Frontiers in Public Health · Oct 21, 2026
Community-based identification of poor sleep quality in middle-aged and older adults with abdominal obesity using psychological and physical fitness indicators
By Qiao C, Yang S, Wang Y et al.
Background: Poor sleep quality, abdominal obesity, psychological symptoms, and declining physical function may cluster during midlife and later life, posing a multidimensional challenge to healthy aging. We aimed to develop and internally validate a community-oriented nomogram integrating readily obtainable psychological and physical fitness indicators to identify current poor sleep quality in community-dwelling middle-aged and older adults with abdominal obesity and to evaluate its added value over routinely available demographic and health information. Methods: This retrospective cross-sectional study analyzed data from 212 adults aged ≥45 years with abdominal obesity assessed at a community health service center in Beijing, China. Poor sleep quality was defined as a Pittsburgh Sleep Quality Index (PSQI) score >5. Ten candidate predictors were evaluated using least absolute shrinkage and selection operator logistic regression with repeated stratified 10-fold cross-validation, followed by multivariable logistic regression. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration measures, the Brier score, and decision curve analysis. Separate 1,000-resample bootstrap procedures assessed variable-selection stability and optimism in the performance of the fixed seven-predictor model. Results: Among the 212 participants, 117 (55.2%) had poor sleep quality. The final model included age, the 9-item Patient Health Questionnaire, the 7-item Generalized Anxiety Disorder scale, grip strength, 30-s chair stand repetitions, chair sit-and-reach distance, and 2-min step count. The apparent AUC was 0.779 [95% confidence interval (CI), 0.718–0.840]. After bootstrap optimism correction, the AUC, calibration slope, and Brier score were 0.751, 0.858, and 0.206, respectively. Compared with the routine-information model (apparent AUC, 0.648), the final model improved the AUC by 0.131 (95% CI, 0.050–0.213; p = 0.002) and showed greater theoretical net benefit across threshold probabilities of 13–80% in the development sample. Conclusion: The nomogram showed moderate internal performance and higher apparent performance than the routine-information model, although content overlap between PHQ-9 and PSQI may have inflated these estimates. Where psychological questionnaires and field-based fitness assessments are already available, it may support targeted sleep assessment in community healthy-aging programs. External validation is required before implementation in public health practice.