第一型糖尿病青少年血糖控制模式的建構：縱貫性研究

摘要理想血糖控制是維持第一型糖尿病青少年健康的重要任務，也是健康照護者需要密切關注的健康議題。過去研究鮮少以縱貫性研究進行第一型糖尿病青少年血糖控制之影響因素的路徑分析。本研究目的為建構基線的社會人口學、糖尿病情緒困擾、同儕反應歸因、父母教養方式和基線後三個月的糖尿病自我管理對基線後六個月的HbA1c值影響的路徑。本研究為縱貫性研究，以方便取樣在中、南部四家醫院的兒童內分泌科門診就診的10至18歲第一型糖尿病青少年為研究對象。研究工具包括自填量表和在病歷紀錄收集HbA1C值。量表包括社會人口學和糖尿病情緒困擾－青少年版本、朋友反應的歸因、父母教養方式測量II和第一型糖尿病青少年自我管理等四份中文版量表。量表在原作者同意翻譯和使用後，進行量表中文化，並經由內容效度、建構效度、內在一致性及再測信度的檢定。統計方法包括頻率、百分比、平均數等描述性統計，以驗證性因素分析檢定各量表的建構效度，並以結構方程模式檢定所收集的資料是否支持所假設的「第一型糖尿病青少年血糖控制模式」。本研究基線時共收203位研究對象，基線後三個月共收177位，回覆率為87.2%。經由三次模式修正後獲得一個資料支持的模式，此模式的卡方值為131.819（自由度 = 68, p < .001）、卡方值/自由度（the ratio of chi-squared to the degrees of freedom, ??2 / df）為1.94、比較適配度指標（Comparative Fit Index, CFI）為0.93、Tucker和Lewis指標（Tucker Lewis Index,TLI）為0.91、漸近誤差均方根（Root Mean Square Error for Approximation, RMSEA）為0.07（90%信賴區間：.054－.092）和標準化殘差均方根（Standardized Root Mean Square Residual, SRMR）為0.06。模式資料顯示： 1. 基線的社經地位對基線後三個月的自我管理呈顯著直接的負向影響（β = -.14, t = -2.13, p = .03），但不會透過基線後三個月的糖尿病自我管理而間接影響基線後六個月的HbA1c值（z = 1.88, p = .06）。 2. 基線的父母教養方式對基線的同儕反應歸因呈顯著直接的正向影響（β = .50, t = 4.57, p < .001），並且透過基線的同儕反應歸因能顯著間接影響基線的糖尿病情緒困擾（z = 2.19, p = .01）和基線後三個月的糖尿病自我管理（z = 2.24, p = .03）。 3. 基線的父母教養方式對基線的糖尿病情緒困擾呈顯著直接正向影響（β = .39, t = 3.57, p < .001），並且透過基線的糖尿病情緒困擾能顯著間接影響基線後六個月的HbA1c值（z = 2.09, p = .04）。 4. 基線的父母教養方式對基線後三個月的糖尿病自我管理呈顯著的直接負向影響（β = -.54, t = -4.82, p < .001），並且透過基線後三個月的糖尿病自我管理能顯著間接影響基線後六個月的HbA1c值（z = 3.01, p < .01）。 5. 基線的同儕反應歸因對基線的糖尿病情緒困擾呈顯著直接的正向影響（β = .27, t = 2.49, p = .01），但不會透過基線的糖尿病情緒困擾間接影響基線後六個月的HbA1c值（z = 1.79, p = .07）。 6. 基線的同儕反應歸因對基線後三個月的糖尿病自我管理呈顯著直接的負向影響（β = -.26, t = -2.57, p = .01），並且透過基線後三個月的糖尿病自我管理能顯著間接影響基線後六個月的HbA1c值（z = 2.14, p = .03）。 7. 基線的父母教養方式對基線後六個月的HbA1c值呈顯著的直接負向影響（β = -.38, t = -2.64, p = .01），而基線的糖尿病情緒困擾對基線後六個月的HbA1c值呈顯著直接的正向影響（β = .25, t = 2.57, p = .010）。 8. 基線後三個月的糖尿病自我管理對基線後六個月的HbA1c值呈顯著的直接負向影響（β = -.50, t = -3.87, p < .001）。本研究建議未來可進行不同樣本的模式檢定以瞭解「第一型糖尿病青少年血糖控制模式」在其他樣本的適用性。本研究結果可供健康專業人員對第一型糖尿病青少年血糖控制的瞭解和照護參考。

關鍵字

第一型糖尿病；青少年；血糖控制；模式建構

並列摘要

Abstract Optimal glycemic control is not only an important mission for adolescents with type 1 diabetes, but also should be paid close attention to health issues by health providers. Previous studies have rarely explored the pathways of glycemic control with longitudinal study design in adolescents with type 1 diabetes. The purpose of this study was to construct a model to examine the effects of demographic characteristics, diabetes distress, attribution of peer reactions, parenting style at baseline, and self-management three months after the baseline on HbA1c levels six months after the baseline in adolescents with type 1 diabetes. A longitudinal survey design was applied. Participants of aged 10 to 18 years were recruited from endocrinology outpatient clinics of four hospitals in Central and Southern Taiwan. The self-report questionnaires and HbA1C levels from medical records were collected. The questionnaires included demographic characteristics and all the Chinese version of Problem Areas in Diabetes-Teen, Attribution of Friend Reactions, Parenting Style Inventory-II, and Self-Management of Type 1 Diabetes in Adolescents. The Problem Areas in Diabetes-Teen, the Attribution of Friend Reactions, and the Parenting Style Inventory-II were translated from English version into Chinese version after they were permitted by all the authors of the original questioonaires. Also, the Chinese version Self-Management of Type 1 Diabetes in Adolescents was used after the permission from the author of this original questionnaire. All questionnaires were examined by content validities, construct validities, Cronbach’s α, and test−retest reliabilities. Descriptive statistics were used to describe all the variables. The hypothesis of a model was examined by Structural Equation Model. A total of 203 adolescents with type 1 diabetes were recruited as participants at baseline and 177 partipants were followed three months after the baseline. The response rate was 87.2%. The final model was supported by the data via three examinations of Structural Equation Model. The fit indices were the chi-squared = 131.819 (degrees of freedom = 68, p < .001), the ratio of chi-squared to the degrees of freedom (??2 / df) = 1.94, the comparative fit index (CFI) = 0.93, the Tucker Lewis index (TLI) = 0.91, the root mean square error for approximation (RMSEA) = 0.07 (90% confidence interval = .054－.092), and the standardized root mean square residual (SRMR) = 0.06. The model indicated: 1. Social economic status at baseline had a significantly direct and negative effect on self-management (β = -.14, t = -2.13, p = .03), but did not indirectly impact on HbA1c level six months after the baseline via self-management three months after the baseline (z = 1.88, p = .06). 2. Parenting style at baseline had a significantly direct and negative effect on HbA1c level six months after the baseline (β = -.38, t = -2.64, p = .01). 3. Parenting style at baseline had a significantly direct and positive effect on attribution of peer reactions (β = .50, t = 4.57, p < .001), and significantly indirectly impacted on diabetes distress at baseline (z = 2.19, p = .01）and on self-management three months after the baseline (z = 2.24, p = .03) via attribution of peer reactions at baseline. 4. Parenting style at baseline had a significantly direct and positive effect on diabetes distress at baseline (β = .39, t = 3.57, p < .001), and significantly indirectly impacted on HbA1c level six months after the baseline via diabetes distress at baseline (z = 2.09, p = .04). 5. Parenting style at baseline had a significantly direct and negative effect on self-management three months after the baseline (β = -.54, t = -4.82, p < .001), and significantly indirectly impacted on HbA1c level six months after the baseline via self-management three months after the baseline (z = 3.01, p < .01). 6. Attribution of peer reactions at baseline had a significantly direct and positive effect on diabetes distress (β = .27, t = 2.49, p = .01), but did not indirectly impact on HbA1c level six months after the baseline via diabetes distress at baseline (z = 1.79, p = .07). 7. Attribution of peer reactions at baseline had a significantly direct and negative effect on self-management three months after the baseline (β = -.26, t = -2.57, p = .01), and significantly indirectly impacted on HbA1c level six months after the baseline via self-management three months after the baseline (z = 2.14, p = .03). 8. Diabetes distress at baseline had a significantly direct and positive effect on HbA1c level six months after the baseline (β = .25, t = 2.57, p = .010). 9. Self-management three months after the baseline had a significantly direct and negative effect on HbA1c level six months after the baselin (β = -.50, t = -3.87, p < .001). This model should be cross-validated in various populations in the future to understand the applicability of this model to other samples. The findings could provide a theoretical basis to help health care providers identify protective and risk factors of glycemic control, and then to develop interventions in glycemic control for adolescents with type 1 diabetes.

並列關鍵字

type 1 diabetes ； adolescent ； glycemic control ； constructing a model

參考文獻

Polit, D. F., & Beck, C. T. (2006). The content validity index: Are you sure you know what's being reported? Critique and recommendations. Research in Nursing & Health, 29(5), 489-497. doi: 10.1002/nur.20147

Wang, Y. L., Brown, S. A., & Horner, S. D. (2013). The school-based lived experiences of adolescents with type 1 diabetes. Journal of Nursing Research, 21(4), 235-243. doi:10.1097/jnr.0000000000000003

中華民國內政部（2017）．106年第10週內政統計通報（我國老年人口數首次超過幼年人口數）．2017年4月14日取自http://www.moi.gov.tw/stat/news_content.aspx?sn=11735