Abstract
Objective
To assess cardiometabolic control among patients with type 2 diabetes mellitus (T2DM) in Shanxi Province and identify variables associated with poor cardiometabolic control.
Methods
Data on demographic characteristics, medical history, Knowledge–Attitude–Practice (KAP) questionnaire scores, and clinical measurements of blood pressure, glucose, and lipid levels were collected from seven hospitals in Shanxi Province between June 2024 and June 2025. Poor cardiometabolic control was defined as failure to achieve comprehensive control across the three cardiometabolic domains (blood pressure, glycemia, and lipid levels), with failure to meet the target in any single domain considered poor control. Variable selection was performed using the least absolute shrinkage and selection operator (LASSO) combined with bootstrap-based stability selection. Firth’s penalized logistic regression was then applied to the stably selected variables to estimate their associations with poor cardiometabolic control.
Results
Among 1,874 patients, only 2.88% achieved comprehensive cardiometabolic control. Most patients had marked deficits in knowledge of blood pressure and lipid management, as well as in self-monitoring and symptom recognition. Bootstrap stability selection combined with Firth regression identified six candidate variables. Among these, BMI, education level, and knowledge score exhibited the strongest, nominally significant associations with poor cardiometabolic control, whereas practice score, hypertension history, and T2DM duration were selected but did not reach nominal significance. These exploratory findings require validation in independent cohorts.
Conclusions
Cardiometabolic control among patients with T2DM in Shanxi Province remains suboptimal. Exploratory analyses suggest that BMI, education level, and knowledge score are the variables most robustly associated with poor cardiometabolic control, whereas practice score, hypertension history, and T2DM duration represent additional, weaker candidate associations warranting further investigation. These findings may inform future diabetes self-management education and intervention strategies aimed at improving glycemic, blood pressure, and lipid control. However, given the cross-sectional and exploratory nature of this study, the results should be regarded as hypothesis-generating and require confirmation in prospective or interventional studies before causal inferences can be drawn.