When I was 11, I told my stepdad about these cool new pants kids were wearing at school: Levi Strauss jeans. He laughed and informed me that they were first sold in 1873, and more than four decades later, he still teases me about “new” things I am discovering. Similarly, in Australia I hear about this new field called digital health and about how AI has been around for a decade. So in addition to blog posts with a theme of research, I offer this post as the first in a series on the history of data, computation, and technology in healthcare. Like my discovery of Levi’s, my descriptions of history will be based on my experience and be partial, so I welcome corrections and enhancements.
A pioneer in the use of data use in healthcare was Ernest Amory Codman , who created the Registry of Bone Sarcoma, in 1920. This excellent article says
His goal was to collect and analyse all of the cases of bone cancer (or suspected bone cancer) from across the US, and to use these to establish diagnostic criteria, therapeutic effectiveness and a standardised nomenclature.
I was lucky to learn about informatics from many of the pioneers of the field of informatics as I started my PhD in 1994 at the University of Utah where Homer Warner was still chair. In 1959 Ledley and Lusted published a seminal article that described the conceptual opportunities of using computers to support reasoning about medical diagnosis. This article inspired Homer Warner who I’m sure agreed with the sentiment expressed by outspoken Scottish epidemiologist Archie Cochrane that medicine was based on
a level of guesswork” so great that any return to health after a medical intervention was more a “tribute to the sheer survival power of the minds and bodies” of patients than anything else.
Homer published in 1961 the first paper to use real patient data to support diagnosis and compared its performance against physicians. An oral history of Homer by Dean Sittig tells the story:
Because Warner had a large number of patients with congenital heart diseases coming through his laboratory, they decided to make a probabilistic model to diagnose patients with one of 35 different forms of congenital heart disease. First, they collected data on how frequently each of 50 different findings such as murmurs of different kinds and cyanosis occurred in each disease and how common the diseases were in the population of patients referred to the laboratory. After collecting several hundred such cases, they built a matrix which showed the diseases on one axis and findings on the other axis, and then at each intersection of the symptom with the disease there was a number representing the frequency of that finding in patients with that disease. This table then formed the basis for diagnosing a particular patient based on the findings observed as recorded by the referring physician.
In The Theory That Would Not Die (one of my favorite non-fiction books), McGrayne writes
Warner showed that Bayes’ could identify their underlying problems quite accurately. “Old cardiologists just couldn’t believe that a computer could do something better than a human,” Warner recalled.
These two articles are attributed with the development of a new field of investigation. In many ways, Homer's pathway did not go the way he hoped. As his colleague Al Pryor used to quote from the Rhinestone Cowboy, "There's been a load of compromisin' on the road to my horizon." But he launched a new field in computing in healthcare, and people came from across the world to work with him. His legacy continues as people build on the foundations he put in place, and I am thankful for the time I overlapped with him.
Comentários