Latent class analysis for quality of life status, sleep quality and anxiety in patients with type 2 diabetes

Abstract

Introduction

Type 2 diabetes (T2D) is a chronic metabolic disorder that is associated with reduced sleep quality and anxiety, and can cause a decrease in quality of life in these patients. Despite previous studies investigating these factors, few studies have examined their co-occurrence in these patients. To address this research gap, the present study aimed to determine the subgroups of patients with type 2 diabetes based on quality of life, sleep quality, and anxiety in the subgroups of using latent class analysis (LCA).

Methods

This cross-sectional study was conducted using multistage random sampling. A total of 308 patients with type 2 diabetes were randomly selected from health centers in Ardabil. All participants completed four sets of checklists and questionnaires, (Demographic characteristics, 12-item Short Form survey, the Pittsburgh Sleep Quality Index and Generalized Anxiety Disorder 7-item). Data analysis was performed using Analysis of Variance (ANOVA), chi square and latent class analysis.

Results

Three latent classes were identified: The first class (good status) included 56.4% of the participants. Also, the second (moderate status) and third (poor status) classes described 16.5% and 27.1% of the participants, respectively. In latent class 1, the probability of having good quality of life and good sleep quality was higher. In latent class 2, the probability of having moderate quality of life and poor sleep quality was higher. However, these patients revealed no anxiety. Those with third latent class membership were more likely to have moderate quality of life, poor sleep quality, and severe anxiety.

Conclusion

This study showed that sleep quality and anxiety is positively related to quality of life in patients with type 2 diabetes. In addition, this study indicated the co-occurrence of sleep quality and anxiety in these patients. Based on these findings, effective and targeted interventions can be designed to improve the health status and quality of life of these patients, taking into account sleep quality and anxiety.

Background

Type 2 diabetes (T2D) is a metabolic disorder and a global health problem that can lead to long-term or irreversible complications [1]. It is among the most important diseases that significantly contribute to a major share of premature mortality and disability worldwide [2]. The prevalence of T2D is increasing worldwide and is expected to exceed 592 million by 2035 [3]. Studies conducted in countries such as China [4], the United States [5], and Iran [6] have shown that more than 30% of the elderly population are diagnosed with diabetes mellitus. Due to its numerous complications, this disease is one of the most significant causes of poor quality of life worldwide [7, 8].

According to the World Health Organization, quality of life is defined as an individual’s perception of their life within the context of the culture in which they live and in relation to their expectations, standards, goals, and concerns [9]. The impact of diabetes mellitus, disease treatment, and healthcare costs in type 2 diabetic patients is largely influenced by this index, and it overall reflects the physical and psychological burden [10].In addition, having a poor quality of life for people with diabetes and their families can have numerous effects on their physical and mental health [7]. The main goal of diabetes management is to improve their health-related quality of life [10]. In recent years, evaluating quality of life has been regarded as one of the most important factors in assessing diabetes care worldwide [11].

Anxiety is a common complication of diabetes, with a prevalence rate of 22 to 75% [3]. Yücel et al. study suggests that diabetic patients’ social life may be affected by anxiety about an uncertain future, fear of dependence on care, changes in appearance, and poor body function [12]. Anxiety disorders are also associated with increased complications, including lower blood glucose levels, poor adherence to treatment, increased health care costs, an increased likelihood of disability and loss of productivity, and an increased risk of death in people with diabetes, resulting in a reduced quality of life [3, 13]. Therefore, emotional disorders, such as anxiety and depression, pose a significant threat to quality of life in people with type 2 diabetes compared to healthy individuals [10].

The natural physiological process of sleep is crucial for psychological and biological functioning, including glucose metabolism in diabetic individuals. Consequently, it plays an important role in the well-being of people with chronic diseases [14].This factor has a two-way relationship with diabetes, and insomnia can sometimes be a cause of diabetes or a risk factor for diabetes [15]. Sleep disorders are a common symptom in people with diabetes, with a prevalence of 30–50% [3]. Poor sleep quality is associated with an increased risk of physical and mental health-related outcomes [14]. Previous research has shown that difficulty falling asleep, disruption of the sleep-wake cycle, and sleep-related disorders are among the main problems of people with diabetes, which are associated with poor blood sugar management and can affect the quality of life of these people [16, 17].

Diabetes is a chronic disease with numerous complications [3]. Therefore, based on evidence, clustering these patients according to sleep pattern and anxiety is necessary to understand their co-occurrence and gain greater awareness of individuals’ health statuses to design preventive interventions and reduce the risk of adverse health outcomes [18]. One of the effective ways of identifying health-related problems is using latent class analysis (LCA) [19]. LCA is a statistical method to detect heterogeneity among groups by analyzing behavioral patterns and latent classes with similar patterns [20]. Based on previous research, a few studies have used this method to identify subgroups of diabetic patients based on biological or behavioral factors [21, 22]. To our knowledge, there have been no studies that have examined quality of life, anxiety, and sleep quality in this patient group at the same time.

Maintaining a good quality of life in the face of challenges and complications of diabetes is one of the main goals of for patients with diabetes. Finding the reasons for the decrease in quality of life, particularly with regard to mental health conditions and sleep quality, is crucial [13]. Given the increasing prevalence of T2D in developing countries, such as Iran [15], despite the conduction of several studies on the relation between quality of life and anxiety, and the relation between quality of life and sleep quality in T2D patients, there are limited number of studies focusing on the relationship between quality of life with anxiety and sleep quality [3]. Previous studies have examined anxiety factors and sleep quality separately, and few studies have examined the co-occurrence of these two factors in disease subgroups. The present study aimed to determine the subgroups of patients with type 2 diabetes based on quality of life, sleep quality, and anxiety in the subgroups of using latent class analysis (LCA).

Methods

This cross-sectional study was conducted in Ardabil (a city in northwest Iran) in 2023 using multistage random sampling. The study population was identified based on the lists of diabetic patients who were available in the Ardabil health centers. There are five municipal districts in Ardabil. In the first stage, in each district, three health centers were randomly selected. In the second stage, in proportion to the number of population in each district and population covered by each health center, 308 patients were recruited for this study (Fig. 1). Informed consent was obtained from all participants and they were assured that any information obtained in connection with the study would remain confidential. The study was approved by the Ethics Committee of Ardabil University of Medical Sciences. In this study all procedures were performed in accordance with declaration of Helsinki.

Fig. 1.

Schematic diagram of study population selection

Several checklists and questionnaires were completed by all participants. In the first section of the questionnaire, demographic characteristics were evaluated. The second section assessed the quality of life, which was measured using the 12-item Short Form survey (SF-12). More details about this questionnaire can be found elsewhere [19]. The validity and reliability of the questionnaire were confirmed in Iran (Cronbach’s alpha:0.73) [23]. Additionally, the Pittsburgh Sleep Quality Index (PSQI) was employed to measure sleep quality among patients. More information has been reported elsewhere [19]. According to Gholi Mezerji et al. the item content validity index was excellent (≥ 0.78), the scale content validity index was excellent (≥ 0.90) and the Cronbach alpha coefficient was 0.65 in Iran [24]. The last section of the questionnaire assessed anxiety status among patients. This variable was measured using the Generalized Anxiety Disorder 7-item (GAD-7). Each item of this scale is scored on a three-point Likert scale ranging from Not at all to Nearly every day, and the response options include “not at all, several days, more than half the days, and nearly every day”. GAD-7 total score for the seven items ranges from 0 to 21. 0–4: minimal anxiety. Also, scores 0–4, 5–9, and 10–14 were respectively considered as minimal, mild and moderate anxiety [25]. The validity and reliability of the questionnaire were confirmed in Iran (Cronbach’s alpha:0.78) [26].

LCA was used to investigate the relationship between quality of life and sleep quality with anxiety. Four obvious variables, including quality of life, sleep quality, sleep duration, and generalized anxiety disorder were selected to classify T2D patients. According to the four observed variables, there could be 16 response options. To select the optimal number of classes, an exploratory approach was used, Which started with the most parsimonious 1-class model and successively increased the number of classes by one, until no improvement was observed. LCA models were fitted from a 1-class model to a 6-class model, and the G2 index (which is an alternative for chi-square), Akaike information criteria (AIC), Bayesian information criteria (BIC), and Entropy were calculated for each model. For the data, the model with the lowest values ​​for G2, AIC, and BIC indices may be the best-fitting model. In addition, the ability to interpret the results can also be a criterion to decide on the best model. Item response probabilities higher than 0.5 were considered to label latent classes and describe the characteristics of each class. Data analysis was performed using Analysis of Variance (ANOVA), chi square and latent class analysis. Statistical analysis was performed using SPSS version 16, SAS version 9.4, and STATA version 104.0. Also, in all analyses, a P-value less than 0.05 was considered statistically significant.

Results

The mean age of the participants was 57.7 (10.2) years. Among the total sample, 55.2% (n = 170) were male, and the majority were married. Most participants were employed, while only 13.3% were illiterate. Regarding sleep patterns, 59.4% (n = 183) of the participants demonstrated good sleep quality, and only 9.1% (n = 28) reported a sleep duration of less than five hours.

Table 1 summarizes the sleep quality, sleep duration, anxiety levels, and demographic characteristics of the patients, stratified by quality of life (QoL) scores. Notably, 22.4% of the participants were found to have severe anxiety.

Table 1.

Sleep quality, sleep duration and general anxiety disorder by quality of life in a sample of patients with type 2 diabetes

Characteristics	Quality of life			P-value	Total (n = 308) N(%)
	Poor(n%)	Moderate(n%)	Good(n%)
	37(12.0)	132(42.8)	139(45.1)
Age, Mean (SD)	58.76(8.1)	57.49(10.1)	57.54(10.7)	0.788	57.67(10.2)
Gender
Male	25(14.7)	79(46.5)	66(38.8)	0.034	170(55.2)
Female	12(8.7)	53(38.4)	73(52.9)		138(44.8)
Marital status
Single	5(16.7)	15(50.0)	10(33.3)	0.364	30(9.7)
Married	32(11.5)	117(42.1)	129(46.4)		278(90.3)
Employment
Employed	8(6.8)	44(37.6)	65(55.6)	0.034	117(38.0)
Retired	5(12.2)	20(48.8)	16(39.0)		41(13.3)
Jobless	24(16.0)	68(45.3)	58(38.7)		150(48.7)
Education
Illiterate	9(22.0)	18(43.9)	14(34.1)	0.094	41(13.3)
Non-academic	23(12.4)	80(43.2)	82(44.3)		185(60.1)
Academic	5(6.1)	34(41.5)	43(52.4)		82(26.6)
Sleep quality
Good	4(2.2)	64(35.0)	115(62.8)	< 0.001	183(59.4)
Poor	33(26.4)	68(54.4)	24(19.2)		125(40.6)
Sleep duration
> 7 h	7(8.6)	31(38.3)	43(53.1)	0.114	81(26.3)
6–7 h	15(10.6)	60(42.6)	66(46.8)		141(45.8)
5–6 h	9(15.5)	25(43.1)	24(41.4)		58(18.8)
< 5 h	6(21.4)	16(57.1)	6(21.4)		28(9.1)
General anxiety disorder
No	2(1.4)	40(28.8)	97(69.8)	< 0.001	139(45.1)
Mild	4(6.9)	24(41.4)	30(51.7)		58(18.8)
Moderate	8(19.0)	28(66.7)	6(14.3)		42(13.6)
Severe	23(33.3)	40(58.0)	6(8.7)		69(22.4)

Latent Class Analysis (LCA) was performed using four key variables. Models ranging from one to six latent classes were evaluated. Model fit was assessed using the likelihood-ratio statistic (G²), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC), as presented in Table 2. Based on these indices and the interpretability of the model, the three-class solution was determined to be the most suitable.

Table 2.

Comparison of LCA models with different latent classes based on model selection statistics

Number of latent class	Number of parameters estimated	G2	Df	AIC	BIC	Entropy	Maximum log-likelihood
1	9	356.2	86	374.2	407.8	--	-1285.5
2	19	140.8	76	178.8	249.7	0.79	-1177.7
3	29	97.4	66	155.4	263.5	0.86	-1156.0
4	39	65.2	56	143.2	288.7	0.83	-1140.0
5	49	55.1	46	153.1	335.9	0.84	-1134.9
6	59	45.5	36	163.5	383.6	0.84	-1130.1

Note. LCA = latent class analysis; AIC = Akaike information criterion; BIC = Bayesian information criterion

The results of the three-class LCA model are shown in Table 3; Fig. 2. Participants were categorized into three distinct latent classes: Class 1 included 56.4% of the sample, Class 2 comprised 16.5%, and Class 3 accounted for 27.1%. Patients in Class 1 exhibited a higher probability of good quality of life and good sleep quality, with a lower likelihood of experiencing anxiety. Class 2 was characterized by a higher probability of moderate quality of life and poor sleep quality, but with no reported anxiety. In contrast, individuals in Class 3 were more likely to exhibit moderate quality of life, poor sleep quality, and severe anxiety.

Table 3.

Three latent class model of quality of life, sleep quality and general anxiety disorder in a sample of patients with type 2 diabetes

Items	Latent class
	Good status	Moderate status	Poor status
Latent class prevalence	0.56	0.17	0.27
Quality of life
Poor	0.01	0.04	0.41
Moderate	0.33	0.50 *	0.59
Good	0.66	0.47	0.01
Sleep quality
Good	0.99	0.03	0.12
Poor	0.01	0.97	0.88
Sleep duration
> 7 h	0.39	0.09	0.11
6–7 h	0.50	0.38	0.42
5–6 h	0.11	0.33	0.26
< 5 h	0.01	0.12	0.20
General anxiety disorder
No	0.63	0.56	0.01
Mild	0.18	0.39	0.08
Moderate	0.09	0.04	0.29
Severe	0.10	0.00	0.62

^* Item-response probabilities > 0.5 in bold to facilitate interpretation

Fig. 2.

Three latent class model of quality of life, sleep quality and general anxiety disorder in a sample of patients with type 2 diabetes

Discussion

In this study, three latent classes were identified based on quality of life (QoL), sleep quality, and anxiety levels among individuals with type 2 diabetes (T2D). Participants were classified into these latent groups according to shared characteristics. Class 1 (C1), representing 56.4% of the sample, included patients with the least reported problems. Class 2 (C2) and Class 3 (C3) comprised 16.5% and 21.7% of the participants, respectively.

Although few studies have applied latent class analysis (LCA) using similar variables, several comparable investigations are noteworthy. Liu et al. conducted an LCA among 1,200 T2D patients and identified three classes: 29.2% of participants were assigned to the “circadian disruption with daytime dysfunction” class, 11.4% to the “poor sleep status with daytime sleepiness” class, and 59.4% to the “favorable sleep status” class [21]. Similarly, Davis et al. identified four latent classes in their study: Class 1 (no symptoms, 65.7%), Class 2 (diabetes-related distress, 14%), Class 3 (subsyndromal depression, 12.6%), and Class 4 (major depression, 7.6%) [27]. Differences in class structure across these studies may stem from cultural attitudes toward mental health, varying levels of stigma, differences in QoL measurement tools, comorbidity profiles, and access to healthcare services. As such, interpretations across settings must be made cautiously, taking contextual factors into account. Ultimately, healthcare interventions should be tailored to meet the heterogeneous needs of diverse populations [10].

In another relevant study, Champion et al. identified three latent classes: Class 1 (“moderate risk,” 52%), Class 2 (“inactive, non-smokers,” 24%), and Class 3 (“smokers and current drinkers,” 24%). Participants in Class 3 exhibited higher levels of anxiety and depression compared to those in Class 1 [28]. Consistent with this, our findings indicated that Class 1 individuals in the current study had better outcomes across multiple domains, including higher QoL, better sleep quality, longer sleep duration (6–7 h), and an absence of generalized anxiety disorder. Conversely, Classes 2 and 3 exhibited greater impairment—Class 2 was associated with moderate QoL and poor sleep quality without anxiety, while Class 3 was characterized by moderate QoL, poor sleep quality, and severe anxiety—highlighting the need for targeted interventions in these groups.

This study also revealed a significant association between QoL and anxiety. Among participants with severe anxiety, 33% reported poor QoL and 58% reported moderate QoL. Furthermore, within Class 3—where severe anxiety was most prevalent—the probability of poor QoL was estimated at approximately 41%. Previous research corroborates this relationship. For instance, one survey found that 50% of individuals with diabetes reported poor QoL, with an anxiety prevalence of 71.7% [10]. Khan et al. reported anxiety in 50.7% and depression in 49.2% of T2D patients, both significantly associated with diminished QoL [29]. Similarly, Obo et al. confirmed the association of anxiety and depression with poor QoL [30]. Abualhamael et al. recommended comprehensive interventions including stress management, psychological counseling, sleep hygiene education, dietary regulation, and physical activity to mitigate the adverse effects of mental health disorders on glycemic control and QoL [10]. Schram et al. emphasized that depression imposes functional limitations in T2D patients, impairing self-management capabilities. They noted, however, that the predominance of cross-sectional designs in the literature limits causal inference, underscoring the need for longitudinal studies to better understand these associations [31].

Additionally, our findings demonstrated that among those with poor sleep quality, 26.4% reported poor QoL, while 54.4% reported moderate QoL. Within Class 3, the probability of poor sleep quality was high, and correspondingly, the likelihoods of poor and moderate QoL were 41% and 51%, respectively. These results align with previous studies. Jeong et al., studying Korean T2D patients in the U.S., reported a negative correlation between sleep quality and QoL [32]. Lou et al. similarly found that poor sleep quality in Chinese T2D patients was significantly associated with lower QoL [33]. Laverty et al. (2023), in a systematic review, confirmed a direct and significant relationship between sleep quality and QoL in individuals with T2D [14]. Poor sleep quality has been linked to increased fatigue, daytime sleepiness, mood disturbances, reduced self-care capacity, and overall lower QoL [34]. One proposed mechanism is dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which can lead to mood disorders and impaired glycemic control, thereby further compromising QoL in diabetic populations [35].

Limitations

This study’s limitation is that the sample was only taken in Ardabil city, not in the population of type 2 diabetes patients in Iran. Therefore, it is suggested that studies on type 2 diabetes patients be conducted in other provinces. In addition, the data were self-reported, and therefore there may be bias in reporting. Finally, the study’s cross-sectional nature makes it impossible to establish a causal relationship.

Conclusion

In this study, latent classes in T2D patients were investigated based on quality of life, sleep quality, and anxiety patterns. In general, the results of the present study showed a cooccurrence of problems in class C2 (16.5%) and C3 (27.1%). Crucial effective interventions are required to improve the health status of T2D patients. Based on this classification, appropriate psychological interventions such as stress and anxiety management and sleep hygiene education will be beneficial for classes 2 and 3. Also, it seems that identifying psychological and behavioral clusters in chronic diseases such as type 2 diabetes can help with appropriate interventions for other endocrine diseases that reduce the quality of life in individuals. The results of this study can be used in prioritizing health programs for these patients as well as emphasizing high-risk groups.

Abbreviations

LCA

Latent class analysis

AIC

Akiaik information criterion

BIC

Bayesian information criterion

Author contributions

Authors FR, DA, SM, NA, and AAG designed and conducted the study. The analysis of data was also performed by AAG, SA and RF. Authors RF, SA, and AAG wrote the first draft of the paper and conducted the literature review. Authors NA, SM, and DA completed and edited the manuscript. All authors approved the final manuscript.

Funding

This research was supported by grant from Ardabil University of Medical Sciences. This university had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Ardabil University of Medical Sciences (IR.ARUMS.REC.1402.112). Permission to conduct the study was obtained from this committee. All staff had signed an informed consent form.

Consent for publication

No applicable.

Clinical trial number

Not applicable.

Competing interests

The authors declare no competing interests.