构建预测模型,以识别HRQoL下降的高风险患者。方法应用横断面调查法,采用便利抽样方式选取海南省中医院2020年1月至2021年6月的135例单侧髋部骨折患者作为研究对象。使用欧洲五维健康量表(EQ-5D)自评HRQoL;采用Logistic回归模型确定HRQoL下降危险因素;使用校正曲线、ROC曲线、决策曲线分析(DCA)评估列线图的预测准确性和临床净收益。结果最终获得有效随访患者122例。术后6、12个月,41例占33.6%患者的EQ-5D指数均低于骨折前,与Barthel指数呈现较强相关性(r=0.833、0.705,均P<0.05)。营养风险筛查2002(NRS-2002)(OR=1.352)、Charlson合并症指数(CCI)(OR=1.121)、髋关节置换术(OR=0.795)和肺部感染(OR=1.328)与髋部骨折患者HRQoL下降独立相关(均P<0.05)。列线图的一致性指数为0.889(0.791~0.955),AUC为0.871(0.784~0.947)。DCA结果显示模型提供良好的临床净收益。结论该预测模型能够有效识别HRQoL下降的高风险患者,有助于确定提供护理策略的优先级患者。

Objective To evaluate the health related quality of life(HRQoL)of hip fracture patients after surgery and build a prediction model to identify high-risk patients with HRQoL decline.Methods A cross-sectional research method was applied to select 135 patients with unilateral hip fractures from January 2020 to June 2021 in Hainan Provincial Hospital of Traditional Chinese Medicine using convenience sampling as the study population.HRQoL was self-assessed using the European Five-Dimensional Health Scale(EQ-5D).Logistic regression models were used to identify risk factors for HRQoL decline.The columnar plots'predictive accuracy and net clinical benefit were assessed using calibration curves,ROC curves,and decision curve analysis(DCA).Results The final number of patients with valid follow-up was 122.At 6 and 12 months postoperatively,the EQ-5D index was lower than before the fracture in 41(33.6%)patients,showing a strong correlation with the Barthel index(r=0.833,0.705,both P<0.05).Nutritional Risk Screening 2002(NRS-2002)(OR=1.352),Charlson Comorbidity Index(CCI)(OR=1.121),hip replacement(OR=0.795),and lung infection(OR=1.328)were independently associated with a decrease in HRQoL in patients with hip fracture(all P<0.05).The consistency index(C-index)of the column line plot was 0.889(0.791 to 0.955),and the area under the curve(AUC)was 0.871(0.784 to 0.947).The DCA results showed that the model provided good net clinical benefit.Conclusions The model can effectively identify patients at high risk of declining HRQoL,which is helpful to determine priority patients to provide nursing strategies.