Predicting Major Amputation Risk in Diabetic Foot Ulcer Patients with Temporal Validation

A clinical dataset used to develop interpretable machine learning models for predicting in-hospital major lower-extremity amputation risk in diabetic foot ulcer patients. The models were developed on a retrospective cohort from 2019–2020 and temporally validated on a later cohort from 2024. The dataset likely contains admission variables including comorbidities, limb/ulcer assessments, and laboratory tests.

Use Cases

• Developing predictive models for major amputation risk based on admission variables
• Comparing performance of logistic regression, elastic net, random forest, and XGBoost models
• Applying Shapley additive explanations for patient-level interpretability of predictions
• Validating clinical risk models using temporal external validation cohorts

Strengths

• Models achieved high discrimination with AUCs of 0.977 and 0.984 in internal testing and temporal validation
• Candidate predictors were routinely available admission variables within 24 hours, suggesting clinical practicality
• The dataset includes temporal external validation, which tests model performance on a later cohort

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• Data is from a single center, which may reflect geographic or institutional bias

Provenance

Source

figshare

Collection Method

Retrospective review of consecutive admissions for diabetic foot ulcers at a single center

Time Range

Development cohort 2019–2020, validation cohort 2024

Freshness

Last updated 2026-05-22 04:29:57

Geography

Single-center study; location is unknown