Knee Osteoarthritis Phenotypes and Racial Disparities in Pain and Surgery

Ahmad Alkhatatbeh published a retrospective cohort study on 2026-05-20. The dataset contains electronic health record and radiograph data for 33,553 adults aged 40+ with knee osteoarthritis, used to derive phenotypes and assess associations with race and incident total knee arthroplasty. It includes a subset of 20,030 participants with pain scores recorded near their index radiograph.

Use Cases

• Clustering patient phenotypes based on age, radiographic severity, pain, and comorbidities as described in the methods.
• Analyzing racial disparities in pain-dominant versus structural/metabolic osteoarthritis phenotypes as indicated in the results.
• Modeling time-to-event for total knee arthroplasty using Cox proportional hazards with features like radiographic severity and sex.
• Validating phenotype-aware approaches for studying disparities and guiding osteoarthritis care as discussed.

Strengths

• Dataset is derived from a cohort of 33,553 adults with knee osteoarthritis, providing a substantial sample size.
• Includes a defined subset of 20,030 participants with pain scores temporally linked to radiographs.
• Analysis identified three clinically distinct phenotypes (structural/metabolic, younger trauma-associated, pain-dominant).
• Released under a permissive CC-BY-4.0 license, facilitating reuse.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the final analytical subsets is unknown, which may limit suitability assessment.
• Data may reflect geographic or institutional bias inherent to the single-source Emory Knee Radiograph dataset.

Provenance

Source

Emory Knee Radiograph dataset.

Collection Method

Retrospective cohort study using electronic health record and radiograph data.

Freshness

Last updated 2026-05-20 05:38:36; freshness should be verified.