Abstract
Background
Type 2 diabetes is a prevalent chronic condition that may pose substantial psychological challenges for patients, particularly in managing long-term glycemic control and its associated complications. Although previous research has examined diabetes distress among individuals with diabetes, studies investigating its latent profiles and associated factors remain scarce. Therefore, this study aimed to identify distinct latent profiles of diabetes distress among patients with type 2 diabetes and explore the factors associated with profile membership.
Methods
A convenience sampling method was employed to recruit 155 patients with T2D from a tertiary hospital in Heilongjiang Province. Participants completed a general information questionnaire, the Diabetes Distress Scale, and the Self-Regulation Fatigue Scale (SRFS). Latent profile analysis (LPA) was used to categorize diabetes distress, and unordered multinomial logistic regression was applied to identify factors associated with each profile.
Results
Diabetes distress was categorized into three latent profiles: low (38.1%), moderate (17.4%), and high (44.5%). The three-class model was identified as the optimal solution based on model fit indices and classification quality, with high entropy (0.905) indicating good separation between profiles. Factors significantly associated with higher distress included being female, older age, more comorbidities, longer duration of diabetes, higher HbA1c levels, and greater self-regulation fatigue (all P < 0.05).
Conclusion
LPA identified three distinct profiles of diabetes distress among patients with type 2 diabetes. Targeted interventions for patients with moderate and high levels of distress, focusing on alleviating psychological burden and self-regulation fatigue, may help improve disease management and overall quality of life.
Supplementary Information
The online version contains supplementary material available at 10.1186/s40359-026-05210-0.
Keywords: Type 2 diabetes, Diabetes distress, Latent profile analysis, Self-Regulation fatigue, Determinants
Research background
Type 2 diabetes (T2D) is a chronic condition that significantly affects a large global population, and its prevalence continues to rise [1]. According to the International Diabetes Federation (IDF), the number of diabetes patients will increase to 643 million (11.3%), and by 2045, it will rise to 783 million (12.2%), with a 46% increase, more than double the estimated population growth rate during the same period (20%) [2]. The Lancet published the results of the Global Burden of Disease Study 2021 (GBD 2021), which estimated the global burden of diabetes from 1990 to 2021 and made predictions for 2050 [3]. The analysis revealed that in 2021, there were 529 million diabetes patients globally, with 96.0% being T2D [3].
In addition to its substantial physiological and epidemiological impacts, T2D is frequently associated with significant diabetes distress [4]. Diabetes distress is a multifaceted psychological construct characterized by diabetes-specific emotional burdens and concerns related to disease management, treatment demands, and potential complications [5]. This distress primarily stems from the chronic and progressive nature of the disease, lifelong requirement for self-management, persistent threat of diabetes-related complications, and continuous concern regarding blood glucose control and future health outcomes [6, 7].
From an epidemiological standpoint, there is increasing evidence indicating that diabetes distress is highly prevalent in individuals with T2D. For instance, research has revealed that among a cohort of 893 Chinese community-dwelling individuals with T2D, the rates of depressive symptoms and anxiety are markedly elevated, at 56.1% and 43.6%, respectively [8]. Similarly, a population-based study involving 1,954 individuals with T2D found that 11.9% of participants experienced high diabetes distress based on the Problem Areas in Diabetes 5-item scale (PAID-5 ≥ 40), further demonstrating the psychological burden associated with long-term diabetes management [4].
Given the significant prevalence and adverse effect of diabetes distress, effective self-management has emerged as a crucial element of T2D care.
Self-management is essential for individuals with T2D to achieve glycemic control and prevent complications [9]. Nevertheless, many individuals encounter difficulties in self-care, often evidenced by suboptimal adherence to medication regimens, dietary adjustments, and lifestyle changes [10, 11]. These self-management difficulties are common and have been associated with adverse health outcomes. Moreover, insufficient self-regulation may contribute to these challenges by limiting patients' ability to maintain long-term self-care behaviors, thereby negatively affecting their psychological well-being and quality of life [12]. Previous studies have reported suboptimal levels of self-regulation among individuals with T2D, with one study indicating that approximately 46.5% of patients demonstrated adequate self-regulation [13].
Emerging evidence suggests that prolonged self-management demands may deplete psychological and volitional resources, leading to self-regulation fatigue—a state of mental and physical exhaustion that undermines sustained self-control and effective disease management [14, 15]. Manifestations include reduced willpower, emotional dysregulation, and diminished motivation for self-care. Both theoretical and empirical research link self-regulation fatigue to adverse behaviors, poor treatment adherence, and impaired capacity to cope with disease-related stress, which may in turn be related to diabetes distress such as anxiety and despair [16, 17]. Thus, self-regulation fatigue may represent an important factor associated with diabetes distress in T2D patients.
However, while a correlation has been identified between self-regulation exhaustion and psychological stress, the manner in which T2D patients perceive stress is inconsistent. Diabetes distress in individuals with T2D is marked by significant variability rather than consistency. This variability is influenced by a range of factors, including demographic characteristics, health conditions, sociocultural contexts, and psychological resources [18, 19]. Within Chinese culture, elements such as familial obligations, the financial burden of medical care, and societal expectations regarding health and functional roles can exacerbate the psychological stress experienced by patients during disease management [20]. Such variations can lead to more complex manifestations of diabetes distress.
Such heterogeneity presents challenges for traditional variable-centered approaches, which generally assume population homogeneity and focus on relationships between variables at the group level. Person-centered approaches, such as latent profile analysis (LPA), allow for the identification of distinct subgroups based on shared response patterns and may therefore better capture the diversity of diabetes distress experiences among individuals with T2D [21, 22]. In diabetes research, person-centered methods have increasingly been used to identify subgroups of patients with different patterns of self-management behaviors, treatment adherence, and emotional responses, providing valuable insights for tailored interventions [23].
Compared with latent class analysis (LCA), which is typically applied to categorical indicators, LPA is more appropriate for continuous variables such as DDS subscale scores and enables more precise characterization of latent psychological profiles [24, 25]. Given that the Diabetes Distress Scale yields continuous subscale scores, LPA was selected as the preferred analytic approach.
Against this backdrop, this study aims to employ LPA to identify potential categories of diabetes distress in individuals with T2D. It also seeks to explore the relationships among self-regulation fatigue, demographic characteristics, and various types of diabetes distress. This study aims to clarify the diversity of diabetes distress and its key psychological determinants by considering individual differences. Additionally, it aims to provide both theoretical and practical insights to support the development of more targeted and culturally relevant psychological intervention strategies for this population.
Methods
Study participants and methods
We used a convenience sampling method to consecutively recruit 155 inpatients with T2D from a general hospital in Harbin, China, between April and August 2024. All eligible patients admitted during the study period were approached for participation.
Participants were eligible if they had a confirmed clinical diagnosis of T2D based on medical records and the Chinese Guidelines for the Prevention and Treatment of T2D (2020 Edition) [25], were aged between 18 and 80 years, and were able to communicate and complete the questionnaires independently.
Participants were excluded if they had type 1 diabetes or other specific types of diabetes, severe acute diabetic complications or unstable physical illnesses that could interfere with questionnaire completion, severe cognitive impairment, psychiatric disorders, or communication difficulties, drug or substance dependence, were pregnant or breastfeeding, or had incomplete questionnaire data.
Sample size calculation
The sample size was estimated according to Kendall’s principle, which recommends including 10–20 participants per predictor in multivariable analyses [26]. Given the 12 predictors included in this study and an anticipated 20% rate of invalid or incomplete responses, the target sample size was estimated to range from 150 to 300 participants.
No formal Monte Carlo simulation was conducted to determine the sample size required for LPA, as there is no universally accepted minimum sample size criterion for LPA and sample adequacy depends on multiple factors, including class separation, entropy, and profile proportions [27–29]. Therefore, the identified latent profiles, particularly the small moderate-distress subgroup, should be interpreted cautiously because limited subgroup size may affect profile stability and the precision of subsequent regression analyses.
Ultimately, a total of 175 eligible participants were approached for this study. Among them, 161 participants completed the questionnaire survey, yielding a participation rate of 92.0%. After excluding six questionnaires due to missing data or invalid responses, 155 valid questionnaires were retained for the final analysis, resulting in an effective response rate of 96.3%.
This study received approval from the Second Affiliated Hospital of Harbin Medical University (approval number KY2024-012).
Research instruments
General information questionnaire
A self-developed general information questionnaire was used to ensure that the questions were highly aligned with the study objectives. The questionnaire included the following 11 items: age, gender, body mass index, location, education level, average monthly household income, smoking status, alcohol consumption, number of comorbidities, duration of diabetes, and most recent HbA1c level.
The specific definitions are as follows.
Body Mass Index [30]: The body mass index (BMI) was calculated as weight (kg) divided by height (m) squared. BMI categories were defined as follows: BMI < 18.5 kg/m2 (underweight), 18.5 ≤ BMI < 24 kg/m2 (normal weight), 24 ≤ BMI < 28 kg/m2 (overweight), BMI ≥ 28 kg/m2 (obese).
Smoking [31]: Non-smokers were those who had not smoked during the last month and had never smoked for longer than one year. Subjects were classified as former smokers when they reported that they had smoked for a whole year, had not smoked during the last month, and had stopped smoking. Those who had smoked for longer than a year and had not stopped smoking were classified as smokers.
Alcohol Consumption [32]: Defined as the consumption of any alcoholic beverage (beer, wine, or spirits) in the past month. Drinking was categorized as having consumed alcohol at least once in the past month, non-drinking indicated no alcohol consumption in the past month.
Duration of Diabetes [33]: The time span (in years) from the initial diagnosis of diabetes to the present date for each patient.
HbA1c Testing [34]: It serves as a biomarker for glucose regulation over the prior 2–3 months. An HbA1c level of 6.5% was used as the diagnostic threshold.
Diabetes distress scale
The Chinese version of the Diabetes Distress Scale (DDS), originally established by Polonsky [35] and adapted by Yang Qing [36], was employed to evaluate the diabetes distress of patients with T2D. The DDS has 17 measures distributed across four dimensions: emotional burden, doctor-related distress, life routine-related distress, and interpersonal relationship-related distress. The scale employs a 6-point Likert approach, with responses ranging from “no problem” to “a very serious problem” (scores 1–6), where higher scores signify increased distress. Examples include “Feeling that diabetes is taking up too much of my mental and physical energy every day” and “I am often failing with my diabetes routine.” In the present study, the scale demonstrated acceptable psychometric properties, with a KMO value of 0.781 and a significant Bartlett’s test of sphericity χ2 = 1250.042, p < 0.001. The Cronbach’s α coefficient for the total scale was 0.810, indicating good internal consistency reliability.
Self-Regulation Fatigue Scale (SRFS)
The Chinese version of the Self-Regulation Fatigue Scale (SRFS), created by Nes [17] and modified by Wang Ligang [37], was utilized. The measure comprises three dimensions: cognitive control (6 items), behavioral control (5 items), and emotional control (5 items), amounting to a total of 16 items. The scale utilizes a 5-point Likert scoring system, ranging from "not at all" to "very much," with a maximum attainable score of 80 points. Higher scores indicate greater self-regulation fatigue. Example items include: “I feel mentally exhausted” and “I find it difficult to persist with planned activities.” In the present study, the scale demonstrated acceptable psychometric properties, with a KMO value of 0.810 and a significant Bartlett’s test of sphericity χ2 = 780.773, p < 0.001. The Cronbach’s α coefficient for the total scale was 0.790, indicating acceptable internal consistency reliability.
Statistical methods
LPA was conducted using Mplus 8.3 to identify distinct latent profiles of diabetes distress based on the four subscales of the Diabetes Distress Scale (DDS). Model selection was based on the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), sample-size adjusted Bayesian Information Criterion (aBIC), entropy, Lo–Mendell–Rubin likelihood ratio test (LMRT), and bootstrap likelihood ratio test (BLRT). Lower AIC, BIC, and aBIC values, significant LMRT and BLRT results, and entropy values greater than 0.80, indicating good classification accuracy, were considered when determining the optimal number of latent profiles [38].
All statistical analyses were performed using IBM SPSS Statistics 26. Continuous variables were summarized as means and standard deviations (M ± SD), and categorical variables were presented as frequencies and percentages. Prior to ANOVA, assumptions of normality and homogeneity of variance were assessed. Q–Q plots indicated no substantial deviations from normality, and Levene’s test confirmed homogeneity of variance (F = 0.916, p = 0.402). Detailed normality assessment results are provided in the Supplementary Materials. Differences in continuous variables among latent profiles were examined using one-way analysis of variance (ANOVA), and categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate.
A multinomial logistic regression analysis was subsequently performed to identify factors associated with latent profile membership. Variables were selected based on theoretical relevance, previous literature, and clinical considerations [39, 40]. Multicollinearity was assessed before model estimation, and no significant multicollinearity was detected (all variance inflation factor VIF values < 2). Odds ratios (ORs) with 95% confidence intervals (CIs) were reported. A two-sided p < 0.05 was considered statistically significant.
Results
Assessment of common method bias
Harman’s single-factor test was employed to evaluate the common method bias. The results showed that the eigenvalues of 13 components exceeded 1, with the first factor accounting for only 20.824% of the variance, which is significantly lower than the recommended threshold of 40% [41]. This indicates that common method bias was not a substantial issue in this study.
Demographic attributes
The participants ranged in age from 30 to 80 years, with a mean age of 61.06 ± 12.55 years. Among the 155 participants, 59.4% were female and 56.8% resided in rural areas. Most participants had a normal BMI (52.3%), completed Middle school education (41.9%), and reported a monthly household income ≥ 3000 RMB (65.8%). In addition, 44.5% reported smoking and 30.3% reported alcohol consumption. Regarding health conditions, 41.9% had three or more comorbidities. The mean duration of diabetes was 14.94 ± 8.38 years, and the mean HbA1c level was 10.25% ± 1.99. Detailed demographic and clinical characteristics are presented in Table 1.
Table 1.
General information of patients (n = 155)
| Variable | Category | n (%)/Mean (SD) |
|---|---|---|
| Age (years) | — | 61.06 (12.55) |
| Duration of Diabetes (years) | — | 14.94 (8.38) |
| HbA1c (%) | — | 10.25 (1.99) |
| Sex | Male | 63 (40.6) |
| Female | 92 (59.4) | |
| BMI | Underweight | 11 (7.1) |
| Normal | 81 (52.3) | |
| Overweight | 59 (38.1) | |
| Obese | 4 (2.6) | |
| Residence | Urban | 67 (43.2) |
| Rural | 88 (56.8) | |
| Education Level | Primary school | 55 (35.5) |
| Middle school | 65 (41.9) | |
| High school or above | 35 (22.6) | |
| Average Monthly Household Income | < 3000 RMB | 53 (34.2) |
| ≥ 3000 RMB | 102 (65.8) | |
| Smoking | Yes | 69 (44.5) |
| No | 86 (55.5) | |
| Alcohol Consumption | Yes | 47 (30.3) |
| No | 108 (69.7) | |
| Number of Comorbidities | 1–2 | 90 (58.1) |
| ≥ 3 | 65 (41.9) |
Descriptive statistics of key variables
The total score for diabetes distress was 33.69 ± 7.68. The scores for each dimension were as follows: emotional burden (9.79 ± 3.30), doctor-related distress (5.51 ± 2.18), life routine-related distress (13.71 ± 3.95), and interpersonal relationship-related distress (4.68 ± 2.31). The total score for self-regulation fatigue was 41.25 ± 9.64, including cognitive control (19.37 ± 4.90), emotional control (13.94 ± 4.49), and behavioral control (7.93 ± 2.56). Detailed descriptive statistics are presented in Table 2.
Table 2.
Descriptive statistical analysis of variables (n = 155)
| Scale/Dimension | Min | Max | Mean (SD) | Skewness | Kurtosis |
|---|---|---|---|---|---|
| DDS Total Score | 17 | 52 | 33.69 (7.68) | −0.19 | −0.29 |
| Emotional Burden | 5 | 18 | 9.79 (3.30) | 0.23 | −0.89 |
| Doctor-Related Distress | 4 | 16 | 5.51 (2.18) | 2.45 | 7.53 |
| Life Routine-Related Distress | 5 | 25 | 13.71 (3.95) | −0.22 | 0.27 |
| Interpersonal Relationship-Related Distress | 3 | 12 | 4.68 (2.31) | 1.35 | 0.81 |
| Self-Regulation Fatigue Total Score | 22 | 61 | 41.25 (9.64) | −0.19 | −1.19 |
| Cognitive Control Dimension | 9 | 40 | 19.37 (4.90) | 0.08 | 0.79 |
| Emotional Control Dimension | 5 | 23 | 13.94 (4.49) | 0.02 | −1.18 |
| Behavioral Control Dimension | 5 | 20 | 7.93 (2.56) | 1.19 | 2.52 |
Analysis of potential characteristics of diabetes distress
Identification of latent distress profiles
This study utilized LPA to explore the four dimensions of the Diabetes Distress Scale and identify potential types of diabetes distress in patients with T2D. In the model fitting process, we assessed models ranging from one to five classes. Detailed descriptive statistics are presented in Table 3.
Table 3.
LPA fit information for diabetes distress in T2D Patients (n = 155)
| Model | LL | AIC | BIC | aBIC | LMRT | BLRT | Entropy | Class Probabilities |
|---|---|---|---|---|---|---|---|---|
| 1 | −1526.240 | 3068.480 | 3092.827 | 3067.506 | - | - | - | - |
| 2 | −1469.417 | 2964.834 | 3004.398 | 2963.250 | 0.0002 | 0.000 | 0.887 | 0.49677/0.50323 |
| 3 | −1421.576 | 2879.152 | 2933.934 | 2876.959 | 0.0002 | 0.000 | 0.905 | 0.17419/0.38065/0.44516 |
| 4 | −1379.234 | 2804.469 | 2874.467 | 2801.667 | 0.0279 | 0.000 | 0.922 |
0.03871/0.36129/ 0.17419/0.42581 |
| 5 | −1351.347 | 2758.695 | 2843.911 | 2755.284 | 0.1137 | 0.000 | 0.934 |
0.27742/0.38065/0.03871/ 0.12258/0.18065 |
The results showed that as the number of classes increased, the AIC, BIC, and aBIC values decreased, suggesting an improved model fit. The entropy values were all above 0.80, with the three-class model reaching an entropy of 0.905, indicating high classification accuracy. The LMRT demonstrated that the three-class model was significantly better than the two-class model, whereas the four-class model showed a significant improvement over the three-class model. However, the five-class model did not show a significant enhancement compared to the four-class model. The BLRT test showed that all k-class models were significantly better than the k-1 class models (p < 0.001). Nevertheless, considering class proportions and classification stability, the four-class model included a small class accounting for only 3.8%, resulting in lower classification stability, whereas the class proportions in the three-class model were reasonable: 
low,
moderate, and 44.5% high distress, which is more conducive to interpretation and application (see Fig. 1 for details).
Fig. 1.
LPA of diabetes distress
Consequently, this study identified three latent profiles of diabetes distress: low 
, moderate 
, and high 
.
Classification accuracy and model reliability
To assess the reliability of the LPA outcomes, the average membership probability for each of the three latent classes was computed. The findings revealed that the probability of correct classification was 96.5% for latent class 1, 93.0% for latent class 2, and 98.2% for latent class 3, all of which exceeded the 80% threshold. These results suggest that the three-class latent model retained in this study was reliable (Table 4).
Table 4.
Average latent class probabilities for the most likely latent class membership (rows) by latent class (columns)
| Latent Class 1 | Latent Class 2 | Latent Class 3 | |
|---|---|---|---|
| Latent Class 1 | 0.965 | 0.024 | 0.011 |
| Latent Class 2 | 0.019 | 0.930 | 0.051 |
| Latent Class 3 | 0.005 | 0.012 | 0.982 |
Differences in diabetes distress scores across latent profiles
One-way ANOVA indicated significant differences in distress scores among the three groups: the moderate distress group had a mean score of 36.22 ± 3.84, the high distress group had a mean score of 39.46 ± 4.43, and the low distress group had a mean score of 25.78 ± 4.55 (F = 160.686, P < 0.001). The results are presented in Table 5.
Table 5.
Comparison of Mean Scores (M ± SD) for different latent profiles on the diabetes distress scale dimensions (n = 155)
| Dimension | High Distress Group | Moderate Distress Group | Low Distress Group | F | P |
|---|---|---|---|---|---|
| Emotional Burden | 12.78 (1.77) | 7.96 (2.07) | 7.12 (1.94) | 158.07 | < 0.001 |
| Doctor-Related Distress | 6.17 (2.64) | 5.22 (1.55) | 4.86 (1.56) | 6.421 | 0.002 |
| Life Routine-Related Distress | 16.61 (2.66) | 14.07 (1.54) | 10.15 (3.01) | 94.476 | < 0.001 |
| Interpersonal Relationship-Related Distress | 3.90 (1.19) | 8.96 (1.48) | 3.64 (1.08) | 206.794 | < 0.001 |
| Total Diabetes Distress Scale Score | 39.46 (4.43) | 36.22 (3.84) | 25.78 (4.55) | 160.686 | < 0.001 |
***p < 0.001
**p < 0.01
Although the high distress group demonstrated the highest overall diabetes distress scores, the moderate distress group showed relatively higher scores in the interpersonal distress dimension compared with the high distress group. This finding suggests that interpersonal distress did not increase linearly across latent profiles and may reflect distinct patterns of distress manifestation among different subgroups.
Univariate analysis of demographic factors in the potential profiles
Univariate analysis was performed to assess the demographic characteristics of the three distress profiles. Statistically significant differences were identified in the following variables: age, gender, place of residence, education level, monthly income, number of comorbidities, duration of diabetes, recent HbA1c, and total self-regulation fatigue score (P < 0.05). No significant differences were observed for the other variables. The detailed results are presented in Table 6.
Table 6.
Comparison of general demographic data among different latent profile categories of diabetes distress (n = 155)
| Variable | Moderate Distress Group | Low Distress Group | High Distress Group | Statistic | P |
|---|---|---|---|---|---|
| Age (years) | 59.56 (14.96) | 55.93 (13.06) | 66.04 (8.76) | 12.081ᶜ | < 0.001 |
| Gender | 51.960ᵃ | < 0.001 | |||
| Male | 8 (29.6) | 45 (76.3) | 10 (14.5) | ||
| Female | 19 (70.4) | 14 (23.7) | 59 (85.5) | ||
| BMI | 5.869ᵇ | 0.41 | |||
| Underweight | 1 (3.7) | 2 (3.4) | 8 (11.6) | ||
| Normal | 13 (48.1) | 35 (59.3) | 33 (47.8) | ||
| Overweight | 13 (48.1) | 21 (35.6) | 25 (36.2) | ||
| Obesity | 0 (0.0) | 1 (1.7) | 3 (4.3) | ||
| Location | 11.384ᵃ | 0.003 | |||
| Urban | 17 (63.0) | 30 (50.8) | 20 (29.0) | ||
| Rural | 10 (37.0) | 29 (49.2) | 49 (71.0) | ||
| Educational Level | 13.388ᵃ | 0.009 | |||
| Primary School | 4 (14.8) | 17 (28.8) | 34 (49.3) | ||
| Middle School | 13 (48.1) | 27 (45.8) | 25 (36.2) | ||
| High School or Higher | 10 (37.0) | 15 (25.4) | 10 (14.5) | ||
| Smoking | 4.998ᵃ | 0.084 | |||
| Yes | 17 (63.0) | 22 (37.3) | 30 (43.5) | ||
| No | 10 (37.0) | 37 (62.7) | 39 (56.5) | ||
| Alcohol Consumption | 5.180ᵃ | 0.076 | |||
| Yes | 13 (48.1) | 17 (28.8) | 17 (24.6) | ||
| No | 14 (51.9) | 42 (71.2) | 52 (75.4) | ||
| Family Average Monthly Income | 8.663ᵃ | 0.013 | |||
| < 3000 RMB/month | 10 (37.0) | 12 (20.3) | 31 (44.9) | ||
| ≥ 3000 RMB/month | 17 (63.0) | 47 (79.7) | 38 (55.1) | ||
| Number of Pre-existing Diseases | 21.231ᵃ | < 0.001 | |||
| 1–2 diseases | 12 (44.4) | 48 (81.4) | 30 (43.5) | ||
| ≥ 3 diseases | 15 (55.6) | 11 (18.6) | 39 (56.5) | ||
| Diabetes Duration (years) | 16.04 (7.50) | 9.03 (5.30) | 19.55 (7.81) | 37.312ᶜ | < 0.001 |
| HbA1c (%) | 10.70 (1.35) | 9.14 (1.76) | 11.03 (1.96) | 18.677ᶜ | 0.004 |
| Self-Regulation Fatigue Total Score | 42.26 (6.29) | 31.83 (6.30) | 48.90 (4.98) | 140.955ᶜ | < 0.001 |
aChi-square test
bKruskal-Wallis test
cANOVA test
Multivariate analysis of factors influencing the characteristics of potential diabetes distress in patients with T2D
An unordered multinomial logistic regression analysis was performed using the low diabetes distress group as the reference category to identify factors associated with latent profile membership. Independent variables included demographic characteristics, disease-related factors, and self-regulation fatigue scores. The variable assignment methods are presented in Table 7, and the regression results are shown in Table 8.
Table 7.
Assignment method for independent variables
| Variable | Assignment Method |
|---|---|
| Age (years) | Original value used |
| Gender | Male = 1; Female = 2 |
| Location | Urban = 1; Rural = 2 |
| Educational Level | Primary school = 1; Middle school = 2; High school or above = 3 |
| Family Average Monthly Income | < 3000 RMB/month = 1; ≥ 3000 RMB/month = 2 |
| Number of Pre-existing Diseases | 1–2 diseases = 1; 3 or more diseases = 2 |
| Diabetes Duration (years) | Original value used |
| Most Recent HbA1c Value | Original value used |
| Self-Regulation Fatigue Total Score | Original value used |
Table 8.
Multivariate analysis of factors influencing latent profiles of diabetes distress in T2D patients
| Profile | B | Standard Error | Wald | P | Exp(B) | Exp(B) 95% CI | ||
|---|---|---|---|---|---|---|---|---|
| Lower | Upper | |||||||
| Moderate Distress Group vs Low Distress Group | Intercept | −15.855 | 5.070 | 9.780 | 0.002 | |||
| Age(years) | 0.014 | 0.036 | 0.147 | 0.702 | 1.014 | 0.945 | 1.087 | |
| Diabetes Duration(years) | 0.211 | 0.095 | 4.950 | 0.026 | 1.235 | 1.025 | 1.488 | |
| Most Recent HbA1c Value(%) | 0.628 | 0.296 | 4.493 | 0.034 | 1.874 | 1.048 | 3.349 | |
| Self-Regulation Fatigue Total Score | 0.228 | 0.068 | 11.296 | 0.001 | 1.257 | 1.100 | 1.435 | |
| Gender | ||||||||
| Male | −2.464 | 0.976 | 6.374 | 0.012 | 0.085 | 0.013 | 0.576 | |
| Female | 0b | |||||||
| Location | ||||||||
| Urban | 0.096 | 0.989 | 0.009 | 0.923 | 1.100 | 0.158 | 7.653 | |
| Rural | 0b | |||||||
| Educational Level | ||||||||
| Primary School | −1.870 | 1.339 | 1.950 | 0.163 | 0.154 | 0.011 | 2.126 | |
| Middle School | 0.732 | 1.194 | 0.376 | 0.540 | 2.080 | 0.200 | 21.583 | |
| High School or above | 0b | |||||||
| Family Average Monthly Income | ||||||||
| < 3000 RMB/month | 0.192 | 1.006 | 0.037 | 0.848 | 1.212 | 0.169 | 8.699 | |
| ≥ 3000 RMB/month | 0b | |||||||
| Number of Pre-existing Diseases | ||||||||
| 1–2 Diseases | −2.690 | 1.232 | 4.763 | 0.029 | 0.068 | 0.006 | 0.760 | |
| 3 or more | 0b | |||||||
| High Distress Group vs Low Distress Group | Intercept | −33.809 | 7.044 | 23.038 | 0.000 | |||
| Age(years) | 0.094 | 0.043 | 4.709 | 0.030 | 1.098 | 1.009 | 1.195 | |
| Diabetes Duration(years) | 0.271 | 0.101 | 7.275 | 0.007 | 1.312 | 1.077 | 1.598 | |
| Most Recent HbA1c Value(%) | 0.891 | 0.340 | 6.882 | 0.009 | 2.439 | 1.253 | 4.747 | |
| Self-Regulation Fatigue Total Score | 0.443 | 0.087 | 26.142 | 0.000 | 1.557 | 1.314 | 1.845 | |
| Gender | ||||||||
| Male | −3.147 | 1.117 | 7.944 | 0.005 | 0.043 | 0.005 | 0.383 | |
| Female | 0b | |||||||
| Location | ||||||||
| Urban | −0.888 | 1.072 | 0.685 | 0.408 | 0.412 | 0.050 | 3.368 | |
| Rural | 0b | |||||||
| Educational Level | ||||||||
| Primary School | −0.541 | 1.431 | 0.143 | 0.705 | 0.582 | 0.035 | 9.620 | |
| Middle School | 1.056 | 1.335 | 0.627 | 0.429 | 2.876 | 0.210 | 39.341 | |
| High School or above | 0b | |||||||
| Family Average Monthly Income | ||||||||
| < 3000 RMB/month | 0.636 | 1.084 | 0.344 | 0.558 | 1.888 | 0.226 | 15.806 | |
| ≥ 3000 RMB/month | 0b | |||||||
| Number of Pre-existing Diseases | ||||||||
| 1–2 Diseases | −2.743 | 1.310 | 4.382 | 0.036 | 0.064 | 0.005 | 0.840 | |
| 3 or more | 0b | |||||||
0b indicates the reference group
Compared with the low distress group, patients with longer diabetes duration (OR = 1.235, 95% CI: 1.025–1.488), higher HbA1c levels (OR = 1.874, 95% CI: 1.048–3.349), and higher self-regulation fatigue scores (OR = 1.257, 95% CI: 1.100–1.435) had significantly higher odds of belonging to the moderate distress group. Male patients had lower odds of belonging to the moderate distress group than female patients (OR = 0.085, 95% CI: 0.013–0.576). In addition, patients with one to two comorbidities had lower odds of belonging to the moderate distress group than those with three or more comorbidities (OR = 0.068, 95% CI: 0.006–0.760).
Similarly, older age (OR = 1.098, 95% CI: 1.009–1.195), longer diabetes duration (OR = 1.312, 95% CI: 1.077–1.598), higher HbA1c levels (OR = 2.439, 95% CI: 1.253–4.747), and higher self-regulation fatigue scores (OR = 1.557, 95% CI: 1.314–1.845) were associated with significantly higher odds of belonging to the high distress group. Male patients showed lower odds of belonging to the high distress group than female patients (OR = 0.043, 95% CI: 0.005–0.383). Furthermore, patients with one to two comorbidities had lower odds of belonging to the high distress group than those with three or more comorbidities (OR = 0.064, 95% CI: 0.005–0.840).
Overall, demographic, clinical, and psychological factors were associated with latent profile membership. Older age, longer diabetes duration, poorer glycemic control, and higher self-regulation fatigue were associated with increased odds of belonging to higher diabetes distress profiles, while male sex was associated with lower odds of membership in these profiles.
Discussion
Analysis of potential characteristics of diabetes distress
This study utilized LPA to classify diabetes distress among patients with T2D into three latent characteristic groups: low (38.1%), moderate (17.4%), and high (44.5%) distress. The findings revealed significant differences among these latent groups across four dimensions: emotional burden, doctor-related distress, life routine-related distress, and interpersonal distress, highlighting the heterogeneity of diabetes distress within the patient population.
The low-distress group exhibited significantly lower scores across all dimensions than the moderate- and high-distress groups, suggesting that the former may have relatively better psychological adaptation and disease self-management abilities, enabling them to cope more effectively with stress. Previous studies have shown that adaptive coping strategies and enhanced self-management behaviors may facilitate psychological adjustment and reduce diabetes-related distress among individuals living with diabetes [42].
Conversely, the moderate- and high-distress groups demonstrated significantly higher scores in the emotional burden and life routine-related distress dimensions, reflecting sustained psychological stress during long-term blood glucose control, treatment adherence, and lifestyle adjustment [43, 44]. Notably, the high distress group did not exhibit the highest interpersonal distress scores; rather, the moderate distress group displayed a higher level of interpersonal distress. This suggests that different levels of diabetes distress do not increase linearly in their manifestations but may present with distinct structural differences. One possible explanation is that patients with high levels of distress may experience reduced social engagement or limited emotional communication, which could influence their perception or reporting of interpersonal distress. Previous research has shown that individuals with depression or social anxiety may exhibit reduced social interactions and difficulties in maintaining interpersonal connections [45]. However, this interpretation should be approached cautiously because interpersonal coping styles and social behaviors were not directly assessed in the present study.
Overall, the moderate- and high-distress groups constituted 61.9% of the total, indicating that diabetes distress is prevalent among patients with diabetes, with a particularly high proportion of highly distressed individuals. This underscores the necessity of stratified assessments and precise interventions based on latent characteristics in clinical settings.
Analysis of factors influencing intergroup variations
The present study identified distinct diabetes distress profiles among patients with T2D, and the multinomial logistic regression results further revealed meaningful differences in demographic, clinical, and psychological characteristics across these latent groups. Rather than reflecting isolated effects of individual variables, these findings suggest that each latent profile may represent a specific combination of disease burden, psychological vulnerability, and self-regulation capacity.
The high distress profile was characterized by older age, longer diabetes duration, poorer glycemic control, a greater number of comorbidities, and higher levels of self-regulation fatigue. In addition, female patients were more likely to belong to this group. These findings suggest that individuals in the high distress profile may experience cumulative psychological and physiological burdens associated with long-term disease management. Previous studies have indicated that diabetes-related multimorbidity and poor glycemic control are associated with greater disease burden and psychological distress, which may further complicate long-term diabetes management [46, 47]. Furthermore, higher self-regulation fatigue was significantly associated with membership in the high distress profile, suggesting that higher self-regulation fatigue may coexist with elevated diabetes distress [48]. Collectively, these characteristics suggest that the high distress group represents a particularly vulnerable subgroup requiring integrated psychological and disease-management support.
Compared with the high distress group, the moderate distress profile appeared to reflect an intermediate state of psychological burden. Although patients in this group also demonstrated longer diabetes duration, higher HbA1c levels, more comorbidities, and higher self-regulation fatigue relative to the low distress group, the magnitude of these associations was comparatively weaker. This pattern may indicate that moderate distress patients still retain partial adaptive capacity despite ongoing disease-related stress.
In contrast, the low distress profile was characterized by relatively better glycemic control, fewer comorbidities, shorter disease duration, and lower levels of self-regulation fatigue. Patients in this group may demonstrate relatively stronger adaptation to long-term disease management. These findings further support the heterogeneity of diabetes distress among individuals with T2D.
Notably, educational attainment, residential location, and household income were not significantly associated with latent profile membership in the multivariable analysis. This may suggest that the effects of socioeconomic characteristics on diabetes distress are indirect and potentially mediated through factors such as disease management behaviors, access to healthcare resources, and social support systems.
Research guidance and significance
These findings indicate that effective diabetes management should extend beyond glycemic control to integrate mental health considerations. LPA revealed that a substantial proportion of patients experienced moderate to high diabetes distress, closely associated with disease-related factors, comorbidities, HbA1c levels, age, gender, and self-regulation fatigue, highlighting high-risk subgroups.
Therefore, clinical management may benefit from stratified psychological support based on latent distress characteristics. Strategies such as cognitive behavioral therapy, emotional management training, and professional counseling may help alleviate the emotional burden and enhance coping skills. Interventions should also address self-regulation and self-efficacy, particularly in patients with long-term, high-intensity self-management requirements.
Overall, a stratified, individualized management approach informed by diabetes distress profiles may help guide precise interventions, improve self-management behaviors, and optimize long-term health outcomes in patients with diabetes.
Research limitations
While this study contributes to understanding the factors associated with diabetes distress, several limitations should be acknowledged. First, diabetes distress and self-regulation fatigue were assessed using self-report questionnaires, which may be affected by recall bias, subjective interpretation, and social desirability bias. Second, participants were recruited from a single general hospital in Harbin, China, using convenience sampling, which may limit the generalizability of the findings to other regions, healthcare settings, or populations with different sociocultural backgrounds. Third, because of the cross-sectional design, causal relationships between self-regulation fatigue and diabetes distress cannot be established, and changes in diabetes distress over time could not be evaluated.
In addition, although the multivariable multinomial logistic regression model was developed based on theoretical considerations, previous literature, clinical relevance, and statistical results, the possibility of residual confounding cannot be completely excluded. Some potentially important confounding variables may not have been measured or adequately controlled. Furthermore, the inclusion of multiple predictors relative to the sample size may have increased the risk of model overfitting, particularly within smaller latent subgroups.
The sample size and latent profile estimation represent important methodological limitations. No formal Monte Carlo simulation-based sample size evaluation was conducted for LPA before data collection. In addition, the relatively small moderate-distress subgroup may have limited profile stability and reduced the precision of multinomial regression estimates. Therefore, the identified latent profiles and associated predictors should be interpreted cautiously and require validation in larger samples.
Future studies should adopt longitudinal, multicenter designs with larger and more diverse samples to improve the stability and generalizability of latent profiles. Incorporating additional variables related to psychosocial resources, coping styles, healthcare access, and sociocultural factors may also help reduce residual confounding and provide a more comprehensive understanding of diabetes distress heterogeneity.
Conclusion
This study employed LPA to classify diabetes distress among patients with T2D into three distinct categories: low, medium, and high, revealing significant heterogeneity. The key determinants identified included gender, age, comorbidity burden, duration of diabetes, HbA1c levels, and self-regulation fatigue. These findings underscore the necessity of incorporating mental health assessments and individualized, stratified interventions into comprehensive diabetes management strategies to support improvements in psychological well-being, self-management, and long-term glycemic control.
Supplementary Information
Acknowledgements
All methods employed in this study were conducted in accordance with the relevant guidelines and regulations. All authors have read and approved the final manuscript and consented to publication.
Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work, the author(s) used GPT to enhance the readability and language quality of the manuscript. Following the use of this service, the authors reviewed and edited the content as necessary and assumed full responsibility for the content of the published article.
Abbreviations
- T2D
Type 2 Diabetes
- LPA
Latent Profile Analysis
- DDS
Diabetes Distress Scale
- SRFS
Self-Regulation Fatigue Scale
- AIC
Akaike Information Criterion
- BIC
Bayesian Information Criterion
- aBIC
Sample-size adjusted Bayesian Information Criterion
- LMRT
Lo-Mendell-Rubin likelihood ratio test
- BLRT
Bootstrap Likelihood Ratio Test
- BMI
Body Mass Index
- IDF
International Diabetes Federation
- GBD
Global Burden of Disease
Authors’ contributions
Siyu Li: Responsible for research design, data collection and organization, clinical procedures and patient management, and drafting the initial manuscript. Pengyue Zheng: Conducted data analysis, statistical processing, manuscript drafting and revision, and literature review. Jie Gao: Contributed to research conception, provided methodological guidance, clinical diagnosis and treatment advice, and manuscript review. Lianheng Xia: Involved in clinical data collection, case management, and research implementation support. Min Liu: Conducted data verification, quality control, and language polishing of the manuscript. Yuhuan Zhang: Oversaw overall study design and project supervision, research process oversight, and final manuscript revision and submission. Zhixin Di: Provided guidance on clinical methods, data analysis oversight, and manuscript review. Siyu Li and Pengyue Zheng contributed equally to this work.
Funding
This study was funded by the Heilongjiang Higher Education Association Special Project on the Third Plenary Session of the 20th CPC Central Committee and the 2024 National Education Conference (24GJZXE003),2023 China Youth Science Fund Project(8230153519), and the 2025 Harbin Medical University Key Project for Comprehensive Reform and Quality Enhancement in Ideological and Political Work (HYDSZZDXM011).
Data availability
The datasets utilized or analyzed during this study are accessible from Zheng upon reasonable requests. Contact: 2,386,663,684@qq.com.
Declarations
Ethics approval and consent to participate
This study was approved by the Second Affiliated Hospital of Harbin Medical University (approval no. KY2024-012). Prior to the commencement of the study, the participants were thoroughly informed of the research objectives and methodologies. Informed consent was obtained from all participants before their inclusion in the study. Participants were assured that their involvement was voluntary, their data would remain confidential, and they could withdraw from the study at any time without any repercussions. Researchers were available to address participant inquiries.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Siyu Li and Pengyue Zheng contributed equally to this work.
Contributor Information
Yuhuan Zhang, Email: 2802262584@qq.com.
Zhixin Di, Email: 44725660@qq.com.
References
- 1.Xie D, You F, Li C, et al. Global, regional, and national burden of type 2 diabetes attributable to dietary factors from 1990 to 2021. Sci Rep. 1990;15:13278. 10.1038/s41598-025-98022-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sun H, Saeedi P, Karuranga S, et al. IDF Diabetes Atlas: global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Res Clin Pract. 2021;183:109119. 10.1016/j.diabres.2021.109119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.GBD 2021 Diabetes Collaborators. Global, regional, and national burden of diabetes from 1990 to 2021, with projections to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2023;402(10397):203–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Riise HKR, Haugstvedt A, Igland J, et al. Diabetes distress and associated psychosocial factors in type 2 diabetes: a population-based cross-sectional study. The HUNT study, Norway. Diabetol Metab Syndr. 2025;17(1):62. 10.1186/s13098-025-01631-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wong MH, Kwan SM, Dao MC, et al. Prevalence and factors associated with diabetes-related distress in type 2 diabetes patients: a study in Hong Kong primary care setting. Sci Rep. 2024;14:10688. 10.1038/s41598-024-61538-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bassi G, Gabrielli S, Donisi V, et al. Assessment of psychological distress in adults with type 2 diabetes mellitus through technologies: literature review. J Med Internet Res. 2021;23(1):e17740. 10.2196/17740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kalra S, Jena BN, Yeravdekar R. Emotional and psychological needs of people with diabetes. Indian J Endocrinol Metab. 2018;22(5):696–704. 10.4103/ijem.IJEM_579_17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sun N, Lou P, Shang Y, et al. Prevalence and determinants of depressive and anxiety symptoms in adults with type 2 diabetes in China: a cross-sectional study. BMJ Open. 2016;6:e012540. 10.1136/bmjopen-2016-012540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Puzhakkal S, Kavanagh S, Conway B, et al. Self-care and self-management in diabetes: concepts, theories and practices. Int J Clin Pharm. 2025;47:2094–100. 10.1007/s11096-025-01941-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Goins RT, Jones J, Schure M, et al. Type 2 diabetes management among older American Indians: beliefs, attitudes, and practices. Ethn Health. 2020;25:1055–71. 10.1080/13557858.2018.1493092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Castillo RR, Pabón Carrasco M, Jiménez-Picón N, et al. Effects of a diabetes self-management education program on glucose levels and self-care in type 1 diabetes: a pilot randomized controlled trial. Int J Environ Res Public Health. 2022;19:16364. 10.3390/ijerph192316364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Estuningsih Y, Rochmah TN, Andriani M, et al. Effect of self-regulated learning for improving dietary management and quality of life in patients with type-2 diabetes mellitus at Dr. Ramelan Naval Hospital, Surabaya Indonesia. Kesmas. 2019;14:51–7. 10.21109/kesmas.v14i2.2257. [Google Scholar]
- 13.Fadli F, Nursalam N, Sjattar EL, et al. An integrative model of self-regulation in type 2 diabetes self-management: a systematic review of individual and family-based interventions. Healthcare (Basel). 2025;13(24):3230. 10.3390/healthcare13243230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang L, Yu Y, Tao T, et al. The self-care dilemma of type 2 diabetic patients: the mechanism of self-regulation resource depletion. PLoS ONE. 2018;13:e0208690. 10.1371/journal.pone.0208690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cameron D, Webb T. Self-regulatory fatigue. In: Gellman MD, Turner JR, editors. Encyclopedia of Behavioral Medicine. Springer; 2013. 10.1007/978-1-4419-1005-9_1178.
- 16.Baumeister RF, Bratslavsky E, Muraven M, et al. Ego depletion: is the active self a limited resource? J Pers Soc Psychol. 1998;74(5):1252–65. 10.1037/0022-3514.74.5.1252. [DOI] [PubMed] [Google Scholar]
- 17.Nes LS, Ehlers SL, Whipple MO, et al. Self-regulatory fatigue in chronic multisymptom illnesses: scale development, fatigue, and self-control. J Pain Res. 2013;6:181–8. 10.2147/JPR.S40014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ravindran OS, Shanmugasundaram N, Madhusudhan S. Optimism, quality of life, and psychological distress in patients with type 2 diabetes mellitus: a case-control study. Indian J Psychiatr Nurs. 2022;19(1):35–42. 10.4103/iopn.iopn_50_21. [Google Scholar]
- 19.Mirzaei-Alavijeh M, Najafi F, Veis RG, et al. Psychological distress and its association with gender, socioeconomic status, education and health conditions. Sci Rep. 2025;15:29142. 10.1038/s41598-025-14848-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shi J, Huang A, Jia Y, et al. Perceived stress and social support influence anxiety symptoms of Chinese family caregivers of community-dwelling older adults: a cross-sectional study. Psychogeriatrics. 2020;20(4):377–84. 10.1111/psyg.12510. [DOI] [PubMed] [Google Scholar]
- 21.Spurk D, Hirschi A, Wang M, et al. Latent profile analysis: a review and “how to” guide of its application within vocational behavior research. J Vocat Behav. 2020;120:103445. 10.1016/j.jvb.2020.103445. [Google Scholar]
- 22.Guo L, Zhang M, Namassevayam G, et al. Identification of sleep quality clusters among stroke patients: a multi-center latent profile analysis study. Sleep Med. 2023;112:203–8. 10.1016/j.sleep.2023.10.019. [DOI] [PubMed] [Google Scholar]
- 23.Pisanti R, Bogosian A, Violani C. Psychological profiles of individuals with type 2 diabetes and their association with physical and psychological outcomes: a cluster analysis. Psychol Health. 2023;38(8):1056–73. 10.1080/08870446.2021.2001469. [DOI] [PubMed] [Google Scholar]
- 24.Oberski DL. Mixture models: latent profile and latent class analysis. In: Robertson J, Kaptein M, editors. Modern statistical methods for HCI. Cham: Springer; 2016. p. 275–87. 10.1007/978-3-319-26633-6_12. [Google Scholar]
- 25.Chinese Diabetes Society. Guideline for the prevention and treatment of type 2 diabetes mellitus in China (2020 edition) (Part I). Chin J Pract Intern Med. 2021;41(8):668–95. 10.19538/j.nk2021080106. (in Chinese). [Google Scholar]
- 26.Ni P, Chen JL, Liu N. Sample size estimation of quantitative studies in nursing research. Chin J Nurs. 2010;45:378–80. (in Chinese). [Google Scholar]
- 27.Shen Z, Yin L, Zhong J, et al. Heterogeneity of ego depletion and its relationship with health promotion behavior among older adults with type 2 diabetes mellitus: latent profile analysis. J Gerontol Nurs. 2025;51(12):44–54. 10.3928/00989134-20251008-02. [DOI] [PubMed] [Google Scholar]
- 28.Tein JY, Coxe S, Cham H. Statistical power to detect the correct number of classes in latent profile analysis. Struct Equ Modeling. 2013;20(4):640–57. 10.1080/10705511.2013.824781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nylund-Gibson K, Choi AY. Ten frequently asked questions about latent class analysis. Transl Issues Psychol Sci. 2018;4(4):440–61. 10.1037/tps0000176. [Google Scholar]
- 30.National Health Commission of the People’s Republic of China. Body weight management guidelines (2024 edition). Beijing: National Health Commission of the People’s Republic of China; 2024. [Google Scholar]
- 31.Slagter SN, van Vliet-Ostaptchouk JV, Vonk JM, et al. Associations between smoking, components of metabolic syndrome and lipoprotein particle size. BMC Med. 2013;11:195. 10.1186/1741-7015-11-195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Slagter SN, van Vliet-Ostaptchouk JV, Vonk JM, et al. Combined effects of smoking and alcohol on metabolic syndrome: the LifeLines Cohort Study. PLoS ONE. 2014;9:e96406. 10.1371/journal.pone.0096406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Zhang D, Zhang Y, Kang J, et al. Nonlinear relationship between diabetes mellitus duration and diabetic retinopathy. Sci Rep. 2024;14:30223. 10.1038/s41598-024-82068-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Diggle J. HbA1c: practicalities and pitfalls. Diabetes Prim Care. 2025;27(3):81. https://diabetesonthenet.com/diabetes-primary-care/hba1c-practicalities-pitfalls/. [Google Scholar]
- 35.Polonsky WH, Fisher L, Earles A, et al. Assessing psychosocial distress in diabetes: development of the diabetes distress scale. Diabetes Care. 2005;28(3):626–31. 10.2337/diacare.28.3.626. [DOI] [PubMed] [Google Scholar]
- 36.Yang Q, Liu X. Evaluation of the reliability and validity of the Chinese version of the diabetes distress Scale. J Nurs Sci. 2010;17(17):8–10. (in Chinese). [Google Scholar]
- 37.Wang L, Zhang J, Wang J, et al. Reliability and validity of the Chinese version of the self-regulatory fatigue scale in assessing fatigue in young people. Chin J Ment Health. 2015;29(4):290–4. 10.3969/j.issn.1000-6729.2015.04.010. (in Chinese). [Google Scholar]
- 38.Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling. 2007;14(4):535–69. 10.1080/10705510701575396. [Google Scholar]
- 39.Busili A, Kumar K, Kudrna L, et al. The risk factors for mental health disorders in patients with type 2 diabetes: an umbrella review of systematic reviews with and without meta-analysis. Heliyon. 2024;10:e28782. 10.1016/j.heliyon.2024.e28782. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Wu R, Luo X, Cui J, et al. Factors influencing psychological insulin resistance among patients with type 2 diabetes in China: a structural equation model analysis. Diabetes Metab Syndr Obes. 2025;18:4385–97. 10.2147/DMSO.S501031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Zhou H, Long L. Statistical remedies for common method biases. Adv Psychol Sci. 2004;12(6):942–50. 10.3969/j.issn.1671-3710.2004.06.018. (in Chinese). [Google Scholar]
- 42.Stanton AL, Revenson TA, Tennen H. Health psychology: psychological adjustment to chronic disease. Annu Rev Psychol. 2007;58:565–92. 10.1146/annurev.psych.58.110405.085615. [DOI] [PubMed] [Google Scholar]
- 43.Aikens JE, Perkins DW, Lipton B, et al. Longitudinal analysis of depressive symptoms and glycemic control in type 2 diabetes. Diabetes Care. 2009;32(7):1177–81. 10.2337/dc09-0175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.de Groot M, Anderson R, Freedland KE, et al. Association of depression and diabetes complications: a meta-analysis. Psychosom Med. 2001;63(4):619–30. 10.1097/00006842-200107000-00015. [DOI] [PubMed] [Google Scholar]
- 45.Villanueva J, Meyer AH, Mikoteit T, et al. Having versus not having social interactions in patients diagnosed with depression or social phobia and controls. PLoS ONE. 2021;16(4):e0249765. 10.1371/journal.pone.0249765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Gregg EW, Pratt A, Owens A, et al. The burden of diabetes-associated multiple long-term conditions on years of life spent and lost. Nat Med. 2024;30(10):2830–7. 10.1038/s41591-024-03123-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Ajele KW, Idemudia ES. The role of depression and diabetes distress in glycemic control: a meta-analysis. Diabetes Res Clin Pract. 2025;221:112014. 10.1016/j.diabres.2025.112014. [DOI] [PubMed] [Google Scholar]
- 48.Liu Z, Li M, Jia Y, et al. Relationship between mindfulness and psychological distress in patients with hepatocellular carcinoma: the mediation effect of self-regulation. Am J Health Behav. 2021;45(6):1041–9. 10.5993/AJHB.45.6.8. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets utilized or analyzed during this study are accessible from Zheng upon reasonable requests. Contact: 2,386,663,684@qq.com.