Key result
Hierarchical models incorporating patient subgroups slightly improve 30-day readmission prediction for COPD and THA/TKA.
Why the study?
A considerable gap exists between the identification of patient subgroups and their modeling and interpretation for clinical applications.
Does the MIPS framework improve the prediction of 30-day hospital readmission in patients with COPD, CHF, and THA/TKA compared to standard logistic regression?
Population
Patients with COPD (n=29,016), CHF (n=51,550), and THA/TKA (n=16,498)
Comparison
Hierarchical prediction using subgroup membership vs standard logistic regression without subgroup membership
Design
Matched case-control study
Authors
Loading...
Warrants no immediate change in readmission prediction; leaves open value of subgroup-informed models pending validation.
Case-Control (n=97,064)
Does the MIPS framework improve the prediction of 30-day hospital readmission in patients with COPD, CHF, and THA/TKA compared to standard logistic regression?
Effect estimate: Net reclassification improvement 0.059 and 0.11
The MIPS framework successfully identified clinically meaningful patient subgroups and achieved high classification accuracy, but provided only small improvements in predicting hospital readmission based on comorbidities alone.
Bhavnani et al. (2022) conducted a case-control in Hospital readmission (COPD, CHF, THA/TKA) (n=97,064). Hierarchical prediction model using patient subgroup membership vs. Standard logistic regression without subgroup membership was evaluated on Hospital readmission within 30 days (Net reclassification improvement 0.059 and 0.11). A hierarchical prediction model incorporating patient subgroup information yielded a small but statistically significant improvement in predicting hospital readmission for COPD and THA/TKA (NRI 0.059 and 0.11).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: