An “artificial intelligence framework” got better and lower-cost patient outcomes than real doctors.
“Artificial intelligence” in this context is not about
humanoid replicants but about creating logical processes that can mechanically reproduce the decision making power of human brains – though we see no reason why these processes can’t be programmed and plunked into the headpiece of a
medical robot.
The researchers’ goal, they state in their
paper, was to get their logic network to determine courses of treatment that would lead to optimal results based on cost-per-unit change (CPUC) – that is, “maximizing patient improvement while minimizing treatment costs” – which they then compared with the actual outcomes achieved by the doctors in the records.
After hundreds of simulations, the model obtained an average CPUC $189 compared with the $497 obtained in the real treatment. Not only that, the researchers said: “Tweaking certain AI [artificial intelligence] model parameters could further enhance this advantage, obtaining approximately 50% more improvement (outcome change) for roughly half the costs.”
You may be saying, yeah, but that’s not the real world – surely the program misses things humans would catch. The researchers, however, say their model can even consider “sources of uncertainty,” such as “limited resources, unpredictable patient behavior (e.g., lack of medication adherence) and variable treatment response time.”
The researchers are excited about the possibilities: “Training a human doctor to understand/memorize all the complexity of modern healthcare, even in their specialty domain, is a costly and lengthy process,” they write. “…Does it make sense to continue to have human clinicians attempt to estimate the probabilistic effects of multiple actions over time, across multitudes of treatment options and variable patient characteristics, in order to derive some intuition of the optimal course of action?”
They insist that this will not replace doctors but instead “free them to focus on delivery of actual patient care” – that is, prescribing, ordering and operating at the thinking-machine’s commands, though we know the machines are
halfway ready to do that too. Soon enough the only part of the job left for meat puppets to perform will be the ethical one – and God help humanity when someone runs an ROI analysis on that.