AI IN MEDICINE -IT'S IN YOUR POCKET - The anesthesia consultant

Author
Recent Posts

Physician anesthesiologist at Stanford at Associated Anesthesiologists Medical Group

Richard Novak, MD is a Stanford physician board certified in anesthesiology and internal medicine.Dr. Novak is an Adjunct Clinical Professor in the Department of Anesthesiology, Perioperative and Pain Medicine at Stanford University, the Medical Director at Waverley Surgery Center in Palo Alto, California, and a member of the Associated Anesthesiologists Medical Group in Palo Alto, California.
emailrjnov@yahoo.com

Latest posts by THE ANESTHESIA CONSULTANT (see all)

AI IN MEDICINE –IT’S IN YOUR POCKET - 30 Apr 2026
THE PITT: 10 MEDICAL DETAILS THE SHOW GETS WRONG, AND 10 DETAILS THEY GET RIGHT - 2 Apr 2026
ANESTHESIOLOGY AND MALPRACTICE LAW - 23 Mar 2026

The future of Artificial Intelligence in Medicine is in your pocket, not in the Electronic Medical Record (EMR). The EMR is now universal in America’s hospitals. To what degree will useful AI be integrated into the EMR? I’m betting the most useful applications of AI are in the anesthesiologist’s pocket—in the smartphones we each carry.

Let’s look at two examples.

Patient #1 has significant lumbar disc disease and is scheduled for a lumbar laminectomy. She’s otherwise healthy. The anesthesiologist anesthetizes the patient safely for her surgery. Intraoperatively, the AI elements of the EMR help in several ways. The EMR: a) reminds the anesthesiologist to administer a preoperative antibiotic because it was not recorded as given; b) reminds the anesthesiologist that the fresh gas flow of 5 liters/minute is too high; c) reminds the anesthesiologist that the blood pressure was not measured for 8 minutes following the disconnection of the BP cuff for the turning of the patient from the supine to prone positioning; and d) alarms and reminds the anesthesiologist that the blood pressure of 79/40 is too low, and needs to be treated. The anesthesiologist follows all these prompts, and the surgery and anesthesia are completed without complication.

Patient #2 has significant lumbar disc disease and is scheduled for a lumbar laminectomy. She has the medical comorbidities of carcinoid syndrome and Wolf-Parkinson-White (WPW) syndrome. Her anesthesiologist has not anesthetized a patient with either condition for several years, and is uncertain what precautions need to be taken. In years past, the anesthesiologist would Google “anesthesia for carcinoid” and “anesthesia for WPW” to educate himself prior to initiating anesthesia care. Today the anesthesiologist could ask OpenEvidence, an AI app on his phone, how to safely conduct this anesthetic by entering a specific prompt such as, “How should I do an anesthetic for a lumbar laminectomy on a 50-year-old woman with carcinoid syndrome and WPW syndrome?” The AI app will summarize and recommend the safest approach for each disease. In 2025 The New York Times reported that OpenEvidence raised $200 million as a ChatGPT for medicine. The New England Journal of Medicine recommended OpenEvidence as the most reliable medical AI app.

Try it yourself right now.

As an alternative to OpenEvidence, a physician could utilize ChatGPT, developed by OpenAI, or Claude, developed by Anthropic, both well-known AI apps on his phone. A critique has been that ChatGPT or Claude could make erroneous recommendations—even though they are summarizing all medical knowledge on the internet—and that ChatGPT and Claude are not doctors, and should not be relied upon. An appropriate response is that you’re a doctor, and like everyone else who is learning to use AI apps, it’s your job to regard smartphone AI input in the context of your education, training, and experience to make the correct decisions. AI apps in our smartphones supplement the knowledge base of every physician, and are synergistic, more powerful, and more useful than any AI contributions from the EMR. This represents a marked advance in anesthesia care.

BIG DATA

Big Data in medicine is a different concept than AI in medicine. Compiling EMR information from large populations of surgical patients provides a vast amount of public health data, often referred to as Big Data. Big Data has attracted the attention of researchers and scientists. The EMR at your hospital compiles information from every anesthetic into an Anesthesia Information Management System (AIMS), and researchers attempt to utilize this data to improve future care. By analyzing this data, computer scientists can design Decision Support Algorithms (DSAs) into the EMR, to help make anesthesia care more safe. An example of a Decision Support Algorithm is a reminder that a patient’s mean tidal volume should be between 6 and 8 ml/kg. For a 70-kilogram patient, the appropriate tidal volume would be between 420 and 560 ml. If the anesthesiologist was using a different tidal volume, the EMR would alert him or her of this guideline. Further examples of DSAs include: a) coordinating a more judicious administration of drugs to prevent postoperative nausea and vomiting; b) a reminder to administer antibiotics, which has been shown to improve the administration of antibiotics on time from 62.5 to 83.9%; and c) monitoring of blood pressure in the operating room was improved by the Smart Anesthesia Messenger (SAM warning system), which warned doctors and nurses of a lack of measurement after 6 minutes. These feedback systems serve to make anesthesia safer.

In Belgium a model AIMS system was integrated into the EMR, and allowed MDs to assess intra-operative, postoperative, and extra-hospital mortality. The Belgian scientists wrote, “the training of medically qualified practitioners in data science and artificial intelligence is of paramount importance, . . . In Belgium, an initiative has been launched to create a certificate in artificial intelligence (AI) in healthcare, open to doctors, engineers and informaticians.”

Within anesthesiology, the Multicenter Perioperative Outcomes Group (MPOG) is a multicenter learning health system compiling EMR data from fifty academic and community hospitals. MPOG aims to be a learning health system focused on perioperative care, with a goal to continuously raise standards for Quality Improvement, research, and patient safety. MPOG states it is “incumbent upon perioperative clinicians to become increasingly familiar with the opportunities and challenges afforded by large perioperative EMR databases.”

But there are drawbacks to analyzing Big Data. EMR problems include the “4 Vs of Big Data (Volume, Velocity, Variety, Veracity)”. Volume refers to the problem that one single anesthetic record in the EMR can include 10+ pages of information. A one-day surgery and hospital stay may result in 1000 pages of EMR. Velocity refers to the instantaneous speed at which massive amounts of data is created. Variety refers to the fact that the data exists as text, images, videos, voice files and other information that doesn’t fit easily into the framework of a spreadsheet. There’s also the lack of a standard EMR—different vendors make different EMR products. The adult hospital and the children’s hospital at my university share adjoining walls, yet the two hospitals have separate EMR systems with different passwords and formats. Veracity refers to whether the data in an EMR can be trusted. Disconnected ECG leads, BP cuffs, and invasive monitoring catheters can result in markedly untrue readings of a patient’s vital signs or clinical data. These potential flaws represent barriers in converting data into knowledge, and converting knowledge into clinical quality improvement.

A Big Data platform is not a magic bullet that answers all questions. Variation in practices exist in different institutions, such as the anesthetic technique employed, medications used, and operating room staffing models used. Yet reputable researchers conclude that, “Opportunities are ripe for coalescing perioperative EMR data across patients, clinicians, institutions, and regions to perform comparative effectiveness research and improve the quality and safety of anesthesia care.”

In 2019 I published the book Doctor Vita, a prescient novel describing the development of Artificial Intelligence in Medicine modules at the fictional University of Silicon Valley. Doctor Vita is true AI—an intelligent force that understands all medical knowledge and synthesizes patient care recommendations based on that knowledge. Current Big Data and the EMG stop far short of Doctor Vita. The AI in your phone approximates what the future of Artificial Intelligence in Medicine can be.

YOUR SMARTPHONE

Will the most useful AI in medicine come from physicians, engineers, and informaticians working to improve the hospital EMR systems, or will it come from that smartphone in your pocket, with an AI app that summarizes all published information on a given topic? Big Data analysis by researchers will eventually be published in reputable medical journals. This content will be available to AI apps on your smartphone. A typical physician will never personally analyze Big Data, and may never choose to read the studies published by the scientists researching Big Data. But when you’re really wondering how to handle a complex unfamiliar case, your smartphone AI app will be your most important ally. Both the EMR and the smartphone will contribute to AI’s future. My prediction, as an anesthesia practitioner in the trenches who’s been doing his own anesthetics day after day for four decades, is that the smartphone app will be more useful.

And . . . to anyone who believes smartphones should be banned from an anesthesia practitioner’s hands intraoperatively, the revelation that AI in your smartphone is a valuable medical tool should be all the rebuttal needed.