PURPOSE: Compliance with computer alerts suggesting oral medication use during computerized order entry of i.v. medications was analyzed. SUMMARY: Using automated computerized clinical decision support (CDS) to suggest converting i.v. medications to oral alternatives can reduce medication costs for hospitalized patients, but prescriber noncompliance limits the effectiveness of such interventions. Clearer understanding of the factors associated with noncompliance to alerts may facilitate the design of more effective CDS systems. Electronic medical record data were retrospectively analyzed to measure the rate of compliance with a CDS alert that suggested converting to an equivalent oral form of a drug at the time of ordering the i.v. formulation. Multiple logistic regression was used to examine the associations among medication type, clinician characteristics, hospital service type, time of order, and compliance with the i.v.-to-oral conversion recommendation. The main outcome was compliance with the alert, measured at the level of the individual medication order. The mean +/- S.E. overall compliance rate was 18.7% +/- 0.6%. Compliance varied among the medications, with methyl-prednisolone having the lowest (8%) and famotidine the highest (32%) (p < 0.05). Nurses had the highest compliance rate (35%) while pharmacists had the lowest (10%) (p < 0.05). Medical house staff (19%) and medical faculty (21%) complied at similar rates. The intensive care units had lower compliance rates than did the medical-surgical ward (15% versus 21%, p < 0.05). CONCLUSION: CDS alerts to convert 12 i.v. medications to oral alternatives were developed and implemented in an urban tertiary hospital. Compliance rates for the alerts were relatively low and varied by medication, location, and clinician type.
PURPOSE: Compliance with computer alerts suggesting oral medication use during computerized order entry of i.v. medications was analyzed. SUMMARY: Using automated computerized clinical decision support (CDS) to suggest converting i.v. medications to oral alternatives can reduce medication costs for hospitalized patients, but prescriber noncompliance limits the effectiveness of such interventions. Clearer understanding of the factors associated with noncompliance to alerts may facilitate the design of more effective CDS systems. Electronic medical record data were retrospectively analyzed to measure the rate of compliance with a CDS alert that suggested converting to an equivalent oral form of a drug at the time of ordering the i.v. formulation. Multiple logistic regression was used to examine the associations among medication type, clinician characteristics, hospital service type, time of order, and compliance with the i.v.-to-oral conversion recommendation. The main outcome was compliance with the alert, measured at the level of the individual medication order. The mean +/- S.E. overall compliance rate was 18.7% +/- 0.6%. Compliance varied among the medications, with methyl-prednisolone having the lowest (8%) and famotidine the highest (32%) (p < 0.05). Nurses had the highest compliance rate (35%) while pharmacists had the lowest (10%) (p < 0.05). Medical house staff (19%) and medical faculty (21%) complied at similar rates. The intensive care units had lower compliance rates than did the medical-surgical ward (15% versus 21%, p < 0.05). CONCLUSION: CDS alerts to convert 12 i.v. medications to oral alternatives were developed and implemented in an urban tertiary hospital. Compliance rates for the alerts were relatively low and varied by medication, location, and clinician type.