3 ms·

I worked for a medical diagnostics company for 4 years and this is an incorrect view. Though I'm not fluent in the regulatory side I do know that most of our te

by dikdik 11y ago

I worked for a medical diagnostics company for 4 years and this is an incorrect view. Though I'm not fluent in the regulatory side I do know that most of our tests were not FDA approved. Most tests/assays were developed in-house, however, a few tests were licensed from other companies and all of those WERE FDA approved. Another interesting caveat, when the FDA approved tests failed we could not dissect the process and try different variations/dilutions and were only able to give general reasons for test failure (it was sometimes referred to as "black box" testing). However, this was not the case for in-house developed non-FDA approved tests. For those, we had much more control and could give pretty definitive reasons for testing failure and sometimes we could re-test failed specimens with slight variations and obtain a valid result.

icegreentea 11y ago
The term you're looking for in FDA speak is 'Laboratory Designed Test', I've included a link to the FDA's guidance document on LDTs below. The reason why FDA approved tests frequently work as 'black boxes' is because the FDA requires that approved tests reach specific performance characteristics. Typically in order to prove these performance characteristics, companies must constrain the parameters of the testing. Once they do that that test is only considered 'valid' over those parameters. The level of error reporting is dependent on the company who develops the test. The FDA is working to subjecting certain LDTs (basically stuff that guides treatment choices) to pre-market review, so a lot of these gaps will likely begin to be closed. One of the things that Theranos kinda trumpeted earlier is that their nanotainer HPV test was technically a LDT and therefore did not require a 510k, but they got it anyways. http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/InVitroDiagnostics/ucm407296.htm http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedur...
raverbashing 11y ago
What would be an example of a test that fails, and how does this happen?
- semi-extrinsic 11y ago
  I'm not very familiar with blood lab tests, but I believe many of them are some form of chromatography, e.g. HPLC, which is "just" a high-tech version of pee-on-a-stick tests. For those, there is a "control" indicator that should always show positive, and failure is when it doesn't show positive.
- dikdik 11y ago
  Examples: Not enough DNA in the sample (this normally comes from the White Blood Cells), the primers used in a PCR test (amplifying the DNA) do not bind properly, other markers for various tests don't bind correctly, etc. Not-binding correctly could be a problem with the whole sample (ie something in the blood is throwing off the test, outside contamination by a bad draw, not documenting draw date properly and sample is expired). With solid tissue samples (often embedded in paraffin wax), there can be extensive problems with the sample being prepared incorrectly. These are normally prepared on-site at your doctor's office and there is a surprising amount of variation.

by dikdik 11y ago

icegreentea 11y ago

The term you're looking for in FDA speak is 'Laboratory Designed Test', I've included a link to the FDA's guidance document on LDTs below. The reason why FDA approved tests frequently work as 'black boxes' is because the FDA requires that approved tests reach specific performance characteristics. Typically in order to prove these performance characteristics, companies must constrain the parameters of the testing. Once they do that that test is only considered 'valid' over those parameters. The level of error reporting is dependent on the company who develops the test. The FDA is working to subjecting certain LDTs (basically stuff that guides treatment choices) to pre-market review, so a lot of these gaps will likely begin to be closed. One of the things that Theranos kinda trumpeted earlier is that their nanotainer HPV test was technically a LDT and therefore did not require a 510k, but they got it anyways. http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/InVitroDiagnostics/ucm407296.htm http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedur...

raverbashing 11y ago

What would be an example of a test that fails, and how does this happen?

semi-extrinsic 11y ago
I'm not very familiar with blood lab tests, but I believe many of them are some form of chromatography, e.g. HPLC, which is "just" a high-tech version of pee-on-a-stick tests. For those, there is a "control" indicator that should always show positive, and failure is when it doesn't show positive.
dikdik 11y ago
Examples: Not enough DNA in the sample (this normally comes from the White Blood Cells), the primers used in a PCR test (amplifying the DNA) do not bind properly, other markers for various tests don't bind correctly, etc. Not-binding correctly could be a problem with the whole sample (ie something in the blood is throwing off the test, outside contamination by a bad draw, not documenting draw date properly and sample is expired). With solid tissue samples (often embedded in paraffin wax), there can be extensive problems with the sample being prepared incorrectly. These are normally prepared on-site at your doctor's office and there is a surprising amount of variation.