What happened
Fifth Eye of Ann Arbor, Michigan, makes the Analytic for Hemodynamic Instability (AHI), software trained on ECG data. Every two minutes it classifies five minutes of a lead-II ECG as stable or unstable. Unstable means signs of low blood pressure combined with a heart rate of 100 or more. Fifth Eye ran a prospective observational study at the University of Michigan Medical Center from November 2019 to January 2020. It included 222 patients in intensive care and emergency units with both ECG and arterial line monitoring. FDA reviewers report a sensitivity of 95.6% and a specificity of 84.9% per window, above goals of 85% and 75%. FDA granted the De Novo request on 1 March 2021. The reviewers write that patients who need more care may be identified earlier with the device. They note that only patients with arterial lines could be studied, and these were sicker than the intended population. About one in four patients had frequent false positives. The study was observational, so the outputs guided no care.
Sources
-
Regulatory document, by authors with no tie to the system.
U.S. Food and Drug Administration, Center for Devices and Radiological Health. De Novo Classification Request for Analytic for Hemodynamic Instability (AHI) (DEN200022): decision summary. Granted 2021-03-01.
The observed sensitivity (95.6%) in the study means there were few false negatives (Type-II error), and thus fewer instances of missed hemodynamic instability.
Details
AI system
- Workflow
-
The software analyzed a continuous lead-II ECG and checked signal quality before each window. Arterial line pressure and heart rate served as the reference for each window.
Follow-up
- Conditions for replication
-
The patient is an adult on continuous ECG monitoring without sustained arrhythmia or pacemaker dependence. Clinicians confirm an unstable output with other measurements before they act.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Fifth Eye built AHI to identify adults on ECG monitoring in hospital who show signs of hemodynamic instability, so that clinicians can watch them more closely.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). AHI flagged unstable circulation from a single ECG lead (AIBD-0783). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0783.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)