Clinical Data for Predicting Platelet-Rich Plasma Therapy Response in Knee Osteoarthritis

A retrospective study of 102 knee osteoarthritis patients who underwent platelet-rich plasma therapy. The dataset includes 33 clinical features such as anthropometrics and blood indices, used to predict a pain score reduction of ≥2 at 6 months. The data was created by Haokang Zhang and last updated on 2026-04-14.

Use Cases

• Predicting clinical response to platelet-rich plasma therapy based on pre-treatment anthropometric and blood index data.
• Identifying key biomarkers like osmotic pressure, lipoprotein(a), and uric acid for treatment efficacy using SHAP analysis.
• Training and validating gradient boosting classifiers for binary outcome prediction in a clinical cohort.
• Benchmarking machine learning models for treatment response prediction using a dataset balanced via SMOTETomek.

Strengths

• Dataset includes 102 patient records with 33 retained clinical features after preprocessing.
• The final predictive model achieved a reported accuracy of 90.3% and an AUC of 0.862 on a held-out test set.
• Data preprocessing steps including imputation, transformation, and balancing are explicitly documented.

Limitations

• Row count for the underlying data file is unknown, limiting scale assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small at 24.7 KB, indicating limited scope.

Provenance

Source

figshare, authored by Haokang Zhang.

Collection Method

Retrospective review of multidimensional clinical data from patients who underwent PRP therapy.

Freshness

Last updated 2026-04-14 04:14:46; freshness should be verified.