Hacker News Reader

Comment by codewiz | Hacker News Reader

Back when I lived in Japan, my employer-sponsored yearly health screening had an optional head MRI exam. I could book it a few days before, and would take 15 minutes with a single technician.

Now it costs around $300: https://www.tokyomidtown-mc.jp/en/pdf/optional_menu.pdf

View on HN

MR screening is a great way to increase unnecessary biopsy and stress etc except in a very narrow range of conditions.

Edit: https://bmjopen.bmj.com/content/12/1/e056572

That’s a problem with the interpretation, and how results are presented to patients. The MRI scanner is working as intended but the humans are getting the probabilities wrong, as usual.

Specificity is not always high. Yes, it’s an interpretation issue, but it’s a know problem. A better radiologist might help, but the issue is that many things look the same or very very similar on MRI. Evidence based medicine is the way to go.

This is a really weird claim to me. It can’t be about the MRI itself, it must be about the surrounding politics or legal environment.

How can it ever be bad to have more objective data? Having a test result to ignore is still strictly better than not testing at all. You still have the option to ignore or discount the test results if you want to. If you’re forced to do otherwise that’s entirely a human problem, not a technological one.

> How can it ever be bad to have more objective data? Having a test result to ignore is still strictly better than not testing at all.

That’s why you base healthcare on data. And that’s what the data shows.

I still don’t get it.

If MRIs lead to unnecessary biopsies, then don’t do the biopsy. What exactly does the data show? If doing a test leads to an unnecessary biopsy, how can it be the test’s fault? Tests aren’t moral agents. The MRI machine can’t make any healthcare decisions for anyone.

Reply on news.ycombinator.com