Abstract
Background: Diabetes self-management is essential for type 2 diabetes care, but cross-sectional associations may be distorted when questionnaire scores are incorrectly derived or converted into arbitrary categories. We evaluated clinical and sociodemographic correlates of the continuous Diabetes Self-Management Questionnaire (DSMQ) total score in a secondary analysis of an outpatient clinical database. Methods: The analysis included 262 adults with type 2 diabetes recruited between 1 November 2025 and 1 May 2026. DSMQ total and subscale scores were independently recalculated from the 16 item-level responses using the published scoring algorithm. The primary analysis retained the DSMQ score as a continuous outcome. The database provided a participant-level count of documented glucose measurements > 250 mg/dL. This operational count was not treated as a standardized clinical event or severity category; individual measurement timestamps, participant-specific observation duration, and the total number of glucose measurements were unavailable. Results: The recalculated DSMQ score was 5.26 ± 2.35 (median 5.10; interquartile range 3.12–7.08). Among 261 participants with an interpretable sex code, 144 (55.2%) were women. In univariable analyses, lower DSMQ scores were concurrently associated with longer diabetes duration, a higher recorded severe-hyperglycemia count, higher HbA1c, and higher fasting glucose. In the adjusted model (n = 261), the recorded severe-hyperglycemia count remained negatively associated with DSMQ score (adjusted coefficient −0.412 points per recorded episode; 95% confidence interval −0.621 to −0.203; p < 0.001). Individualized versus hospital-based education was not independently associated with the corrected score. A beta-regression sensitivity analysis produced a concordant direction of association (coefficient −0.152 on the logit mean scale; p < 0.001). The linear model explained 17.2% of observed variance (adjusted R2 = 0.136), with an optimism-corrected R2 of 0.099. Conclusions: The recorded severe-hyperglycemia count was a concurrent correlate of lower DSMQ score. Because the study is cross-sectional and glucose-measurement counts were preaggregated without observation-time or monitoring-frequency data, the findings do not establish temporal direction, causality, or individual-level clinical prediction. Item-level score verification and continuous-outcome analysis materially changed the interpretation of the dataset.