AI in critical care, from the device to the record
What it can do, what it cannot do yet, and whether the thing in front of you is real.
I spent forty years connecting medical devices to the record, at Alaris, CareFusion and BD, and on the committees that wrote the standards. I spent the last year building live reasoning applications on a frontier model and writing down, issue by issue, what worked and what did not. I take a few executive advisory engagements a year with people who can use that.
For investors
Due diligence on AI claims in medical devices and clinical software. Does the model do what the deck says, on whose patients, with what oversight? What is the regulatory path, and what do the state laws now arriving monthly do to it? What does it cost to make it real, and when does it pay?
For device and health IT executives
Connectivity that holds up: ISO/IEEE 11073, IHE Devices, HL7 and FHIR, and the places integrations fail between the pump and the chart. Oversight a board can stand behind, built as stages an auditor can check. And a working demonstration when an argument needs settling.
How it works
- Executive level: the CEO, the board, the investment committee.
- A quarterly retainer for a set number of days, renewed by agreement.
- A few engagements a year, chosen for fit.
- A written letter and a conflict check first. My own inventions stay out of the room.
Record
Co-inventor on US patent 7,375,647, with five provisional applications filed on the reasoning method behind Ai MedAgent™. AAMI BI&T Best Article 2014, Worth the Effort? Closed-Loop Infusion Pump Integration with the EMR. Co-author of the Lifescan EEG algorithm, J Clin Monit 1986. Contributor to IHE Patient Care Device standards and the EMRAM medication criteria work with HIMSS Analytics. Founding member of the iMetrikus team. Author of The Technology Was Never the Problem and Inside the Loop.