Purpose: To identify meaningful, measurable process measures for high-quality MR obtainable from electronic health records (EHR), and to determine if performance on these measures responded to a campaign focused on improving MR.
Description: Our project defined MR as the process of identifying an accurate home medication list, and recording those data correctly in our EHR (Epic). Baseline MR audits of hospitalized patients on the Hospital Medicine Service using in-person interviews and chart review revealed high rates of medication discrepancies.
Using our direct observations, we then developed process indicators easily measured in our EHR. Instances of poor MR were more likely in patients with no listed “prior to admission” (or home) medications and in patients who the prior to admission medication list was not used to generate an inpatient medication list (and instead often had home medications entered as new hospital medications).
We created an Epic data report that we used to determine whether: 1) reconciliation actions (e.g. order/don’t order/change) were performed for each “prior to admission” medication, 2) a patient had medications entered into the “prior to admission” list, and 3) patients were discharged with duplicate medications on their discharge medication list.
Beginning in July 2015, housestaff, with support from an interdisciplinary team, led improvements, which included provider education, audit and feedback, financial incentives, and EHR prompts to reconcile prior to admission medications. Observations revealed some providers were unaware of the need to consult multiple sources to obtain the best possible medication list, while many others were unfamiliar with the appropriate EHR workflow. An EHR flag was created in provider’s patient lists indicating that all prior to admission medications had not been addressed (Figure 1). Pocket-sized cards and a video detailing best practices for MR were incorporated into monthly housestaff orientation.
After the intervention, analyses with statistical process control charts showed special cause variation for acting on all prior to admission medications, having prior to admission medications entered into the chart, and patients with no duplicate medications in the discharge list. These changes coincided with the improvement efforts (Figure 2).
Conclusions: In this study, strategies proven in prior studies to improve MR (education, developing standard work, and audit and feedback) increased performance on our EHR-based MR audits. This suggests these measures, all easily acquired from most EHRs, may be surrogate measures for MR that enable dynamic audit.