I'm a doctor working in clinical trials, specifically on treatments for rare and incurable illnesses. Since I have been in this industry I have seen barely any enthusiasm for AI, let alone actual implementation. There is a lot of work going on at the startup level, and big Pharma are keen, but this is currently limited to drug discovery for the most part - which is great, but doesn't change the fact that it takes 10-20 years for a newly discovered, promising molecule to reach patients.
People point towards the animal and human testing phases - yes, there is a big lag there, but the duration a participant spends on a trial is minimal compared to the time spent planning a study, gathering and processing the data from that study and so on. A Phase 1 study might involve a month of recruiting and 2 months of actually monitoring a participant, but it took a year lead time to get there and could be 2 more years until Phase 2 gets going, and many of the bottlenecks are exactly the kinds of tasks that AI can already trivialise.
Medical data is not as clean as maths, but it is cleaner than a lot of other domains and totally amenable to even existing AI models. We don't need GPT 8 to uncover life changing treatments in the vast amounts of unused, already anonymised, medical data that the pharmaceutical industry sits on, or to speed up the process of getting it through Phase 1, 2 and 3.
You would be shocked how much time in this industry is wasted on manual data entry, manual pattern seeking, manual projections etc. Exactly the kid of work that has been trivialised within certain areas of big tech, and yet this is a domain that will impact every single human life on this planet in a positive way.
So yes, I just wanted to rant. I am sick and tired of medical textbooks listing in great detail the cause and effect of xyz horrible progressive illness only for a tiny stub at the end of the page titled "Treatment" and the words "There is currently no known cure". It's not good enough.
[link] [comments]