Can Patient-Reported Outcome Measures Help Predict Unplanned Hospital Readmission?

Area of Science:

• Health Services Research
• Clinical Epidemiology
• Health Informatics

Background:

• Administrative data often omit crucial patient symptom and functional status information for predicting hospital readmissions.
• Patient-reported outcome measures (PROMs) offer a valuable supplement to administrative data.
• Integrating PROMs can potentially enhance the accuracy of readmission risk prediction models.

Purpose of the Study:

• To evaluate the added predictive capability of PROMs for unplanned hospital readmissions.
• To inform post-discharge monitoring strategies and ongoing patient care management.
• To assess the incremental value of specific PROMs (EQ-5D-5L, VR-12) in readmission prediction.

Main Methods:

• A population-based retrospective cohort study utilizing linked administrative and PROMs data from British Columbia, Canada.
• Inclusion of adult patients discharged from acute care who completed EQ-5D-5L and VR-12 surveys.
• Application of Cox proportional hazards models to predict readmission risk at 30, 180, and 360-day horizons, with subgroup analysis for Ambulatory Care Sensitive Conditions (ACSCs).

Main Results:

• Observed unplanned readmission rates were 5.6% (30-day), 18.4% (180-day), and 25.0% (360-day).
• PROMs modestly improved predictive discrimination across all time horizons.
• For the 180-day horizon, adding VR-12 to administrative data increased the C-index from 0.762 to 0.782; EQ-5D-5L increased it to 0.774.

Conclusions:

• PROMs significantly enhance the predictive value for unplanned hospital readmissions.
• The inclusion of PROMs is particularly beneficial for patients with Ambulatory Care Sensitive Conditions (ACSCs).
• These findings support the routine integration of PROMs into risk prediction models for improved patient management.