SPA Spring Meeting Reviews
Saturday Session I: Innovation Leads to Collaboration
By Sean S. Barnes, MD, MBA
The Digital Disruption of Medical Knowledge: Ending Bias, Borders, and Boredom
Todd Ponsky, MD, (Cincinnati Children’s Hospital, Cincinnati) discussed the digital disruption of medical knowledge. He opened his talk with examples of industries being disrupted – Uber disrupting the taxi industry, Netflix disrupting the film industry, and Apple disrupting the camera industry. One of the most disrupted industries happens to be the publishing industry. Examples shared were the UK magazine industry, decreasing from roughly 64,000,000 magazine sales per year to 48,000,000 sales per year in the span of a decade, due to mobile devices becoming more common. Newspaper advertising revenues decreased by 50 billion dollars after Google was founded. Noting these examples, it is surprising that the medical publishing industry has not been disrupted in the same way.
An interesting article titled “Challenges and Opportunities Facing Medical Education” described how quickly the total amount of medical publications doubles. In 1950 the doubling time was 50 years, 1980 – 70 years, 2010 – 3.5 years, and in 2020 – 70 days. The exponential growth of medical publications has created a new problem; it is now much harder to identify what knowledge being shared is important. Traditionally, physicians allowed for new medical knowledge acquisition through textbooks, medical societies, and journals. Unfortunately, these methods are now inefficient and outdated.
The concept of staying up to date on new medical knowledge requires us to harness our digital future. This can be achieved through artificial intelligence and improved distribution of medical knowledge. Through the process of machine learning, Dr. Ponsky was able to improve the editorial efficiency for the American Pediatric Surgical Association. Leveraging ‘Natural Language Processing Algorithms’ and ‘Word Clusters’, the artificial intelligence program reviewed 520 abstracts with outcomes comparable to human reviewers (81% sensitivity and 85% NPV). Additionally, through surgical telemonitoring, surgical technique can be taught, and possibly improved, through the addition of artificial intelligence. Perhaps the greatest barrier to adoption will be the lack of expert surgeons willing to take time away from work to travel and teach.
Acquisition of new medical knowledge should rely on efficient distribution. Currently, distribution of medical knowledge is inefficient and relies completely on publishing houses. Dr. Ponsky proposes removing the publishing houses and leveraging open access crowd sourcing in a manner similar to what social sciences did in creating the Social Science Research Network. Dr. Ponsky concluded with a discussion on leveraging video and social media channels to improve the efficiency of medical knowledge distribution. He gave an example of his trainee giving an invited lecture at a surgical conference with an estimated reach of 300 people; that same talk recorded and made available on the APSA video channel reached over 50,000 viewers and tallied 1,600 unique engagements. Academic medicine will be disrupted, it is simply a matter of how and when.
Near Infrared Spectroscopy Collaboration – From Engineering Research to Clinical Use in Pediatric Anesthesia
Justin Skowno, FCA, FANZCA, PhD, (Children's Hospital at Westmead, Sydney) discussed his experience with Near Infrared Spectroscopy (NIRS) and the potential utility of this technology both in research and clinical practice. NIRS is a noninvasive technology that continuously monitors regional tissue oxygenation. Dr. Skowno commented that initially, NIRS were used for assessment of cerebral oxygen saturation; however, its use has now expanded to evaluation of oxygenation of tissues other than the brain. Research evaluating NIRS can broadly be divided into Cardiac and Non-Cardiac.
The body of evidence for NIRS in non-cardiac patients is primarily limited to observational studies and case reports. Dr. Skowno discussed an international, multicenter, observational study by Olbrecht et al from Anesthesiology 2018. In this study of cerebral oxygenation during Infant and neonatal anesthesia, it was observed that mild and moderate low cerebral saturation occurred frequently, whereas severe low cerebral saturation was uncommon. Spinal anesthesia demonstrated similar outcomes in infants undergoing surgery (Froyshteter et al Journal of Anesthesia 2018). The authors of this study concluded that despite changes in heart rate and blood pressure, there were no clinically significant changes in cerebral oxygenation. Lastly, Dr. Skowno commented on a study by Dewhirst et al that demonstrated cerebral oxygenation declined slightly during general anesthesia with the transition from normocarbia to hypocarbic conditions. Moreover, the cerebral oxygen saturation decrease related to hypocarbia was easily reversed with a return to baseline values by the administration of supplemental oxygen (60% vs. 30%).
In children undergoing cardiac surgery, NIRS monitoring is becoming the standard of care in both the intraoperative and postoperative settings. There are more studies investigating NIRS monitoring in children undergoing cardiac surgery and this has resulted in many “believers” however agnostics remain. At a minimum, NIRS can be considered a clinically useful “extra pair of eyes” during cardiopulmonary bypass. Dr. Skowno went on to discuss the study by Kussman et al published in Anesthesia & Analgesia (2017) titled “Cerebral Oxygen Saturation And Children With Congenital Heart Disease And Chronic Hypoxemia”. In this study, the authors concluded that children with adequately compensated chronic hypoxemia appear to have cerebral oxygen saturation values within the normal range. Furthermore, a low cerebral oxygen saturation value in the setting of chronic hypoxemia should not be interpreted as a ‘normal’ baseline for that patient.
As NIRS monitoring becomes more prevalent, it is important to recognize the current limitations of this technology. Dr. Skowno highlighted many of these limitations, including depth of interrogation, validity, data analysis, and physiologic meaning. Continuous wave devices typically are cheap and provide a robust amount of data, however, they may be imprecise, have poor spatial resolution, and have both scattering and absorption-based issues. Data issues can arise when one uses cerebral oxygen saturation as a single point measurement. Other potential data issues include absence of interrogation of other signals (MAP, Sa O2, CO2) and/or poor clinician judgment - “Oh, it’s just a saturation monitor.” Lastly, Dr. Skowno closed his talk discussing the potential future of monitoring, including cellular oxygenation via cytochrome 3 oxidase, diffuse correlation spectroscopy, time resolved spectroscopy, functional NIRS/imaging, and autoregulation.