Statistical Analysis of Hip Hypermobility and Psychosocial Outcomes
Project Overview
This project investigates whether psychosocial factors such as pain catastrophizing, pain interference, and pain self-efficacy predict hip hypermobility (Beighton Score) in a clinical population. Leveraging robust statistical methods and reproducible analytics, the analysis provided actionable findings for clinicians and researchers. Using both multiple linear regression and multiple imputation, we found:
- Key predictors of hypermobility are age and gender, not psychosocial scores
- Younger females tend to have higher hypermobility
- Pain and self-efficacy scores showed little/no association with Beighton Score
Methodology
- Data: 445 patients from a Baylor Scott & White hip clinic with Beighton Score, age, gender, and three psychosocial surveys (PCS, PROMIS-PI, PSEQ)
- Outcome: Beighton Score (continuous; dichotomized for logistic regression)
- Predictors: Age, gender, pain catastrophizing (PCS), pain interference (PROMIS-PI), pain self-efficacy (PSEQ)
- Handling Missing Data: Used Multiple Imputation by Chained Equations (MICE) to address 47-64% missingness in psychosocial variables
- Statistical Analysis:
- Multiple Linear Regression (MLR) for continuous Beighton Score
- Post-hoc Logistic Regression for dichotomized outcome (>=4 vs <4)
- Age and gender included as covariates; all models checked for multicollinearity and validated via diagnostic plots
Key Results
- Age and gender were the only significant predictors:
- Younger females exhibited higher Beighton Scores (males scored ~3.7 points lower, age slope ~-0.06 per year)
- Psychosocial outcomes: No significant association found in either linear or logistic regression, regardless of missing data strategy
- Model diagnostics: No problematic multicollinearity; regression assumptions verified- No significant association found between psychosocial measures (PCS, PROMIS-PI, PSEQ) and hypermobility, regardless of imputation method
Tools & Technologies
- Languages: R
- Libraries: mice, broom, ggplot2, modelsummary
Visualizations & Report
- Regression coefficient plots, imputation diagnostics, confidence interval visualizations
- Full Project Report (PDF)
Interactive Dashboard
Explore all results and rerun analyses with new data:
Further Reading
Conclusion & Next Steps
- Psychosocial screening did not enhance risk assessment for hypermobility in this clinical sample.
- Future research: Apply methods to general-population data to clarify external validity; explore additional biopsychosocial predictors.