Chris McMaster had his pre-confirmation presentation for his PhD, and his topic is domain-expert led machine learning for healthcare. His LinkedIn page says “clinician and clinical data scientist.” Chris made the point that there are thousands of people around the world working on shared machine learning datasets to improve their algorithm and although the datasets are useful for data scientists, very few of the outputs created from those datasets can be applied in practice. We need more datasets that reflect real practice in order to support real practice. Another interesting topic arose in the discussion after his presentation: A major emphasis in computer science is generalizability of the technique to other datasets, and local datasets may not be looked on as favorably. However, I believe that a less generalizable output that can be integrated into practice is potentially more valuable than yet another a generalizable output that can’t be used. My summary of Chris’s hypothesis is that having more domain expertise guiding machine learning projects from the beginning will result in AI-supported clinical decision support being used in everyday care. Watch his work at the Austin to see if his hypothesis will be borne out.
This made me think a lot about the types of expertise needed in digital innovation for healthcare. In Chris’s world, clinical expertise is underutilized in favor of methodological expertise. Brian gave an expanded version of his AIDH talk to the LHS Academy fellows about patients needing not only data but knowledge to be true partners in their healthcare. I have a previous student from the University of Utah who came to the field as a patient with a lifetime of expert experience in how healthcare has worked or not for him. Like many others, he has seen his expertise ignored and dismissed by his healthcare teams.
We need all kinds of expertise to navigate the complexity of the human body/mind along with the social, political, and system complexities of healthcare delivery. As a linguist who entered the field, I’m always cognizant of being an outsider AND of the viewpoint and knowledge I bring to the table that is unique. Every one of us has expertise and experience that matters, and I hope you will embrace yours!
My daughter Clare is visiting us for the first time since we moved here, and she and I explored the wildlife on Raymond Island over the weekend (after my 7 days of isolation from COVID). She is discovering many of the things I’m loving about Australia: kangaroos, koalas, kookaburras, black swans, and the really nice people who stop and talk to us along the path.
Regional Linguistic Quirks (RLQ): Clare heard a few Australianisms on the road but didn’t notice some of them that I I pointed out, like “mozzies” and “pretty as”. Is that phrasing a form of ellipses, where the comparator is left off--whatever it is, it’s new to me, and I really like it.
Comments