Hello,
I am try to author a paper, and I am looking for a reputable source that reports the standard deviation of lab tests. Most of the sources I have found are not super solid and only list reference ranges. Any help would be appreciated.
If I can't find anything that explicitly states standard deviations, I know I can calculate them based on reference ranges because they represent 4 SD's. I would feel better if I could just find the norms plainly stated.
Thank you!
If you're talking about reference range standard deviations for certain analytes, this information really isn't going to exist for the most part. A significant part of my job is determining reference ranges for assays used at a major medical center/global reference lab.
First, reference ranges are both unique to the patient population as well as the analytical method and even specific analyzer used for that analyte. There can be huge (like 20-30% differences in results between methods in some cases, such as a hormone measured by immunoassay vs. LC tandem mass spec.) This is why you aren't going to find one set of ranges and standard deviations for say alk phos, even for a photometric method by two different manufacturers. Even a slight reagent reformulation on an instrument can cause a shift.
Then there isn't always consensus on if/how the ranges are stratified, such as by age or gender and at which age breaks, etc. There's tons of variability there.
Second, reference ranges aren't always based on calling anything within +/- 2 SDs of the mean as within the reference range. It used to be that the central 90% was used to define the reference range, but now there's a trend to move to the central 95%. Despite that, many manufacturers' package inserts still report ranges of the central 90% and most labs just verify package insert ranges as it isn't feasible for them to establish their own ranges. So depending on the source of the ranges you're looking at, trying to back calculate SDs is going to be hard because it might not be explicitly stated how that range was arrived at. The older the literature source the less details you tend to get.
Plus manufacturers' ranges can be off due to the fact that they tend to use convenience samples from presumed healthy people. We've had instances where we couldn't verify their ranges using samples from a better screened population. In the literature you'll also frequently find pretty different ranges for the same type of patient population on the same analyzer just because of differences in the inclusion and exclusion criteria used for the study. Then you get to scrutinize that to try to figure out which one seems better.
Then you have things like cardiac markers like troponin, where the central 99% is used for the reference range.
Labs really struggle for reference ranges and every time we have a reagent formulation or slight change somewhere, we're going back and evaluating where our ranges came from and whether or not they are still appropriate as things done to determine them 10 years ago aren't always ideal.
If you can explain a little bit more about what it is you're trying to do or how you're planning in using this info, I can probably provide you with some additional guidance as this has become a bit of an area of expertise for me. PM if you'd like and would rather not go into detail in here.
TLDR: reference ranges are very complex and kind of a pain in the ...
