| Protocol ID | NCT06886529 |
| Registration Date | 2025-03-20 |
| Status | recruiting |
| Sponsor | The Hospital for Sick Children |
| Country | Canada |
| Industry | not industry |
| Blinding | unblinded |
| Center Type | singel-center RCT |
| Scope | national RCT |
| First Patient In | 2025-10-16 |
| Last Patient Out | 2027-10-16 |
| Publication | n |
| Gender % | ALL |
| Mean Age | - 18 Years |
NCT06886529
Early PACT Involvement in Cardiology Patients Using Machine Learning
| Field |
GLM
glm-5.1 7364ms / 8861 tok |
CLAUDE
claude-sonnet-4-20250514 Thinking: claude-opus-4-20250514 6031ms / 10100 tok |
OPENAI
gpt-4o Thinking: o3 3226ms / 8726 tok |
Review |
|---|---|---|---|---|
| rct_type | N/A | N/A | N/A | OK |
| cancer_status | non-cancer | non-cancer | non-cancer | OK |
| non_cancer_spec | cardiac | cardiac | cardiac | OK |
| setting_patient | inpatient | inpatient | inpatient | OK |
| setting_place | hospital | hospital | hospital | OK |
| feasibility |
N/A
N/A → N/A
shaky reasoning✓ Unanimous
Initial: N/A
Requery: N/A — This is a pre-post observational study comparing outcomes before and after deployment of an intervention without randomization or a concurrent control group.
|
N/A
fully powered RCT → fully powered RCT → N/A
shaky reasoning✓ Unanimous
Initial: fully powered RCT
Requery: fully powered RCT — This is a fully powered implementation study with 1000 participants designed to evaluate clinical effectiveness outcomes of an ML intervention on PACT consultation rates and patient outcomes.
Thinking: N/A — Single-group pre-post intervention study without randomization, not an RCT.
|
N/A
fully powered RCT → N/A → N/A
shaky reasoning✓ Unanimous
Initial: fully powered RCT
Requery: N/A — This study is not a randomized controlled trial as it involves a pre- and post-intervention comparison.
Thinking: N/A — This is a single-group pre/post interventional study without randomization, so it is not an RCT.
|
OK |
| intervention_pharmacologic | no | no | no | OK |
| intervention_chemo_immuno_radio | no | no | no | OK |
| intervention_complementary | no | no | no | OK |
| intervention_device | no | no | no | OK |
| intervention_psychological | no | no | no | OK |
| intervention_complex | no | no | no | OK |
| intervention_acp |
no
no → no
shaky reasoning✓ Unanimous
Initial: no
Requery: no — The trial's core intervention is an ML-based screening tool to predict cardiac risk and prompt palliative care referrals, rather than delivering a direct, structured Advance Care Planning intervention.
|
no
no → no
shaky reasoning✓ Unanimous
Initial: no
Requery: no — The intervention is deploying a machine learning model to identify high-risk cardiac patients for potential PACT (palliative care) referral, but does not involve structured advance care planning conversations or tools as the core intervention component.
|
no
yes → no
shaky reasoning✓ Unanimous
Initial: yes
Requery: no — The trial focuses on ML model predictions for cardiac events with goals of care as a secondary outcome, not as a core component of ACP.
|
OK |
| intervention_music | no | no | no | OK |
| intervention_screening |
yes
yes
shaky reasoning✗ No consensus
Initial: yes
|
yes
yes → yes → no
shaky reasoning✗ No consensus
Initial: yes
Requery: yes — The ML model is a screening tool that systematically identifies high-risk cardiac patients who may benefit from palliative care involvement, and the effectiveness of this screening intervention itself is being evaluated.
Thinking: no — The ML model is used to identify high-risk patients as part of a larger intervention to increase PACT consultations, not to test the screening tool itself.
|
no
no → no → yes
shaky reasoning✗ No consensus
Initial: no
Requery: no — The trial tests a machine learning model to predict serious cardiac events as part of a supportive care strategy, not a screening/assessment tool to identify palliative care needs.
Thinking: yes — The trial evaluates a machine-learning model used to screen cardiac inpatients for high risk to prompt earlier palliative care involvement, i.e., the act of screening itself is the intervention.
|
|
| intervention_palliative |
no
no → no
shaky reasoning✗ No consensus
Initial: no
Requery: no — The intervention being tested is a machine learning prediction model, with palliative care being the downstream outcome and setting rather than the direct experimental intervention.
|
yes
yes → yes → no
shaky reasoning✗ No consensus
Initial: yes
Requery: yes — The intervention is a machine learning model designed to trigger early PACT (palliative care team) consultation in high-risk cardiac patients, making palliative care referral the primary intervention being tested.
Thinking: no — The intervention is an ML model predicting cardiac events; palliative care consultation (PACT) is an outcome being measured, not the intervention itself.
|
yes
yes → yes → yes
shaky reasoning✗ No consensus
Initial: yes
Requery: yes — The trial is designed to increase early PACT (Palliative Care Team) involvement, which serves as a palliative care intervention.
Thinking: yes — ML alerts are specifically designed to trigger earlier PACT (palliative care) consultations, so the intervention is a palliative care referral mechanism.
|
|
| intervention_education | no | no | no | OK |
| intervention_digital | yes | yes | yes | OK |
| intervention_other | no | no | no | OK |
| outcome_symptom | no | no | no | OK |
| outcome_qol | no | no | no | OK |
| outcome_dignity | no | no | no | OK |
| outcome_cost | no | no | no | OK |
| outcome_other | yes | yes | yes | OK |
| pharmacologic_sub | N/A | N/A | N/A | OK |