That's not the correct way to think about it. The correct way to think about it is whether or not there's more harm with more data than the status quo.
Most medical professionals seem to think there's more harm than the status quo, but that doesn't change the fact that you're thinking about it wrong.
It is much easier to say after surgery: “All is good; no cancer, we took out a small bleb of tissue we thought was a tumor” than to say “Sorry, we thought that was a benign growth and that is why we did not recommend surgery 6 months ago, but it turns out we were wrong—it was a malignant tumor that has now spread.”
(source: am doctor)
In reality, having data people are biased towards action.
But I agree, a holistic interpretation - including the population level "bad decisions" - is the ultimately correct approach. The caution I would urge is not to allow "no true scotsman" arguments, ie "well it's a good thing as long as people are careful".
Many people aren't going to be, which is part of the calculus.
For some reason many people will argue ad nauseum that this very obviously true and simple fact isn't true
Data is meaningless without context and without knowing how reliable it is.
PSA tests, for example, provide data that sent a lot of people to an early grave, or to years of unnecessary treatment and anxiety.
I think about that a lot when people in various forums (including this one) cite academic papers or other information from areas they aren't experts in. They may be correct (technically correct), and have the correct data, but context might they be missing?
(The 50% number was of samples sent to a lab. Doctors aren't in the habit of taking random healthy people's lymph nodes and sending them for biopsy, so those patients had other symptoms of lymphoma that triggered the need for a biopsy of suspicious lymph nodes.)