Randomized trial to improve prescribing safety in ambulatory elderly patients

Background

Inappropriate medication in elderly patients can potentially cause adverse events. Prevalence of these inappropriate prescriptions ranges from 12-40% in patients over 65 years of age. Few effective interventions exist to reduce this inappropriate medication usage in this vulnerable patient population.

Objective

To determine whether a computerized tool that alert pharmacists to potentially inappropriate medications for patients over 65 was effective in decreasing the rate of dispensing of these medications

Setting

Kaiser Permanente Colorado (KPCO), group model HMO.

Study Design

Randomized trial

Patients

59680 members of KPCO over 65 years of age living in the Denver-Boulder-Longmont area. Equal numbers (29,840) were randomized to the intervention and control group.

Intervention

The list of potentially inappropriate medications was pre-determined by a multidisciplinary group of experts. The Pharmacy Information Management System (PIMS) sent an electronic alert to the pharmacist when an intervention group patient was newly prescribed a potentially inappropriate medication. This alert prevented the prescription label from printing until the pharmacist intervened. The pharmacist consulted the prescribing clinician and either dispensed the medication if the indication was deemed acceptable or dispensed a safer alternative based on indication. This procedure of collaboration and consultation was also established by consensus before the trial began.

Outcome Measures

Primary outcome was the first dispensing of a targeted medication that deviated from the clinical guidelines used in the intervention. Secondary outcome was the first dispensing of targeted medications for inappropriate indications.

Results

For primary outcome, 543(1.8%) of intervention group patients were newly dispensed targeted medications compared with 644 (2.2%) in the usual care group. (P=0.002). For secondary outcome, 327 (1.1%) of intervention group compared with 447 (1.5%) in the usual care group were newly dispensed a targeted medication for inappropriate indications. Amitriptyline and diazepam were dispensed significantly less often for inappropriate indications in the intervention group.

Conclusions

This study demonstrated that a computerized pharmacy alert system in collaboration with pharmacists is effective in decreasing the rate of dispensing of potentially inappropriate medications for elderly patients over 65 years of age.