A research team in China developed an explainable AI model intended to forecast disability trajectories for hospitalized older adults with chronic back pain, aiming to provide predictions while patients are still in the hospital. The coverage positions the model as addressing a key operational issue in geriatric care: clinicians need early risk stratification to identify who is likely to decline in function and independence. Explainability is highlighted as a requirement for clinical usability rather than a purely retrospective score. While the excerpt does not provide performance metrics, the framing indicates a shift from symptom-focused documentation toward trajectory-aware prediction during acute care. For biotech and digital-health developers, this is another example of clinical AI focusing on actionable prediction windows—where model outputs can inform care planning and rehabilitation intensity.