Exploratory associations of psychological disorders with heart failure among adults with diabetes in NHANES

Background Diabetes mellitus (DM) with heart failure (HF) is a serious public health problem, and its occurrence and progression are related to a variety of factors, including lifestyle and sociodemographic characteristics. Recent studies have shown that psychological factors (such as panic disorder, depression, and anxiety) may be associated with an increased disease prevalence through physiological and behavioral mechanisms, but the specific association still needs to be further explored. Based on data from the National Health and Nutrition Examination Survey (NHANES), this study aimed to explore the association between psychological disorders and diabetes mellitus complicated with heart failure and to provide exploratory evidence for future research on integrated chronic disease management. Methods A total of 1503 participants from NHANES were included, including 29 patients with both diabetes mellitus and HF, 29 patients with DM only, 2 patients with HF only, and 1443 participants without either DM or HF. Sociodemographic information (age, gender, education level, marital status) and lifestyle data (smoking, alcohol consumption) were collected through family interviews. Python 3.8, SPSS 27.0, and RStudio 4.3.1 were used for statistical analysis, including descriptive statistics, logistic regression analysis, and a random forest multi-classification model. Logistic regression was used to evaluate the association between psychological factors and DM-HF, and the random forest model was used as an exploratory approach to assess feature importance rather than to establish a validated prediction model. The analyses were conducted without incorporating NHANES sampling weights, stratification, or clustering; therefore, the estimates should be interpreted as associations within the included analytic sample rather than nationally representative estimates. Results Based on NHANES data, this study explored the association between psychosocial factors and diabetes mellitus complicated with heart failure (DM-HF). The prevalence of panic disorder, depression, and anxiety disorder was higher in the DM-HF group ( P < 0.001). Logistic regression analysis showed that psychiatric problems, such as panic disorder (OR = 14.807; 95% CI 4.223–51.912), were positively associated with DM-HF. However, these estimates should be interpreted cautiously because of the limited number of DM-HF cases and the relatively wide confidence intervals. In the exploratory random forest analysis, alcohol consumption and psychological variables ranked among the features with relatively high importance. However, the apparent model performance should be interpreted cautiously because of the small number of DM-HF cases, class imbalance, and lack of internal validation. Conclusions Psychological disorders, including panic disorder, depression, and anxiety, were associated with DM-HF in this exploratory NHANES-based analysis. Alcohol consumption and psychological variables showed relatively high feature importance in the random forest model. However, because the analyses did not incorporate NHANES sampling weights, stratification, or clustering, the findings should not be interpreted as nationally representative estimates. These associations should be interpreted cautiously and require confirmation in larger, survey-weighted, and longitudinal studies.