I love 2001: A Space Odyssey as much as the next person, but sorry, not sorry, despite what it and so many other movies would lead you to think, the “computer uprising” is the least of our concerns.
Actually, scratch that, it’s not concern at all—it’s a benefit.
Computers are getting smarter (or we’re getting smarter about how we use them), and you can find the advantages everywhere from fancy cars to healthcare technology. Case in point: Researcher Nigam Shah recently developed an algorithm to help locate people with undiagnosed familial hypercholesterolemia (FH).
This project was made possible using machine learning, a field of computer science in which computers are trained to act without directly being programmed to do so.
Sounds creepy when you think about it, right? Computers thinking and behaving on their own?
Some people consider machine learning to be the first step toward Hal-style artificial intelligence (aka smart and EVIL), but for now, it’s just a really high-tech way to take advantage of data patterns.
Shah’s FH algorithm, for example, uses data patterns by shifting through electronic health records.
It links information like age and reported cholesterol levels, and creates country-wide maps showing the likelihood that someone has FH.
Of course, being a fairly new program, the FH algorithm isn’t infallible. It’s intentionally set up to be broader in its searches, pinging people with regular-old high LDL cholesterol as well as those with FH.
“But hey,” the people in charge of the FH algorithm argue, “What’s wrong with that?”
It’s not a bad question. Why is it wrong for an algorithm to warn people they have dangerously bad cholesterol, no matter the cause? Especially if it means warning people with FH sooner.
However, detractors have raised the relevant concern that the algorithm is really just a way for big pharma companies to screw people over by pushing their medications on them earlier. The fact that the project is sponsored by the FH Foundation, which is largely supported by pharma donations, doesn’t help.
Personally, all I know are 3 things:
- More information tends to be better than no information when it comes to rare disease.
- Early diagnosis can drastically improve quality of life.
- This algorithm wouldn’t refuse to open the pod bay doors.
For me, that’s enough to warrant being excited about this technology.