The somatic symptom clusters may be associated with increased risk of adverse outcomes in frail elderly people. Relevant studies in China have mainly adopted a cross-sectional design with neglect of the trajectory of somatic symptom clusters in this group.

To explore the characteristics of somatic symptom clusters at different time points and influencing factors in elderly people with frailty in nursing homes in Chengdu.

From November 2019 to January 2020, 206 frail elderly people were selected from 6 nursing homes in Chengdu by convenience sampling, and surveyed using the general data questionnaire and Memory Symptom Assessment Scale (MSAS) for 3 times〔at baseline (T0) , 6 (T1) , and 12 months later (T2) 〕. Exploratory factor analysis was carried out for symptoms with an incidence of ≥20% at different time points. Latent growth mixture model (LGMM) was used to identify the change trajectory of somatic symptom clusters across the above-mentioned three time points. Logistic regression analysis was used to identify the potential factors associated with the trajectory category.

By exploratory factor analysis, 5 factors were extracted at each of the three time points. Neurological symptom cluster, energy deficiency symptom cluster, respiratory symptom cluster and digestive symptom cluster all appeared at the three time points. In addition, senescence-related symptom cluster also occurred at T0 and T1, and other symptom cluster occurred at T2. The MSAS score of each symptom cluster differed significantly across three time points (P<0.05) . Four heterogeneous trajectories of frailty symptom clusters were obtained by LGMM model fitting, which were named as "high decline" "low rise" "medium maintenance" and "high rise", accounting for 16.5%, 12.5%, 66.0% and 5.0%, respectively. Multivariate Logistic regression analysis showed that the number of chronic diseases was independently associated with the "high decline" or "high rise" trajectory, and the number of medications was independently associated with the "high rise" trajectory (P<0.05) .

There are various trajectories of somatic symptom clusters in frail elderly people in nursing homes, and each of the trajectories has a different independently associated factor. To provide more appropriate services for this population, medical workers in nursing homes can dynamically adjust nursing services according to the trajectories and associated factors of somatic symptom clusters.

衰弱老年人的躯体症状群可增加不良结局风险，我国相关研究以横断面为主，忽视了对衰弱老年人躯体症状群变化轨迹的探讨。

探讨成都市养老机构衰弱老年人不同时间点躯体症状群的变化轨迹及其影响因素。

2019年11月至2020年1月，采用便利抽样法选取成都市6所养老机构的206例衰弱老年人作为研究对象，分别在基线（T0）、6个月（T1）、12个月（T2）时采用一般资料调查表、记忆症状评估量表（MSAS）对其进行调查。对不同时间点上发生率≥20%的症状进行探索性因子分析，使用潜变量增长混合模型（LGMM）识别养老机构衰弱老年人躯体症状群在3个时间点的变化轨迹，采用Logistic回归分析识别轨迹类别的影响因素。

探索性因子分析在3个时间点均分别提取了5个因子，3个时间点均有神经系统症状群、能量不足症状群、呼吸道症状群及消化道症状群，此外，T0和T1还包含衰老相关症状群，T2还包含其他症状群。3个时间点上各症状群MSAS得分比较，差异有统计学意义（P<0.05）。LGMM拟合得到"高下降""低上升""中维持""高上升"4条躯体衰弱症状异质性轨迹，4条轨迹占比分别为16.5%、12.5%、66.0%、5.0%。多因素Logistic回归分析结果显示，患慢性病数量多是"高下降"和"高上升"轨迹的独立预测因素，服药种类多是"高上升"轨迹的独立预测因素（P<0.05）。

养老机构衰弱老年人的躯体症状群存在不同变化轨迹，且不同轨迹的独立预测因素不相同，养老机构医务人员可根据其症状群变化轨迹及独立预测因素动态调整护理措施，为其提供更为精准的护理服务。