Hospitals use clinical pathways to decide important
decisions on how to treat a patient. The Flagler hospital in Saint Augustine,
Florida has found a way to create improved pathways to enhance care and cut
treatment costs by using unsupervised learning in AI. Unsupervised learning in
AI is simple, it does not know the expected output like Supervised learning.
Supervised learning takes in input data and output data for the AI to learn,
but in Unsupervised it is just given the input data to make inferences from.
The most commonly used unsupervised learning is clustering, it finds hidden
patterns or groups in a dataset by finding high measures of similarity data
within it.
I believe that this is a wonderful use of technology being
used, especially since it is being used in a hospital: it will speed up
treatment and cut the costs by thousands of dollars per patient, which in
return lowers patient’s medical fees. From the data that this hospital has been
able to track, the AI was able to analyze and learn from thousands of patient
records where the commonalities lie between patients with life threatening
diseases such a pneumonia and sepsis.
I personally know how long the process is at being in the
hospital. My mother is sickly and has been taken to the ER multiple times, and
I’ve sat in the ER for hours waiting for simple tests to be done and for their
results to come back. I believe that if the hospital we went to were to use an
AI program to do what Flagler hospital was able to get done, it would’ve been faster
to get my mother’s treatment done. She ended up spending an additional month
after her surgery for the doctors to find the correct dose of Warfarin to keep
her INR in the healthy range.
When you’re a patient, or a patient’s family, that waiting
time can be devasting. We don’t know if the results will be okay or not, so we
can only wait in anticipation. If you think about that stress, it is mentally
helpful for both the patient and its family to have less days of hospital
admission. The Flagler hospital was able to reduce the length of days by 2 days
and even reduce readmission percentages from 2.9% to 0.4%. The physician who
recently became a computer scientist said that he had never seen something like
this before at a hospital. Using the AI they bought from Ayasdi Inc, the
hospital is working efficiently and are finding themselves treating one
condition per month, rather than their previous expectation of 12 conditions
over 3 years.
IT is finding its way even into the medical field, just
like machine learning can be found improving businesses and firms in multiple
areas of efficiency. Although we can’t always say that technology is improving
everything, it is definitely creating a doorway for the future of how we
approach medical treatment.
References:
1 comment:
Seemab Kazmi
Extra Credit Comment 2
AI in Hospitals cut thousands of treatment costs for patients By Yon Su Kim
I did a similar blog on IBM’s Watson and their initiative to use big data and AI like Watson to make hospitals more efficient. But, the study was actually ineffective in completing its intended purpose. Watson is a supercomputer that can store data and answer questions. It took the industry by storm in 2012 when it won on Jeopardy by answering any question thrown its way. IBM made many promises on how Watson would be implemented into healthcare and business. Unfortunately, it has not lived up to its promises. The goal was to use Watson’s data management abilities to hold data records on all patients and their medical history, assist doctors by reading results and helping doctors, and diagnosing patients as well. But, the hospitals that partnered with Watson found that it gave the same diagnose as the doctor and did not expediate the process for the patient. It was also not successful is acquiring all patient records.
I had done research on Watson in my junior year of high school and was very impressed with what IBM planned to do with their technology. Finding out that they were not as successful in doing so was very disappointing for me because I thought it had a lot of potential. But, I am glad to know that AI can still be utilized in healthcare to make the experience of customers better. Its also difficult to transition for many hospitals and medical environments because the medical field still has such a large grey area of research that is being done. IBM’s endeavors also cost hospitals a lot of money and the hospitals in question were already some of the most leading health facilities in America and Canada. Unlike Canada, America lacks universal healthcare and as a result the cost of treatment at these locations would have been a lot higher. But, its also good to know that AI can make the healthcare system cheaper if used properly. Although Watson’s small failure was a set-back for IBM, AI still has a lot of hope in healthcare.
Post a Comment