4 ms·

I feel like the cost of a false positive in medicine is lower than the cost of a false negative.

jetti 9y ago
That entirely depends on the condition being tested and how many other ways there are available to test for that condition.
stordoff 9y ago
That entirely depends on, e.g., the false positive rate and the specificity of the test. Couple of quick examples of the top of my head: * A false positive for a terminal condition or cancer may have a poor effect on the patient's mental well-being AND make them less likely to submit to testing in the future (see e.g. https://www.ncbi.nlm.nih.gov/pubmed/23540978 https://www.ncbi.nlm.nih.gov/pubmed/23540978 - "We conclude that the experience of having a false-positive screening mammogram can cause breast cancer-specific psychological distress that may endure for up to 3 years, and reduce the likelihood that women will return for their next round of mammography screening."). Depending on your false positive rate, that may result in worse outcomes overall. * I had an MRI on my head a few years ago, and a _very_ small growth (<2mm IIRC, small enough that some future scans weren't even certain it was there) was identified. Everyone involved is certain that it is an irrelevant finding, but I'm now stuck with yearly MRIs - both a cost to the NHS and a nightmare for me (it sets off an anxiety disorder). * Parent suggests getting a hsCRP test. Whilst potentially useful, it's also non-specific - it'll be raised by pretty much any recent illness/injury/general inflammation, and will always be raised if, e.g., you have arthritis. You would likely need further testing to determine the cause (not necessarily free of risk/cost), and your false positive rate is likely to be higher if you are testing without other reasons to suspect it will be useful. That's not to say all preemptive tests are a bad idea -- that's clearly not the case -- but as I initially said, it's a balancing act. Blindly added more early testing will not necessarily improve patient outcomes overall; they may in fact have a detrimental effect.

That entirely depends on the condition being tested and how many other ways there are available to test for that condition.

stordoff 9y ago

That entirely depends on, e.g., the false positive rate and the specificity of the test. Couple of quick examples of the top of my head: * A false positive for a terminal condition or cancer may have a poor effect on the patient's mental well-being AND make them less likely to submit to testing in the future (see e.g. https://www.ncbi.nlm.nih.gov/pubmed/23540978 https://www.ncbi.nlm.nih.gov/pubmed/23540978 - "We conclude that the experience of having a false-positive screening mammogram can cause breast cancer-specific psychological distress that may endure for up to 3 years, and reduce the likelihood that women will return for their next round of mammography screening."). Depending on your false positive rate, that may result in worse outcomes overall. * I had an MRI on my head a few years ago, and a _very_ small growth (<2mm IIRC, small enough that some future scans weren't even certain it was there) was identified. Everyone involved is certain that it is an irrelevant finding, but I'm now stuck with yearly MRIs - both a cost to the NHS and a nightmare for me (it sets off an anxiety disorder). * Parent suggests getting a hsCRP test. Whilst potentially useful, it's also non-specific - it'll be raised by pretty much any recent illness/injury/general inflammation, and will always be raised if, e.g., you have arthritis. You would likely need further testing to determine the cause (not necessarily free of risk/cost), and your false positive rate is likely to be higher if you are testing without other reasons to suspect it will be useful. That's not to say all preemptive tests are a bad idea -- that's clearly not the case -- but as I initially said, it's a balancing act. Blindly added more early testing will not necessarily improve patient outcomes overall; they may in fact have a detrimental effect.