背景：

目前，中国糖尿病患病率约13%，2型糖尿病患者占总糖尿病人群的90%以上。糖尿病患者特有的因糖尿病特定压力源导致的负性情绪反应或情感体验称为糖尿病痛苦，在我国2型糖尿病人群中的检出率达50%，可直接或间接降低患者治疗依从性和自我管理能力，导致血糖控制不佳、并发症和死亡风险增加。因此，在明确痛苦相关因素的基础上构建糖尿病痛苦轨迹预测模型，对及时评估和预防糖尿病痛苦、改善2型糖尿病管理至关重要。

目的：

1. 描述2型糖尿病患者糖尿病痛苦现状，确定糖尿病痛苦相关因素。

2. 描述2型糖尿病患者糖尿病痛苦在一年内的变化轨迹类别，构建糖尿病痛苦轨迹预测模型。

方法：

1. 采用横断面研究方法。便利抽取2型糖尿病患者为研究对象，使用糖尿病痛苦量表调查糖尿病痛苦水平，描述糖尿病痛苦现状。使用一般资料调查表、疾病情况调查表、糖尿病自我管理行为量表、失眠严重指数量表、D型人格量表、疾病感知问卷简易版、慢性病自我效能感量表、简易应对方式量表、Herth希望指数量表、社会支持评定量表及杨氏生活事件量表对糖尿病痛苦相关因素进行调查，并分别采用Spearman相关分析、Kruskal-Wallis H检验、Mann-Whitney U检验筛选与糖尿病痛苦相关的连续和有序多分类变量、无序多分类变量、二分类变量。

2. 采用纵向研究方法。从横断面研究的患者中选取自愿且能按时参加纵向研究的患者，在横断面研究基线测量的基础上，使用糖尿病痛苦量表测量患者3个月、6个月、9个月及12个月时的糖尿病痛苦水平。通过拟合带协变量和预测变量的增长混合模型确定糖尿病痛苦轨迹类别并构建糖尿病痛苦轨迹预测模型。

结果：

1. 横断面研究结果：便利抽取936名2型糖尿病患者进行问卷调查，（1）一般人口学资料：研究对象中男性514人，女性422人，中位年龄57.00 (44.00, 67.00) 岁。（2）糖尿病痛苦现状：①糖尿病痛苦中位得分1.59 (1.24, 2.24)，处于轻度痛苦水平，维度得分从高到低排列依次为情绪负担、治疗方案相关痛苦、医生相关痛苦、人际关系相关痛苦。②糖尿病痛苦检出率为34.40%，中度痛苦占总人数的24.14%，高度痛苦占10.26%。（3）糖尿病痛苦相关因素：本研究得到有统计学意义（P<0.05）的糖尿病痛苦相关因素共四大类28个，包括①一般人口学特征1个，为经济状况满意度，与糖尿病痛苦显著负相关（ρ=-0.237）；②2型糖尿病疾病情况10个，包括2型糖尿病病程（ρ=0.106）、并发症及合并症数量（ρ=0.094）、糖化血红蛋白（ρ=0.192）、三个月内平均空腹血糖（ρ=0.155）、三个月内平均餐后2小时血糖（ρ=0.140）、治疗方案（H=14.571）、胰岛素种类数（ρ=0.143）和注射次数（ρ=0.145）、治疗花费的时间（ρ=0.088）和费用（ρ=0.116），均与糖尿病痛苦显著正相关；③生活方式3个，自我管理行为（ρ=-0.181）与糖尿病痛苦显著负相关，失眠严重程度（ρ=0.356）及吸烟情况（H=8.521）与糖尿病痛苦显著正相关；④心理社会因素14个，D型人格（Z=9.441）、感知疾病后果（ρ=0.492）、感知疾病持续时间（ρ=0.110）、感知疾病症状（ρ=0.385）、对疾病的情绪反应（ρ=0.438）、消极应对方式（ρ=0.110）、生活事件总刺激量（ρ=0.188）等7个因素与糖尿病痛苦显著正相关，感知个人控制（ρ=-0.281）、感知治疗控制（ρ=-0.220）、对疾病的理解感知（ρ=-0.185）、自我效能感（ρ=-0.424）、积极应对方式（ρ=-0.151）、希望（ρ=-0.257）、社会支持（ρ=-0.159）等7个因素与糖尿病痛苦显著负相关。

2. 纵向研究结果：从横断面调查的936名研究对象中选取自愿且能按时参加随访的研究对象443人进行随访问卷调查，研究对象包括男性243人，女性200人，中位年龄54.00 (41.00, 64.00) 岁。（1）糖尿病痛苦轨迹类别：一年内，糖尿病痛苦呈现出从轻度上升到中度（21.22%）、持续中度略有波动（30.93%）和持续轻度（47.86%）3种轨迹类别。（2）糖尿病痛苦轨迹预测模型：以持续轻度轨迹组为参照组，①从轻度上升到中度轨迹的显著正向预测因素为消极应对方式（b=11.790）和吸烟情况（b=3.801）；负向预测因素为社会支持（b=-0.811），P值均小于0.05。②持续中度略有波动轨迹的显著正向预测因素按重要性排序依次为D型人格（b=25.753）、感知疾病后果（b=5.279）、空腹血糖（b=4.449）、对疾病的理解感知（b=2.668）、对疾病的关注（b=2.236）和治疗花费的时间（b=0.064）；显著负向预测因素依次为积极应对方式（b=-20.329）、自我效能感（b=-13.736）、经济状况满意度（b=-8.950）和治疗费用（b=-0.381），P值均小于0.05。

结论：

我国2型糖尿病成人中，超过三分之一的患者出现了中度或高度糖尿病痛苦。糖尿病痛苦动态变化，其变化轨迹存在异质性，一年内呈现出从轻度上升到中度、持续中度略有波动、持续轻度三种轨迹类别。以持续轻度糖尿病痛苦为参照轨迹，对从轻度上升到中度轨迹，消极应对方式的正向预测作用最强，社会支持的负向预测作用最弱；对持续中度略有波动轨迹，D型人格的正向预测作用最强，治疗花费时间的正向预测作用最弱；积极应对方式的负向预测作用最强，治疗费用的负向预测作用最弱。

Background:

Currently, the prevalence of diabetes in China is about 13%, and type 2 diabetes mellitus accounts for more than 90% of the patients with diabetes. The unique negative emotional reaction or affective experience caused by diabetes-specific stressors in patients with diabetes is called diabetes distress. In China, the prevalence of diabetes distress in patients with type 2 diabetes mellitus is about 50%. Diabetes distress is a set of negative emotional response or affective experience that is unique to people with diabetes due to diabetes-specific stressors, and has an average prevalence of 50% in China. Diabetes distress can directly or indirectly reduce patients’ treatment adherence and self-management behaviors, leading to poor glycemic control and increased risk of complications and death. Therefore, it is essential to construct a diabetes distress trajectory prediction model based on the identification of correlates of diabetes distress for the timely assessment and prevention of diabetes distress, and the improved management of type 2 diabetes mellitus.

Objectives:

1. To describe the current status of diabetes distress in patients with type 2 diabetes mellitus, and to identify the factors associated with diabetes distress.

2. To describe the trajectory categories of diabetes distress within one year in patients with type 2 diabetes mellitus, and to construct a trajectory prediction model of diabetes distress.

Methods:

1. A cross-sectional study was conducted. The study participants with type 2 diabetes mellitus were selected by convenience sampling. The Diabetes Distress Scale was used to investigate the level of diabetes distress and to describe its current status. Correlates of diabetes distress were investigated by general demographic information questionnaire, disease condition questionnaire, Summary of Diabetes Self-Care Activities, Insomnia Severity Index, Type D Personality Scale, Brief Illness Perception Questionnaire, Self-Efficacy for Managing Chronic Disease, Simplified Coping Style Questionnaire, Herth Hope Index, Social Support Rating Scale and Yang’s Life Event Scale, respectively. To identify correlates of diabetes distress, continuous variables and ordered multiple categorical variables, unordered multiple categorical variables, and binary variables were screened for their associations with diabetes distress using Spearman’s correlation analysis, Kruskal-Wallis H test, and Mann-Whitney U test, respectively.

2. A longitudinal study was conducted. Patients who volunteered to participate in the longitudinal study on time were screened from the patients in the cross-sectional study. Based on the baseline measurement from the cross-sectional study, diabetes distress was measured at 3, 6, 9 and 12 months by the Diabetes Distress Scale. Growth mixture modeling was fitted to derive the trajectory categories and the trajectory prediction model for diabetes distress.

Results:

1. The cross-sectional study results: 936 patients with type 2 diabetes mellitus were selected through convenience sampling for questionnaire survey using. (1) General demographic information: The study population consisted of 514 males and 422 females, with a median age of 57.00 (44.00, 67.00) years. (2) Current status of diabetes distress: (i) The median score of the Diabetes Distress Scale was 1.59 (1.24, 2.24), which was at the level of little diabetes distress, and the dimensions in descending order were emotional burden, regimen distress, physician distress, and interpersonal distress. (ii) The prevalence of diabetes distress was 34.40%, with moderate and high distress accounting for 24.14% and 10.26% of the study population, respectively. (3) Correlates of diabetes distress: there were 28 correlates with statistical significance (P<0.05) in four categories. (i) In general demographic characteristics, there was only one correlate that the satisfaction with economic status was negatively correlated with diabetes distress (ρ=-0.237); (ii) In type 2 diabetes mellitus disease conditions, there were 10 correlates that the duration of type 2 diabetes mellitus (ρ=0.106), number of complications and comorbidities (ρ=0.094), glycosylated hemoglobin (ρ=0.192), mean fasting blood glucose over three months (ρ=0.155), mean 2-hour postprandial glucose over three months (ρ=0.140), treatment regimen (H=14.571), number of insulin types (ρ=0.143) and injection times (ρ=0.145), time (ρ=0.088) and money (ρ=0.116) spent on treatment were positively associated with diabetes distress; (iii) In Lifestyle, there were 3 correlates that the self-management behavior (ρ=-0.181) was negatively correlated with diabetes distress, whereas the severity of insomnia (ρ=0.356) and smoking status (H=8.521) were positively related to diabetes distress; (iv) In psychosocial factors, there were 14 correlates that the following 7 factors including the type D personality (Z=9.441), perceived illness consequences (ρ=0.492), perceived illness timeline (ρ=0.110), perceived illness identity (ρ=0.385), emotional response to illness (ρ=0.438), negative coping style (ρ=0.110), and total stimulus of life events (ρ=0.188) were positively associated with diabetes distress, whereas another 7 factors including the perceived personal control (ρ=-0.281), perceived treatment control (ρ=-0.220), illness comprehensibility (ρ=-0.185), self-efficacy (ρ=-0.424), positive coping style (ρ=-0.151), level of hope (ρ=-0.257), and level of social support (ρ=-0.159) were negatively associated with diabetes distress.

2. The longitudinal study results: From the 936 participants in the cross-sectional study, 443 participants who volunteered to attend the follow-ups on time were selected for the follow-up surveys. The study population consisted of 243 males and 200 females, with a median age of 54.00 (41.00, 64.00) years. (1) Trajectory Categories of diabetes distress: Within one year, there were three trajectory categories of diabetes distress, namely increasing from little to moderate (21.22%), persistently moderate with slight fluctuations (30.93%), and persistently little (47.86%). (2) Trajectory prediction model of diabetes distress: Using the “persistently little diabetes distress” group as the reference group, (i) The significant positive predictors of the trajectory “increasing from little to moderate diabetes distress” were, in order of importance, negative coping style (b=11.790), and smoking status (b=3.801)；negative predictors were social support (b=-0.811), all with P-values less than 0.05. (ii) For the trajectory “persistently moderate with slight fluctuations”, significant positive predictors in order of importance were type D personality (b=25.753), perceived disease consequences (b=5.279), fasting blood glucose (b=4.449), perceived understanding of the disease (b=2.668), concerns about the disease (b=2.236) and time spent on treatment (b=0.064); the significant negative predictors were, in order, positive coping style (b=-20.329), self-efficacy (b=-13.736), satisfaction with financial status (b=-8.950), and money spent on treatment (b=-0.381), all with P-values less than 0.05.

Conclusion:

More than one-third of Chinese adults with type 2 diabetes mellitus experience moderate or high levels of diabetes distress. Diabetes distress is dynamic, and its trajectories are heterogeneous. Within one year, there are three trajectory categories, namely continuous little diabetes distress, increase from little to moderate diabetes distress, slight fluctuations in continuous moderate diabetes distress. Taking the trajectory “continuous little diabetes distress” as the reference trajectory, negative coping style was the strongest positive prediction and social support was the weakest negative prediction for the trajectory “increase from little to moderate diabetes distress”. For the trajectory “continuous moderate diabetes distress”, type D personality was the strongest positive prediction, and time spent on treatment was the weakest; whereas positive coping style was the strongest negative prediction, and treatment cost was the weakest.