長者功能狀態對死亡之影響

Abstract

研究背景 人口高齡化為全球普遍趨勢，意味著長者照護負擔與醫療資源分配，以及對老年人行動、孤立和心理健康問題的關注。2020年COVID-19疫情爆發，長者多免疫系統低下、器官功能衰退並伴隨慢性病等健康問題，易於感染COVID-19後出現不良的預後，嚴重者甚至導致死亡。 研究目的 本研究為探討COVID-19疫情期間、65歲以上社區長者身體功能狀態、衰弱、運動能力、骨質密度、失能程度、憂鬱與認知程度與死亡之相關性。 材料與方法 研究設計為回溯性世代追蹤研究，以2019年至2021年4月設籍在彰化縣且參與高齡者整合式健康檢查及體適能檢測之65歲以上長者為研究對象，追蹤其至2023年12月31日止.事件為全死因的發生。使用高齡者整合式健康檢查、社區據點體適能檢測、骨質密度巡迴篩檢與傳染病通報系統中COVID-19感染資料整合後進行多元邏輯式迴歸和Cox比例風險迴歸存活分析，探討各功能變項與死亡的關聯性，並應用機器學習中的隨機森林，探討各類變項對於死亡重耍順序綜合性影響。 結果 本研究納入3,214位65歲以上社區長者，其中死亡人數佔6.44%(N=207)。由多元邏輯式迴歸結果發現，男性死亡風險高於女性 (adjusted odds ratio (aOR) =2.48, 95%信賴區間=1.58-3.89)，年齡大、BMI過輕、與看護同住或住機構、目前有吸菸者或已戒酒之生活習慣者的長者有較高的死亡機率，而曾感染COVID-19的長者死亡風險較未曾感染者低(aOR=0.54, 95%信賴區間=0.37-0.78)。個人病史則皆未與死亡有顯著相關。另外，在各類功能狀態中，僅慣用手握力異常與死亡顯著相關，死亡機率增加64% (aOR=1.64，aOR 95%信賴區間=1.09-2.45)。 由Cox比例風險多元迴歸存活時間結果發現，在基本人口學與生活型態方面，呈現與未納入時間邏輯式迴歸分析時相似的結果，而有糖尿病、腎臟病或心臟病病史的長者有較高的死亡風險。另外，在各類功能狀態中，SOF功能衰弱前期的長者死亡風險增加61%( aHR =1.61，aHR 95%信賴區間=1.17-2.21)，運動能力中未能於12秒內完成坐站5次、慣用手握力異常或6公尺步行速度慢者，死亡風險較高，分別增加39%(aHR =1.39、aHR 95%信賴區間=1.01-1.90)、60%(aHR =1.60、aHR 95%信賴區間=1.09-2.35)、97%(aHR =1.97、aHR 95%信賴區間=1.29-3.02)。中重度失能或認知異常的長者，死亡風險分別增加82%(aHR =1.82、aHR 95%信賴區間=1.21-2.74)、55%( aHR =1.55，aHR 95%信賴區間=1.02-2.35)。 運用機器學習方式找出影響死亡的重要因子，由邏輯式迴歸與隨機森林模型中發現預測死亡事件是否發生的共同影響因子為年齡、BMI與6公尺步行速度。另在邏輯式迴歸模型中，性別、COVID-19感染、糖尿病、心臟病與慣用手握力也是顯著影響死亡的重要變項，而在隨機森林模型中，顯著影響死亡的重要變項則還包含失能狀態、社會心理變項B、坐站5次與居住狀況。 最後將重要變項變項放入Cox比例風險迴歸模型預測個人存活機率及個人特徵分類。 結論 利用社區多元資料在COVID-19疫情期間，找出影響65歲以上社區長者死亡最重要的因素為性別、年齡、BMI、個人病史、運動能力、失能狀態與社會心理變項B，未來可以從這些面向，進行長者健康促進活動或長者照護服務政策在面對疫新興傳染病疫情曝發的規劃及參考。

Background The aging population is a global trend, indicating increased burdens in elderly care, allocation of medical resources, and concerns regarding mobility, isolation, and mental health issues among the elderly. The outbreak of COVID-19 in 2020 exacerbated these concerns, as the elderly often have weakened immune systems, declining organ functions, and underlying health conditions, making them more susceptible to adverse outcomes after contracting COVID-19, with severe cases even leading to death. Research Purpose This study aims to investigate the associbations between the physical functioning, frailty, locomotor, bone density, disability, depression, cognitive status, and mortality among community-dwelling elderly aged 65 and above in Changhua, Taiwan, under the scenario of the COVID-19 pandemic. Materials and Methods This study design is a retrospective cohort study that enrolled individuals aged 65 and older who were residents of Changhua from 2019 to April 2021 and participated in integrated health examinations and physical fitness tests for the elderly.The study aims to track mortality events among the participants until December 31, 2023.Data from integrated health examinations for the elderly, community-based fitness tests, bone density screenings, and COVID-19 infection records from the Infectious Disease Notification System,were collected and linked to mortality data from the household registration system.Multiple logistic regression and Cox proportional hazards regression were used to explore associations between various functional status and mortality.Additionally, random forest, a machine learning technique, was applied to examine the ranking of important variables associated with mortality. Results This study included 3,214 community elders over 65 years old, of whom 6.44% (N=207) died during the study period.Results from multiple logistic regression revealed that males had higher mortality than females (aOR=2.48, 95%C.I =1.58-3.89), while older age, living with caregivers or in institutions, and current smokers or former alcohol consumers were associated with higher mortality rates. The elderly who had been infected with COVID-19 had lower mortality compared to those who hadn't (aOR=0.54, 95%C.I = 0.37-0.78).Personal medical histories didn’t show significant associations with mortality.However, among various functional status, only abnormal grip strength in the dominant hand was significantly associated with a 64% increased risk of death (aOR=1.64, aOR 95% CI=1.09-2.45). Results from multiple Cox proportional hazards regression, adding survival time, showed similar associations in demographic and lifestyle factors as seen in the analysis without time inclusion.The elderly with histories of diabetes, kidney disease, or heart disease had higher risks of mortality.Additionally, among various functional status, pre-frailty assessed by the SOF criteria was associated with a 61% increased risk of death (aHR 95% CI=1.17-2.21).Those unable to complete five chair stands within 12 seconds, with abnormal grip strength, or with slow walking speed over 6 meters had increased risks of death by 39%(aHR=1.39,aHR 95% CI=1.01-1.90),60% (aHR=1.60,aHR 95% CI=1.09-2.35),and 97%(aHR=1.97,aHR 95% CI=1.29-3.02)respectively.The elderly with moderate to severe disability or cognitive impairment faced increased risks of death by 82% (aHR=1.82,aHR 95% CI=1.21-2.74) and 55%(aHR=1.55, aHR 95% CI=1.02-2.35) respectively. Using machine learning methods to find important factors affecting death.Logistic regression and random forest models revealed that age, BMI, and 6-meter walking speed were consistently important predictors of mortality.In logistic regression models, gender, COVID-19 infection, diabetes, heart disease, and grip strength were significant predictors of death. In random forest models, additional significant predictors included disability status, frailty , ability to complete five chair stands, and living arrangements. The classification of personal features in the light of individualized survival probability pursuant to the Cox regression model was further performed. Conclusion Leveraging multifarious community-based data with the COVID-19 pandemic scenario, the present study identified significant factors influencing mortality among community-dwelling elderly aged 65 and above including gender, age, BMI, personal medical history, physical fitness, disability status, and frailty . All these findings can serve as focal points for developing elderly health promotion activities or policies for eldercare services when facing the oiutbreaks of emergiang infectious disease.