Shallow linguistic analysis of a large corpus of drug prescriptions.

We report on first experiences from linguistic analyses of patient instructions from 19,8404 actual drug prescriptions regarding seven pharmaceutical products frequently prescribed in Sweden. The analysis includes expressions for amount, dose unit, dose interval, mode of administration, purpose and a few further details. Even simple processing seems useful to extract information from these short, rather formal text strings. We estimate the potential for calculation of prescribed dose from this material, and collected material gives a good starting point for more advanced linguistic analyses.