My experience with a "liberated" Flir E4 is that the temperature resolution is more than sufficient to detect elevated skin temperatures (you can set temp windows to get a nice colour gradient in the relevant temp range), yet the absolute calibration lacks. What that means: You could very well use it to spot "hotter" individuals in a crowd of people that are supposed to have normal body temperature. You cannot do this reliably by using the E4 to measure spot temperature individually. The internal calibration that the E4 is doing once every few minutes will adjust absolute temperature easily by + / - 1K, so the readings will spread by this amount. I cannot tell for other thermal cameras, but I assume that in the lower price range, the situation would be similar.
Oh, just noticed, that a very old one... Why bumping this stuff, I guess it rather belongs in the "Thermal Imaging" section anyway.