照護失智患者，飲食能力量表發展與驗證

Development and validation of the scale of the ability to care for the diet of patients with dementia

目標：發展與驗證照護失智患者飲食能力量表之信效度，以評價家庭照顧者照護失智患者飲食的能力。方法：第一階段為量表題項建構，以立意取樣邀請29位失智家屬進行焦點團體、專家內容效度檢定，初擬失智飲食照護能力題項。第二階段為量表信效度檢測，以探索性及驗證性因素分析檢定量表建構效度，內在一致性檢定量表信度。運用橫斷式調查、方便取樣於臺灣北、中、南地區之失智症共同照護中心、失智症門診、社區失智症服務據點招募418位失智家屬為研究對象，運用IBM SPSS 26 Amos 27版程式進行資料分析。結果：6個焦點團體共12次訪談達資料飽和，初擬73個題項經專家效度檢定後修訂為50題，CVI值達0.97。經因素分析，量表第一部分「失智飲食照護概念」萃取出三因素8題，模式適配度指標GFI=0.97、CFI=0.97、TLI=0.95、AGFI=0.94、NFI=0.93、RMSEA=0.05、SRMR=0.04；8題SMC均高於0.25；除「因素三地中海食材選擇」CR值為0.49，其他皆高於0.6；「因素二營養與認知功能」AVE為0.4、「因素三地中海食材選擇」AVE為0.33，其餘題目AVE均高於0.5；三個因素的AVE皆高於其他兩兩因素的積差相關係數平方；信度KR-20值為0.70。第二部分「自覺失智飲食照護能力」萃取出四因素17題，模式適配度指標GFI=0.88、CFI=0.94、TLI=0.93、AGFI=0.83、NFI=0.92、RMSEA=0.08、SRMR=0.04；SMC值均高於0.25、CR值高於0.6、AVE值均高於0.5；四個因素的AVE皆高於其他兩兩因素的積差相關係數平方；信度Cronbach’sα值為0.94，量表兩部分SEM模式檢定達良好適配度且具收斂效度與區辨效度。結論：本量表具有良好信效度，可作為家庭照顧者其照護失智患者飲食能力的評估工具。（台灣衛誌2021；40(6)：660-679）

"Objectives: To develop and verify the reliability and validity of the Scale of the Ability to Care for the Diet of Patients with Dementia to assess the family caregivers’ability to care for the diet of such patients. Methods: The first stage involved constructing the items for the scale. Through purposive sampling, 29 family members of patients with dementia were invited to conduct focus group interviews and to conduct expert content validity verification, thereby drafting the items of the ability to care for the diet of patients with dementia. The second stage involved verifying the reliability and validity of the scale. Exploratory and confirmatory factor analyses were conducted to assess the construct validity of the scale. Internal consistency was used to assess the reliability of the scale. A cross-sectional survey was adopted. Through convenience sampling, 418 family members of patients with dementia from a center for integrated dementia care, dementia outpatient clinics, and community-based support sites for dementia care in northern, central, and southern Taiwan were recruited. IBM SPSS 26.0 and AMOS 27.0 were used for data analysis. Results: After conducting 12 interviews with 6 focus groups, data reached saturation. We drafted 73 items, which were revised into 50 items after expert validity verification, reaching a content validity index value of 0.97. After factor analysis, 3 factors and 8 items in the first part of the scale,“Dementia Diet Care Concepts,”were extracted. The model goodness of fit indices were as follows: goodness of fit index = 0.97, comparative fit index = 0.97, Tucker–Lewis index = 0.95, adjusted goodness of fit index = 0.94, normed fit index = 0.93, mean square error of approximation = 0.05, and standardized root mean squared residual = 0.04. The square multiple correlation values of all 8 items were greater than 0.25. With the exception of Factor 3 (Choosing Mediterranean ingredients), whose critical ratio (CR) value was 0.49, the CR values of all the other factors were greater than 0.6. The average variance extracted (AVE) of Factor 2 (Nutrient and cognitive function) was 0.4, and that of Factor 3 (Choosing Mediterranean ingredients) was 0.33. The AVE of the rest of the factors were greater than 0.5. The AVE of 3 factors were all greater than the squared value of the product–moment correlation coefficient of any 2 factors. The Kuder–Richardson Formula 20 reliability was 0.70. In the second part of the scale, Self-reported Ability of Dementia Diet Care, 4 factors and 17 items were extracted. The model goodness of fit indices were as follows: goodness of fit index = 0.88, comparative fit index = 0.94, Tucker–Lewis index = 0.93, adjusted goodness of fit index = 0.83, normed fit index = 0.92, mean square error of approximation = 0.08, and standardized root mean squared residual = 0.04. The square multiple correlation values of all items were greater than 0.25. The CR values were all greater than 0.6, and the AVE values were all greater than 0.5. The AVE of 4 factors were all greater than the squared value of the product–moment correlation coefficient of any 2 factors. The reliability Cronbach’sαwas 0.94. The structural equation modeling model verification of the 2 parts of the scale revealed favorable goodness of fit, and they exhibited convergent validity and discriminant validity. Conclusions: This scale had favorable reliability and validity, and it can be used as an assessment tool for assessing family caregivers’ability to care for the diet of patients with dementia. (Taiwan J Public Health. 2021;40(6):660-679)"