Using Machine Learning to Support Early Palliative Care Involvement in Pediatric Cardiology Patients

What this Trial Is About

This research aims to evaluate the effectiveness of a machine-learning ML model in predicting serious cardiac events within three months among pediatric cardiac inpatients. The study focuses on determining whether the ML model increases palliative care team PACT consultations, reduces time to consultation, decreases ICU deaths, and improves documentation of care goals. The study is conducted at The Hospital for Sick Children, leveraging their strong collaboration between cardiology and palliative care. The ML model identifies high-risk cardiology patients daily. If a patient was seen by PACT in the past year, the update goes to the PACT team otherwise, the cardiology physician in charge receives the alert to decide on consultation. The study compares patient outcomes before and after deploying the ML model, with a pre-deployment period covering 12 months starting 15 months before deployment, and a post-deployment period covering 12 months beginning 3 months after deployment. Participants are pediatric cardiology inpatients. Researchers will monitor PACT consultations within three months of admission, time to consultation, ICU death rates, and documentation of care goals. Outcomes among identified high-risk patients will be compared between the pre- and post-deployment periods to assess the ML models impact on care processes and patient outcomes.

PACT Involvement in Cardiology Patients