Sig / Sig Line

Sig (from the Latin signa, meaning “write”) or sig line is the part of a prescription that contains directions for use. It describes how the patient should take the medication: how much, how often, by what route, for how long, and any special instructions.

A typical sig line might read: 1 tab PO TID x 10d (one tablet by mouth three times daily for ten days), or handwritten in plain language: Take one tablet by mouth three times a day for ten days.

Why sig lines are hard for OCR-only tools:

OCR can read the characters in a sig line but cannot reliably interpret them. Sig notation uses abbreviations (QID, BID, PRN, NTE, HS), shorthand (1-2 tabs, ½ tab, q4-6h), and freehand variations that differ across prescribers and clinics. OCR produces the characters as written; it does not know that 1 tab PO TID means Take 1 tablet by mouth three times daily.

In Kroll, sig lines must be entered in a specific format for the label to print correctly. If an automation tool pastes the raw OCR output into the sig field, the result is often an unusable or incorrect label.

What AutoRx does with sig lines:

AutoRx parses sig lines using a comprehension model — not just character extraction. It interprets abbreviations, resolves ambiguous shorthand, and maps the result to Kroll’s sig format. Unusual or unresolvable sig lines route to the exception queue with the original document attached, rather than producing a malformed entry.