Using a machine-learning alert to prompt palliative care for high-risk pediatric cardiology inpatients
Early PACT Involvement in Cardiology Patients Using Machine Learning
This project will test whether a daily machine-learning alert that predicts serious cardiac events can help palliative care (PACT) reach high-risk pediatric cardiology inpatients within three months.
Quick facts
| Phase | Not applicable |
|---|---|
| Study type | Interventional |
| Enrollment | 1000 (estimated) |
| Ages | N/A to 18 Years |
| Sex | All |
| Sponsor | The Hospital for Sick Children Academic / other |
| Locations | 1 site (Toronto) |
| Trial ID | NCT06886529 on ClinicalTrials.gov |
What this trial studies
Researchers will deploy an ML model that runs each morning on SickKids' curated EHR dataset (SEDAR) to flag pediatric cardiology inpatients at high risk of a serious cardiac event within three months. The flagged patients will generate notifications to the PACT team when appropriate, and patient outcomes during a 12-month post-deployment period will be compared with a 12-month pre-deployment baseline. Primary measures include rates and timing of PACT consultations or visits, ICU deaths, and documentation of goals of care. The analysis excludes patients expected to be discharged the day of admission and focuses on inpatient cardiology admissions at The Hospital for Sick Children.
Who should consider this trial
Good fit: Pediatric cardiology inpatients at SickKids who remain hospitalized beyond the day of admission and who are identified by the ML model as at high risk for a serious cardiac event are the intended participants.
Not a fit: Patients discharged the same day, those not identified as high risk by the model, or patients already recently managed by PACT are unlikely to gain additional benefit from the intervention.
Why it matters
Potential benefit: If successful, this approach could lead to earlier palliative care involvement for at-risk children, improving symptom management, clearer goals-of-care discussions, and potentially reducing ICU deaths.
How similar studies have performed: Machine-learning risk prediction has shown promising results in other inpatient settings, but using ML specifically to trigger palliative care in pediatric cardiology is relatively novel.
Eligibility criteria
Show full inclusion / exclusion criteria
Inclusion Criteria: * Pediatric inpatients admitted to cardiology Exclusion Criteria: * Expected to be discharged prior to midnight on the day of admission
Where this trial is running
Toronto
- The Hospital for Sick Children — Toronto, Canada (Recruiting)
Study contacts
- Principal investigator: Lillian Sung, MD, PhD — The Hospital for Sick Children
- Study coordinator: Lillian Sung, MD, PhD
- Email: lillian.sung@sickkids.ca
- Phone: 4168135287
How to participate
- Review the eligibility criteria above with your treating physician.
- Visit the official trial page on ClinicalTrials.gov for the most current contact information and recruitment status.
- Contact the listed study coordinator or principal investigator to request pre-screening. Pre-screening is free and never obligates you to enroll.