AI MedAgent Smart Formulary and Order Management has two halves. Smart Formulary makes sure the formulary says the same thing to the order screen (CPOE), the pharmacy, the label and the pump. Smart Orders takes the next step: it reads the patient and drafts the order, using published pharmacokinetic models (the math of how a body moves a drug: how fast it arrives, where it spreads and how fast it clears). The clinician reviews and accepts every order. Synthetic patients only.
The open-source alignment check inside Smart Formulary. Drop in a hospital continuous-infusion formulary and get back the findings that stand between it and infusion pump auto-programming. Runs in your browser; nothing is uploaded. Apache 2.0.
Check a formularyReconcile the formulary against a pump drug library export, care area by care area, and see every entry that will auto-program with dose limits, drop to basic mode, or needs review.
Match a libraryPick a synthetic patient: a 79-year-old with heart failure and a slow circulation, a healthy 42-year-old, or an 86-year-old with sepsis who sits outside the model's range, and the page says so. The chart sets up the model, the model does the math, AI MedAgent explains it, and the order comes out with a goal, a flow pattern and a do-not-repeat time. One drug today: propofol, an anesthesia drug.
Draft an orderDr. N. Ty Smith (1932 to 2015) taught anesthesia at Stanford and then for 24 years at UC San Diego and the VA San Diego. He was the founding president of the Society for Technology in Anesthesia and a mentor of mine.
In 1972, with Anton Zwart and Jan Beneken in the Netherlands, Ty coupled a model of the circulation to a model of how an anesthetic is taken up, on an analog computer. The point was simple and it has never stopped being true: the circulation changes the drug, and the drug changes the circulation. That work grew into BODY, a PC simulation with more than 200 variables in which a teacher could build a patient and watch the same drug land differently in a different body. In 2005 Ty aged BODY's patients by changing more than 50 parameters, and the elderly fared worse at induction.
BODY could teach the lesson. It could not be used on a real patient, because there was no electronic chart to read and nothing that could reason over it. Smart Orders is my attempt to carry Ty's lesson to the bedside with what did not exist then: the patient's record in electronic form, free public drug data, and AI that can reason over both. It does not run Ty's equations. It uses published models, cited on the demonstration page, and gives Ty the credit for the idea. His physiology lives on in its own right: BODY formed the backbone of the HumanSim engine, which led to BioGears and then to Kitware's open-source Pulse.
Sources: Zwart A, Smith NT, Beneken JEW. Multiple model approach to uptake and distribution of halothane: the use of an analog computer. Comput Biomed Res 1972;5(3):228-238. · Smith NT, Zwart A, Beneken JEW. Anesthesiology 1972;37(1):47-58. · Smith NT. PC-based “BODY” simulation presents anesthesia learning chance, challenge. APSF Newsletter 1999;14(2). · Smith NT, Starko KR. The physiology and pharmacology of growing old, as shown in Body Simulation. Stud Health Technol Inform 2005;111:488-491. · Shafer SL. The Measures of a Man. Ty Smith Lecture handout, Society for Technology in Anesthesia, 2015. · Pulse Physiology Engine documentation, Kitware.
Will an order placed from each CPOE order sentence reach the pump with dose limits on? The same data, one step further.
Check the loopThe exact HL7 messages that carry one aligned order to the pump and its answer back.
See the messagesDaniel Pettus spent forty years in medical device and health IT leadership at Alaris, CareFusion and BD, contributed to IHE Patient Care Device interoperability standards, and is named on two United States patents. He writes Inside the Loop at insidetheloopdp.substack.com and is the author of The Technology Was Never the Problem, at pettusbook.com.
AI MedAgent is a research demonstration of a patent-pending method. Advisory only. Not a medical device. Not for clinical use. Synthetic data only.