From the 2023 HVPA National Conference
Alexandra Diaz-Barbe, BA (University of Chicago Pritzker School of Medicine) Emma He, BA, Misha Tran MD, MS, Matthew Cerasale MD, MPH
Improving medication reconciliation is a difficult, yet crucial task in improving the safety of hospitalized patients. As part of implementing a best practice medication history, documentation of the level of confidence in an admission medication reconciliation was added to the history and physicals of our hospital medicine group. The intent was to identify patients that would need additional time spent on the reconciliation later in the encounter and to potentially flag for review by pharmacy. After documenting medication reconciliation confidence for over a year, the veracity of the statements had not been evaluated to ensure it was an effective tool to leverage additional resources.
The aim of this study was to evaluate the accuracy of medication reconciliation confidence levels and to determine if any patient factors were associated with 1) confidence in reconciliations and 2) accuracy of the medications.
The study was completed at a large, urban, academic tertiary referral center with a large hospital medicine group. The timeframe of the study was December 31, 2022 through February 2023. Demographics of patient encounters plus medication reconciliation confidence (rated as high, low, or incomplete) were obtained for all hospital medicine admissions. For analyses, low confidence and incomplete were grouped together. A random sample of approximately 60 high confidence and 60 low confidence/incomplete reconciliations were reviewed. Medication reconciliations were identified as inaccurate if discharge medications contained “new” or “stopped” medications that could not be linked to the patient’s medical course as documented in the discharge summary/progress notes. If there were additional inconsistencies between discharge medications and prior clinical documentation, the cases were flagged as potentially inaccurate. Demographic information between high confidence and low confidence/incomplete reconciliations were compared for the entire population and sample group.
A total of 811 encounters with history and physicals written by the hospital medicine team were completed during the study period. Medication reconciliations were noted as high confidence in 92.0% of encounters, compared to 3.2% for low confidence, and 4.7% for incomplete. Incomplete or minimal confidence medication reconciliations were significantly more likely to be Black (9.4% vs. 4.0%, p= 0.0097), ≥65 years old (17.7% vs. 6.3%, p< 0.0001), and on Medicaid (11.9% vs. 6.4% p= 0.011). Of the random sample from chart review, 57 high confidence and 53 low/incomplete were included in the analysis. In low confidence/incomplete cases, no medication changes were noted following admission in 43.4% of cases. In low confidence/incomplete cases, 52.8% had medication changes that were unaccounted for, compared to 35.1% in high confidence medication reconciliations. For low confidence/incomplete cases, 75% had “possible inaccurate medications,” while this was true for 54% of high confidence cases. No factors were statistically associated with unaccounted for medications (Table 1), but low confidence/incomplete trended towards significance.
Providers were more likely to identify their medication reconciliations as low confidence/incomplete for patients who were Black, elderly, and on Medicaid. Medication reconciliations were unlikely to be updated during hospitalization, regardless of provider confidence. Low confidence/incomplete medication reconciliations were more likely to be inaccurate, but a third to half of high confidence reconciliations also likely had errors.
Ensuring patients leave the hospital with their medications reconciled is an essential component of high-value care. Compiling an accurate medication list prior to discharge arms patients with knowledge about their treatments and prevents subsequent medication-associated errors by healthcare providers.