Long-term impact of a geriatric prescribing context

The medication-related death of a hospitalized older adult elucidated the inappropriateness of medication default doses in our electronic health record (EHR) for older adults. In response, we created and implemented the Geriatric Prescribing Context (GPC), an EHR-based set of age-specific dose and frequency defaults for patients 75 years and older, in July 2017. Inpatient medication orders aligned with GPC defaults and showed significant dose decreases at one year for nine of ten most commonly used medications. This follow-up investigation examined GPC alignment of dose and frequency over the 42-month time period after its implementation.

Order data for the ten most commonly used medications at OHSU Hospital were collected retrospectively from July 2016 through December 2020. We used Statistical Process Control charts to assess the proportion of medication orders aligning with the GPC's recommendations. Signals of special cause were evaluated to identify time periods when shifts in process averages likely occurred and suspected shifts were assessed using binomial proportion tests. We used RStudio (RStudio, Inc., version 1.2.5001) and Microsoft Excel (2016) to perform statistical analyses and control charts, respectively.

The preimplementation phase of all medications displayed no special causes. After significant initial improvement in 2017, control charts revealed three different patterns of performance. Eight medications maintained the initial improvement with one medication displaying a second significant improvement at a later date. Two medications showed a subsequent decline in performance not statistically different from baseline. Overall, eight of the ten medications were prescribed at more age-friendly doses and frequencies compared to baseline after 42 months.

The GPC is an effective method to support safer prescribing for hospitalized older patients, but long-term impacts may be medication-specific. Further investigation is needed to ensure appropriate prescribing across drug classes and understand the GPC's impact on patient outcomes like adverse drug events.