Predicting One-Year Mortality in Elderly Hip Fracture Patients Using Clinical Data

2,604 elderly patient records from Sheba Medical Center in Israel between 2017 and 2023 were used to develop and compare machine learning models for predicting one-year mortality. The dataset includes clinical, demographic, perioperative, and laboratory variables. Adi Shuchami published the study on figshare in 2026.

Use Cases

• Benchmarking automated machine learning pipelines against manually optimized models based on clinical prediction tasks.
• Identifying key mortality predictors like serum albumin and urea levels from patient health records.
• Developing interpretable clinical risk models using SHAP and permutation importance analyses described in the study.

Strengths

• Includes 2,604 patient records from a single medical center, providing a focused cohort.
• Models were validated using stratified 5-fold cross-validation and techniques like SMOTE to address class imbalance.
• Performance metrics for multiple algorithms are reported, including AUC, accuracy, precision, recall, and F1-score.

Limitations

• Row count for the underlying dataset is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect geographic and temporal bias inherent to a single hospital in Israel from 2017-2023.

Provenance

Source

Sheba Medical Center, Israel

Collection Method

Retrospective cohort study

Time Range

January 2017 to November 2023

Freshness

Last updated 2026-06-01 04:17:40; freshness should be verified.

Geography

Israel