Silicon Collar looks at machines and humans at work in over 50 settings across industries and countries. On this blog I will excerpt many of those settings over the next few weeks. On Deal Architect I will excerpt more of the policy parts of the book.
At the University of California, San Francisco (UCSF) Medical Center, a research and teaching hospital. In an interview, Dr. Michael Blum, the hospital’s Chief Medical Information Officer, discussed a wide range of automation
“Just by having electronic health records, we have seen significant benefits that don't get widely reported or celebrated. For instance, as part of our automation we put in a robotic pharmacy (in photo) that's linked to our electronic health records. The machine bundles appropriate pills and ensures that exactly what was ordered ends up in a sealed, bar-coded package. Then a nurse can scan and deliver it to the right patient at the right time and in the right dosage. There's a whole other story in the move from just hanging IV drips, where nurses would estimate and use mechanically alarmed devices to establish the flow rate in the drug administration. We now have automated infusion pumps.”
“If you look at AI and machine learning around cancer treatments, they are scaling rapidly. The focus on genomics and proteomics [the study of proteins] is leading to much more rapid discovery than has ever happened before, with progress in previously untreatable cancers or those that were incredibly difficult to manage. Matching genomics to appropriate targeted therapies was not doable five years ago. We are seeing rapid progression and drug development in those spaces, plus targeted therapies and immunotherapies. Cancer is an obvious space because it's relatively easy to look at specific mutations and match the somatic mutations to targeted therapies.”
“During surgeries, robots can be positioned at angles which are impossible for humans. They can also make more precise incisions, which can lead to less blood loss and quicker patient recovery. It's different than robotic automation in manufacturing where it's mostly about reliability and repeatability. We get that from our TUG robots which carry supplies and our pharmacy robots where we have dramatically lowered errors.”
“For decades, doctors would dictate their surgical notes, procedure notes, and discharge summaries. Armies of humans would then transcribe these notes into a digital document. Gradually, speech-to-text technology from companies like Nuance has now evolved to a 99% or better accuracy in well-implemented systems. The ability to train to a particular voice has gotten much better. The ability to recognize dialects and accents has gotten much, much better. Radiology is one of the best examples of its use because it has a somewhat constrained medical vocabulary.”
“There's a perception that doctors and nurses are technophobic and push back on computers and automation because they separate the providers from the patient. I think that's untrue. Yes, we are in a caring profession where we need to have a human bond with the patient to best care for them. However, we also appreciate tools that help us do our best for patients. The magic happens when we find technologies and automation that simply disappear—they help us provide better care for our patients without being in the way.”
Photo Credit UCSF