Predictors of social risk for post-ischemic stroke reintegration

Area of Science:

• Neuroscience
• Social Science
• Health Services Research

Background:

• Successful reintegration post-stroke is vital for patients' quality of life.
• Social risks, including inadequate support systems, significantly impact long-term outcomes.
• Understanding these social risks is critical for developing effective interventions.

Purpose of the Study:

• To predict social risk at rehabilitation discharge using admission data.
• To identify key demographic, clinical, and social factors contributing to social risk.

Main Methods:

• Applied Gradient Boosting decision tree modeling to a cohort of 217 ischemic stroke survivors.
• Utilized 16 variables including sociodemographic, clinical, and social support data.
• Employed data subsampling techniques to address class imbalance for robust model training.

Main Results:

• Models demonstrated strong performance with high AUC values for training (0.831-0.843) and validation (0.881-0.909).
• Social support and economic status were identified as the most significant predictors of social risk.
• Sex and age also contributed to social risk prediction, highlighting a multifactorial nature.

Conclusions:

• Predictive models for social risk in stroke survivors are effective and reliable.
• Social support and economic factors are primary drivers of social risk, influencing reintegration.
• Combined factors, including social and economic aspects, collectively determine an individual's social risk profile.