I'd add that errors happen especially with the most accurate assays. The mass spectrometry-based tests are known to be temperamental and tricky to do right. I've had the sensitive estradiol test botched on a couple occasions, while the immunoassay test has remained consistent and useful, albeit a few percentage points higher.
Do you inspect your injection sites closely when injecting subcutaneously? Depending on the dose, the site and the how frequently the site is used, it's possible to have significant loss of the medication in the minutes after injecting. I'd assume this loss is more common with SC than with IM, making SC appear less effective in some cases.
