The Future is Now: How Advances in Technology are Already Affecting Anesthesiology Practice

Mark Arcario, MD, PhD

While much discussion at the IARS 2021 Annual Meeting was centered around the “future of anesthesiology” in an eloquent, yet abstract sense, the session entitled “How Advances in Technology will Advance the Practice of Anesthesiology” tackled this topic in a much more concrete sense. Detailing how current advances in computational power and other technologies are currently affecting and will continue to affect the everyday practice of anesthesiology, the four panelists provided an exciting preview into the future of the specialty and how embracing technology will not only make the field more efficient, but also more valuable to patients and colleagues.

Ronald Pearl, MD, Professor and Chair of Anesthesiology, Pain and Perioperative Medicine at Stanford University, framed the session with introductions and moderated discussion. 

Gaganpreet Grewal, MD, Co-Chief of Anesthesiology at Christus Santa Rosa – San Marcos, presented a lecture on how advances in surgical techniques and technology will affect the practice of anesthesiology. Starting with more recent advances in surgical technique, such as endovascular repair of aneurysm, robotic surgery, and increasing number of non-operating room procedures (i.e. catheterization laboratory, interventional radiology, gastroenterology procedure rooms), Dr. Grewal briefly discussed how anesthetic practice has advanced and changed alongside these innovations. With that background in mind, Dr. Grewal moved to what is on the horizon for surgical technology. It seems most advances are centered around further development and extension of the robotic surgery systems and smart detection systems. One particularly interesting development Dr. Grewal mentioned was the iKnife, which is able to detect malignant cells in the smoke created by electrocautery, possibly limiting dissection and reducing resection areas in cancer surgery.

While these futuristic medical advances are interesting in and of themselves, participants were eager to learn how this might affect their practice in the near future. The overarching theme was that more flexibility and expertise in more areas of anesthesiology will be required. Already, invasive robotic procedures have been performed under regional anesthesia with minimal sedation and the expectation is that the standard of care for these cases will shift more towards this direction with proven decreased hospital stays and improved pain scores. Additionally, as more focus is shifted to non-operating room anesthesia (NORA), anesthesiologists must be involved in designing new spaces as well as guidelines for selecting patients that can be safely anesthetized in these locations, as very likely sicker and sicker patients will be pushed to “safer and less invasive” procedures. With less invasive procedures and NORA becoming more common, it is imperative that training in these areas be bolstered, likely via a combination of hands-on experience and simulation education.

Next, Michael Mathis, MD, Assistant Professor of Anesthesiology at the University of Michigan, discussed the impact of artificial intelligence and machine learning on anesthesiology practice. He explained why adoption of machine learning techniques has been so slow in medicine compared to other fields, including the large downside to incorrect predictions, general lack of transparency, and lower tolerance of computer errors. He framed how machine learning can be leveraged to improve the practice of anesthesiology. Using two recent examples of successful machine learning in the literature, the COVID deterioration scale and prediction of postinduction hypotension, Dr. Mathis described how the power of machine learning can be utilized to help guide the daily practice of anesthesiology. While the algorithms presented were impressive, what was more important was the innovation they bring to the field. The COVID deterioration scale, while being moderately good at predicting which COVID positive patients would further deteriorate, proved useful because it actually provided a breakdown of what the algorithm had identified as the risk factors for this deterioration as well as actionable items to possibly prevent further deterioration.

Shifting gears, Dr. Mathis discussed the postinduction hypotension algorithm. While it performed moderately well, the real power lies in the downstream clinical judgment of managing patients with predicted postinduction hypotension. He very eloquently laid out a scenario in which postinduction hypotension could be a potentially adverse outcome and the options to treat this and how many times; treatment options are a balance of risk and benefit. Although this was not included in the algorithm, it does show a path forward into making machine learning more useful for practicing clinicians and the power it can demonstrate.

Following Dr. Mathis, Jeanine Wiener-Kronish, MD, Henry Isaiah Dorr Professor of Anesthetics and Anesthesia at Harvard Medical School, outlined advances in and the trajectory of remote monitoring. Remote monitoring is becoming so widespread that it has even made its way into everyday wearable technology (think of the health apps on your smart watch or smart phone), but these have generally come with quite a few false positives and false negatives; the signal-to-noise ratio is not large enough to be useful.

Also, how can these devices and technology be utilized to prevent disastrous outcomes? The answer is: it’s complex. Utilizing retrospective data to understand why a particular group of patients (head and neck tumor patients, in this case) required ICU care or had catastrophic outcomes on the floor, Dr. Wiener-Kronish and colleagues tailored the remote monitoring and alerts in AlertWatch to notify clinicians when a patient would require escalation of care. In almost 90% of alerts, the condition change in the patient necessitated the clinician to change their treatment plan. She then introduced the HAVEN system, aimed at predicting catastrophic events. Taking data directly from the electronic medical record and processing it through a machine-learning algorithm in real time, HAVEN was able to predict 40% cardiac arrest or unplanned ICU admissions at least 12 hours prior to the event. The benefit here was that the system was constantly taking in the updated patient data and able to refine its models and predictions, but the algorithm remains complex, utilizing many hundreds of pieces of data and changes in that data to make these predictions. The takeaway from this talk is that to make a significant improvement in patient outcomes using remote monitoring, all clinical data available needs to be utilized to make these predictions.

To finish the session, Alex Evers, MD, Henry Mallinckrodt Professor of Anesthesiology at Washington University in St. Louis, delivered an insightful talk on developments in pharmacology that will affect anesthetic practice in the near future. The field of anesthesiology grew out of the need for physician care to manage the physiological effects of anesthetics (chloroform and ether at the dawn of the specialty) on surgical patients. While anesthesiologists have become very adept at managing the adverse effects of the drugs provided during an anesthetic, Dr. Evers envisions an era during which therapeutics are refined such that there are minimal or even no side effects. Describing advances in every field of drug discovery, including identification of physiologic targets, resolution of drug binding site via structural methods, and both in silico and in vivo drug screening methods, the time to identify and test a new therapeutic has shrunk considerably in the last decade. Using the opioid receptor as a model, Dr. Evers showed that using advances in basic science technology has produced a new compound that provides analgesia in an in vivo model, without causing respiratory depression. Moreover, he provided an example of computational screening against GABA receptor which detected a compound with the sedative effects of propofol or etomidate without the adrenal or cardiovascular adverse effects of those drugs. This could prove to be an exciting new era in anesthesiology.