OBJECTIVES: Linking indication and prescribed medicine is central to comparative effectiveness research and evidence based medicine. Little research exists regarding mechanisms for acquiring indications for inpatient medication use.  The main objective of this pilot study was to assess an intervention during computerized physician order entry (CPOE) for gathering indications for inpatient medication use. METHODS: Three medications were selected based on clinical interest and likelihood of off-label use: the proton pump inhibitor (PPI), lansoprazole, the intravenous immune globulin (IVIG), Flebogamma®, and recombinant Factor VIIa (rFVIIa).  Alerts were triggered at the time of CPOE for each medication.  To minimize physician burden, alerts were not triggered when appropriate indications/diagnoses were already listed in the electronic problem list. The alerts requested clinicians to enter either a labeled or off-label indication for the medication order. The  lansoprazole, IVIG, and  rFVIIa interventions ran for 60, 175, and 93 days, while 873, 56, 25 alerts were displayed to clinicians, respectively.  A random sample of 100 alerts was chosen for the lansoprazole analysis.  Expert chart review was used as the actual indication gold-standard.   RESULTS: For lansoprazole, 81 % of the medication orders were off-label and 71% were off-label if all indications of the PPI drug class were considered.  For Flebogamma®, 100% of the orders were off-label and 86% were off-label when considering all indications of the IVIG class. All orders for rFVIIa were off-label. The match between indications entered through the alert system and chart review was 61% for lansoprazole, 46% for IVIG, and 40% for rFVIIa . CONCLUSIONS: This pilot of indication based prescribing during CPOE illustrates many challenges: number of indications, coding (ICD9 codes are limited), and the tradeoff of information versus nuisance. The indication data generated was not highly accurate.  The high rate of off-label use noted suggests the value of efficiently defining prescribing indications.