Harvard Dataverse hosts a dataset of patients diagnosed with ischemic stroke at Massachusetts General Hospital between 2003 and 2017. It combines clinical AF risk scores from the Recalibrated Cohorts for Heart and Aging Research in Genomic Epidemiology model with a polygenic risk score derived from 1,093,050 genetic variants. Patients are categorized into clinical and genetic risk tertiles for analysis.
Use Cases
- Train Cox proportional hazards models using clinical AF risk tertiles and genetic risk tertiles to predict atrial fibrillation onset after stroke.
- Assess the improvement in risk prediction models by calculating C indices and percentile-based net reclassification index after adding the AF polygenic risk score.
- Analyze the interaction between clinical risk tertiles and genetic risk tertiles across different follow-up windows for patient stratification.
Strengths
- Data originates from a well-defined cohort at a major academic hospital, Massachusetts General Hospital.
- Genetic risk assessment is based on a contemporary polygenic risk score incorporating 1,093,050 variants.
- Clinical risk is quantified using a validated model, the Recalibrated Cohorts for Heart and Aging Research in Genomic Epidemiology Atrial Fibrillation model.
- Analysis employs established statistical methods including Cox models and net reclassification index.
Limitations
- The dataset is limited to a single hospital's patient population, which may limit generalizability.
- Specific sample size, number of features, and data completeness are not provided in the input.
- Data collection ended in 2017, which may affect the relevance of genetic risk scores based on contemporary variants.
Provenance
- Source
- Harvard Dataverse, contributed by author Anderson, Christopher.
- Collection Method
- Retrospective study of ischemic stroke patients at Massachusetts General Hospital; clinical risk modeled, genetic risk estimated from a polygenic risk score.
- Time Range
- 2003 to 2017
- Freshness
- null
- Geography
- Massachusetts General Hospital, United States.