Sleep and Emotion Symptoms Split Adults With Obesity Into Four Distinct Classes

Adults living with obesity do not experience the psychological burden of their condition in one uniform way. A new cross-sectional study from Wuhan, China, published in BMC Psychology, has used latent class analysis to show that sleep disturbance, anxiety, and depressive symptoms combine into four distinct symptom patterns within this population, and that these patterns are linked in measurable ways to weight self-stigma, social support, and eating self-efficacy. The findings, drawn from 469 adults recruited at two tertiary hospitals between March and December 2025, suggest that clinicians treating obesity may need to assess sleep and mood symptoms jointly rather than in isolation.

The research team, led by Gao-song Guo and Jia-xin Han of Hubei University of Chinese Medicine, together with colleagues from the Hubei Provincial Hospital of Traditional Chinese Medicine, set out to answer a deceptively simple question: do sleep problems and emotional symptoms cluster together in recognizable subgroups among adults with obesity, and if so, do those subgroups differ in their psychological resources? Rather than treating each symptom dimension as a separate score, the investigators applied a person-centered statistical technique that sorts individuals into classes based on their combined response patterns, an approach increasingly favored in psychiatric epidemiology because it captures the heterogeneity of real patients.

Technically, the analysis drew on seven dichotomized components of the Pittsburgh Sleep Quality Index, a validated instrument covering domains such as subjective sleep quality, sleep latency, sleep duration, and daytime dysfunction, alongside dichotomized subscales of the Hospital Anxiety and Depression Scale. Latent class analysis then searched for the smallest number of classes that could reproduce the observed joint distribution of these nine indicators. The preferred solution identified four classes with high classification precision, reflected in an entropy statistic of 0.923, where values approaching 1.0 indicate that individuals are assigned to their most likely class with little ambiguity.

The four classes were far from equal in size. The largest, comprising 56.5 percent of the sample, was a low-symptom group with minimal sleep or mood pathology. The remaining 43.5 percent fell into three higher-symptom classes: 18.6 percent showed sleep disturbance accompanied by daytime impairment, 12.6 percent displayed depressive symptoms with daytime impairment, and 12.4 percent exhibited a combined high-symptom profile spanning sleep disturbance, anxiety, and depressive symptoms simultaneously. In other words, roughly one in eight adults with obesity in this clinic-based sample carried the heaviest combined burden of sleep and emotional symptoms.

One of the most technically important results concerned the fragility of the class structure itself. When the researchers removed daytime dysfunction from the indicator set, the preferred solution shifted from four classes to three, meaning that the daytime impairment component was doing substantial work in separating the sleep-disturbance class from the depressive-symptom class. The authors are explicit that this sensitivity analysis means the four-class structure should not be regarded as a fixed clinical classification tool. It is a description of the sample’s symptom architecture, one that depends on which components of sleep quality are included in the model.

The study then asked whether membership in these classes was associated with psychosocial variables, using exploratory hierarchical multinomial logistic regression with the low-symptom class as the reference category. The strongest and most consistent finding involved weight self-stigma, the internalized experience of shame and self-blame related to body weight. Each additional point on the weight self-stigma measure was associated with roughly 10 percent higher odds of belonging to the combined high-symptom class rather than the low-symptom class, with an odds ratio of 1.100 per point and a 95 percent confidence interval of 1.047 to 1.156 after full adjustment for demographic and clinical covariates.

Social support showed an inverse pattern across all three higher-symptom contrasts: higher perceived support was associated with lower odds of membership in the sleep-disturbance, depressive-symptom, and combined high-symptom classes alike. The authors caution, however, that the social support measure used in the study showed limited internal consistency, meaning its items did not hang together as reliably as ideal psychometrics would demand, so this exploratory association warrants replication with a stronger instrument. Eating self-efficacy, the confidence in one’s ability to resist maladaptive eating, was associated with lower odds of depressive-symptom class membership specifically, while exercise self-efficacy showed no statistically significant adjusted association with any class.

Demographic variables also differentiated the classes. Male sex and younger age were associated with lower odds of belonging to the sleep-disturbance class, and male sex and no alcohol use were associated with lower odds of combined high-symptom membership. To guard against the statistical uncertainty inherent in assigning individuals to latent classes, the team repeated the principal analyses using the R3STEP approach, a sensitivity method that accounts for classification error when estimating the associations between covariates and most-likely class membership. The core psychosocial findings, particularly those involving weight self-stigma and social support, held up under this more conservative procedure.

What makes these results compelling is the convergence of sleep science and obesity psychology. Sleep deprivation and fragmentation are known to alter appetite-regulating hormones, heighten emotional reactivity, and erode the self-regulatory capacity that successful weight management requires. At the same time, weight self-stigma operates as a chronic psychological stressor, and the stress physiology it engages can disrupt sleep architecture and deepen depressive symptoms. The cross-sectional design of this study cannot disentangle the direction of these arrows, and the authors are careful to state that the findings do not establish causal relationships. A person stigmatized by their weight may sleep poorly because of that stress, or sleep poorly and then feel more stigmatized, or both processes may reinforce each other in a loop.

The practical implication is that assessment in obesity care may benefit from differentiation. A patient whose profile is dominated by sleep disturbance with daytime impairment may need a different intervention pathway than one whose profile centers on depressive symptoms, even if both present with the same body mass index. Screening for sleep quality and mood symptoms together, and paying attention to psychosocial resources such as social support and eating self-efficacy, could help identify which adults with obesity are carrying the highest combined symptom burden. The authors suggest these exploratory cross-sectional findings may inform further evaluation of differentiated assessment approaches, while emphasizing that the four-class solution’s dependence on the daytime dysfunction component means any clinical application must await validation in independent, ideally longitudinal, samples.

Subject of Research: Associations of sleep–emotion symptom classes with weight self-stigma and psychosocial resources among adults with obesity

Article Title: Associations of sleep–emotion symptom classes with weight self-stigma and psychosocial resources among adults with obesity

Article References: Guo, G.-S., Han, J.-X., Chen, F., Xiao, M.-Z., Zhang, Q., Wei, S.-S., Lu, L.-B., Li, Y., Zheng, J.-J., & Cai, Y. (2026). Associations of sleep–emotion symptom classes with weight self-stigma and psychosocial resources among adults with obesity. BMC Psychology. https://doi.org/10.1186/s40359-026-05820-8

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05820-8

Keywords: obesity, sleep, mental health, latent class analysis, weight self-stigma, Associations, emotion, symptom, classes, weight, self-stigma, psychosocial