This is a prospective, multicenter, cluster randomised trial which will assess the feasibility and acceptability of AI-assisted point of care echocardiography on HF detection and downstream care processes for patients with suspected HF discharged from the Emergency Department.

Heart failure remains underdiagnosed in frontline care settings, in part due to limited access to timely echocardiography. Recent advances in artificial intelligence (AI)-assisted echocardiography may facilitate scalable bedside screening through automated image acquisition support and interpretation.

This is a prospective, multicenter, cluster randomised trial which will assess the feasibility and acceptability of routine implementation of AI-assisted point of care echocardiography on HF detection and downstream processes of care for patients with suspected HF being discharged from the Emergency Department.

This study will be carried out at 2 participating Emergency Departments (EDs) in Alberta, Canada. Each ED site will alternate monthly between Intervention (AI-assisted echocardiography) and Control (Usual Care) periods.

During intervention periods, patients with suspected HF will undergo AI-assisted echocardiography using a handheld point of care EchoNous Kosmos device by the treating ED physician. The US2.ai algorithm will generate an AI-automated echo report. During control periods, patients will be managed according to usual care.

Inclusion Criteria:

1. Adult patients age ≥18 years presenting to the emergency department (ED);
2. with signs or symptoms of HF as per the treating physician and
3. elevated natriuretic peptides (NT-proBNP >125pg/mL or BNP >50pg/mL), drawn as part of standard of care of evaluation in the ED.

Exclusion Criteria:

1. prior documented diagnosis of HF;
2. any patient receiving renal replacement therapy;
3. any patient receiving comfort or palliative care;
4. pregnant or nursing patients;
5. patient is planned for admission.