Most lab reports are PDFs with a text layer, so the readings are lifted straight out of the file, character for character. A photograph or a scanned image saved as a PDF is not read at all today, which is why a text-based PDF is what you want to upload. Nothing is interpreted at this stage — the result is a list of marker names, values and units, plus whatever reference interval your lab printed, kept exactly as written.
Every reading is shown beside what your lab printed, so you are checking against the paper in front of you rather than against our conversion. Correct anything that looks wrong, delete a duplicate, and confirm the date the sample was taken — that date orders everything that follows, so it cannot be guessed. A reading we could not interpret is not an error to resolve; it stays in the table, uninterpreted, and never blocks you.
At most three markers are put forward, each with the reason it ranked there. A marker inside our range is never one of them, and neither is one we have no sourced range for, or one still carrying a question from the check step. Our reference range appears beside the one your lab printed, because the two often disagree and you should be able to see that they do.
A plan comes with a suggested retest window rather than a daily routine. When the next report arrives it becomes another column beside this one — nothing is overwritten, and nothing is superseded. Where the same marker appears in both, the change is shown with its caveats: different labs, different methods and different times of day all move numbers on their own.