Background: Accuracy and transportability of the recorded outpatient medication list are important in the continuum of patient care. Classifying discrepancies between the electronic medical record (EMR) and actual drug use by the root cause of discrepancy (either system generated or patient generated) would guide quality improvement initiatives.
Objectives: To quantify and categorize the number and type of medication discrepancies that exist between the medication lists recorded in EMRs and the comprehensive medication histories obtained through telephone interviews conducted by a team of nurses providing advice to health plan members at the Palo Alto Medical Foundation in Palo Alto, California.
Methods: The study was conducted as a retrospective comparison of EMR medication lists with information obtained by patient interview. Interview data were obtained by a review of telephone calls made to a nurse advice line by health plan members seeking information about sinusitis, urinary tract infection, acute conjunctivitis, pharyngitis, emergency contraception, or mastitis. As part of the advice protocol, a medication reconciliation process was conducted between July 1 and December 31, 2006. Changes to the medication list made during the telephone visit were extracted, categorized, and evaluated by the study's principal investigator. Data extraction included the number and type of identified medication discrepancies, patient age, gender, and condition that prompted the telephone contact. A modified version of the Medication Discrepancy Tool (MDT) was used to categorize medication discrepancies as either system generated (e.g., failure of the provider to update a medication list) or patient generated (e.g., failure of the patient to report use of an over-the-counter product).
Results: A total of 233 discrepancies were identified from 85 medication reconciliation phone visits, averaging 2.7 per medication list. The most common type of discrepancy was a medication recorded in the EMR but no longer being used by the patient (n=164, 70.4%), followed by omission from the EMR of a medication being taken by the patient (n=36, 15.5%). 79.8% (n=186) of the discrepancies were attributed to system-generated factors, whereas 20.2% (n=47) were patient generated. Approximately half (n=118, 50.6%) of the discrepancies fell into 4 broad American Hospital Formulary System therapeutic classifications: anti-infective agents (14.2%), anti-inflammatory agents (14.2%), analgesics (12.4%), and vitamins (9.9%). The most common patient-generated discrepancy was omission of a multivitamin (n=13, 27.7%), and the most common system-generated prescription drug discrepancy was expired entry for the intranasal corticosteroid mometasone furoate (n=12, 6.5%).
Conclusion: Discrepancies in the outpatient setting were common and predominantly system generated. The most common discrepancy was the presence in the EMR of a medication no longer being taken by the patient. Adding foreseeable end dates to prescription drug orders at computerized order entry might be considered in an effort to improve the accuracy of the outpatient medication list. Reliable systems to involve patients in routinely reconciling EMRs with actual medication use may also warrant examination. The MDT methodology served as a useful qualitative guide for evaluating discrepancies and developing targeted means for resolution.