Designing a Consumer-centric Care Management Program by Prioritizing Interventions Using Deep Learning Causal...

Area of Science:

• Health Services Research
• Artificial Intelligence in Healthcare
• Patient Engagement Strategies

Background:

• Care management programs aim to reduce health risks and improve patient outcomes.
• Post-Discharge Management (PDM) at Elevance Health targets 30-day readmission risk but faces low patient engagement.
• Current intervention selection relies on case managers' limited experience.

Purpose of the Study:

• To analyze the impact of initial interventions on patient engagement in PDM.
• To provide data-driven recommendations for case managers to enhance patient participation.
• To leverage deep learning causal inference for intervention optimization.

Main Methods:

• Deep learning causal inference was employed to assess intervention effects.
• Analysis focused on interventions conducted during the first post-discharge call.
• Three cross-validating experiments were conducted to ensure result reliability.

Main Results:

• Interventions requiring significant nurse-patient interaction on the first call increase consumer engagement.
• Less interactive, more technical interventions on the first call correlate with lower engagement.
• Findings align with clinical intuition and prior research on patient engagement.

Conclusions:

• Prioritizing interactive interventions in early post-discharge calls is crucial for improving PDM program engagement.
• Data-driven insights can guide case managers in selecting effective interventions.
• Optimizing initial patient contact enhances the overall success of care management programs.