We noticed that after several years of slowly rising code counts in LNC versions, there was a largish drop of ~35 000 codes between 2.42 and 2.46 (we skipped loading 2.44). Our QA suggests this is a real finding, not a bug in our loading process. We were curious if the experts have any comments that would explain this change, or could verify that it is both true and offer an explanation on what changed to cause this drop. Was this a one-time QA process?
Our counts agree with yours when the codes for Loinc parts are not included in the data set; we do include them and saw the large drop between 2.42 and 2.46. I don’t know if this reflects a cleanup of the Regenstrief data, or a change in the UMLS files from which we derived the codes or some other issue.
I appreciate the reassurance that we aren’t missing any of the key LOINC codes.