What do the data say? The question is central to engineering and reflective of its motto “In God we trust; all else we measure!” Smart technologies, sensors, and their ability to quantify natural and engineered processes have changed who (and what) gets to decide. People and organizations are frequently forced by the deluge of data and time constraints to relinquish decision-making to smart systems. In many applications, human operators cannot respond in time to override AI automations despite misguided assumptions to the contrary[i],[ii]. As a result, artificial intelligence technologies are increasingly treading on once exclusively human turf. Medicine successfully shielded itself from most of these intrusions, which landed it on the list of least technologically innovative industries. It’s no mystery[iii]. The issue came up repeatedly in Harvard’s Digital Health cohort[iv],[v], which I attended last year.
The challenges and opportunities for providers and clinicians innovating and digitally transforming their services begin with an element they don’t control - instrumented patients who arrive armed with physiological, fitness, and environmental data, trends, and analysis. In theory, these insights should convey more objective patient stories that help clinicians make better and more informed diagnoses and clinical decisions. In practice, clinicians and clinical systems are often overwhelmed by their scope and volume. I attended a briefing by a medical practice consultant who described the frustrations felt by doctors coping with patients who showed up better informed and up to date than they were. Their frustrations extended to insurance constraints that limit how much time they can spend with their uber-informed patients.
Biobeat’s sensors also generate large quantities of data. Fortunately, their clinical models are engineered to buffer clinicians from the deluge. Instead, these systems judiciously exploit advanced analytics and artificial intelligence to deliver targeted sets of actionable information, trends, and clinical insights. The strategy empowers clinicians without undermining their roles in diagnosing, treating, and monitoring patient outcomes. It’s a good solution because powerful as AI tech has become, it is far from expressing, much less replacing human judgment. I will illuminate the benefits and tradeoffs in the next post which will focus on how the 24BP and ABPM systems improve diagnostic performance and lower costs by eliminating white coat hypertension syndrome.
REFERENCES
[i] Ozzie Paez, Dean Macris, Safety Implications of Self-Driving Cars, LinkedIn, April 3, 2018, https://www.linkedin.com/pulse/safety-implications-self-driving-cars-ozzie-e-paez/ [ii] Ozzie Paez, Dean Macris, The Fatal Uber Self-Driving Car Crash – Update, Ozzie Paez Research, June 12, 2018, https://www.ozziepaezresearch.com/post/2018/07/12/the-fatal-uber-self-driving-car-crash-update [iii] Risa Ravitz, Why Healthcare Can Be Slow to Adopt Technological Innovations, April 2, 2020, https://www.med-technews.com/medtech-insights/why-healthcare-is-slow-to-adopt-technological-innovations/ [iv] Ozzie Paez, Harvard Digital Health Wrap-Up, Ozzie Paez Research, December 9. 2021, https://www.ozziepaezresearch.com/post/harvard-digital-health-wrap-up [v] Digital Health, Harvard Online, https://www.harvardonline.harvard.edu/course/digital-health, accessed September 21, 2022.
Comments