Background Inappropriate prescribing has been estimated to be as high as 40% in long-term care. The purpose of this study was to develop a computer program that identifies potentially inappropriate drug prescriptions and to test its reliability.

      Methods
      Potentially inappropriate prescriptions were identified based on modified McLeod guidelines. A database from one pharmacy servicing long-term care facilities in Ontario was utilized for this cross-sectional study. Prescription information was available for the 356 long-term care residents and included: the date the prescription was filled, the quantity of drug prescribed and the eight-digit drug identification number. The pharmacy database was linked to the computer-based program for targeting potential inappropriate prescriptions. The computer program's reliability was assessed by comparing its results to a manual search conducted by two independent research assistants.


      Results
      There was complete agreement between the computer and manual abstraction for the total number of potentially inappropriate prescriptions detected. In total, 83 potentially inappropriate prescriptions were identified. Fifty-three residents (14.9%) received at least one potentially inappropriate prescription. Of those, twenty (37.7%) received two potential inappropriate prescriptions and eight (15.1%) received 3 or more potential inappropriate prescriptions. The most common potential inappropriate prescriptions were identified as long-term use of non-steroidal anti-inflammatory agents and tricyclic antidepressants with active metabolites.


      Conclusion
      A computer program can accurately and automatically detect inappropriate prescribing in residents of long-term care facilities. This tool may be used to identify potentially inappropriate drug combinations and educate health care professionals.