Diabetic Foot Ulcer Pre-Screening Model with Clinical Indicators, 2012-2025

A clinical prediction model for diabetic foot ulcer identification was developed and validated using data from 1,854 hospitalized and 678 outpatient type 2 diabetes patients. The model uses four predictors: age, history of injury, alcohol consumption, and the albumin-to-glycated hemoglobin ratio. Author Weidi Wang published the model on figshare in 2026 under a CC-BY-4.0 license.

Use Cases

• Developing a simplified pre-screening tool for diabetic foot ulcers based on routinely available clinical indicators.
• Validating a predictive model using the albumin-to-glycated hemoglobin ratio as a composite marker of nutritional status and glycemic control.
• Analyzing synergistic risk interactions between patient age and the Alb/HbA1c ratio for diabetic foot ulcers.
• Assessing model performance metrics like AUC, calibration, and Brier score for clinical prediction models.

Strengths

• Model was validated on a distinct outpatient cohort of 678 patients, demonstrating external validity.
• Performance metrics are reported, including an AUC of 0.817 (95% CI 0.782–0.852) in the validation set.
• The dataset underpinning the model includes a training set of 1,854 patients with a 20.9% DFU prevalence.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying clinical data is unknown, which may limit suitability assessment.
• Data may reflect geographic and temporal bias inherent to a single hospital system.

Provenance

Source

figshare, author Weidi Wang.

Collection Method

Retrospective analysis of hospitalized patient records and outpatient data from a hospital system.

Time Range

Training data from January 2012 to December 2020; validation data from January 2021 to December 2025.

Freshness

Last updated 2026-05-29 04:47:38; freshness should be verified.

Geography

Data originates from a geographically distinct branch of a hospital system; specific location is unknown.