Background: Our single academic medical center implemented a standardized Outside Hospital (OSH) communication process, encouraging physician completion of a dedicated templated patient transfer note for admission to any Department of Medicine (DOM) specialty. Prior outcomes analysis revealed that documentation with any transfer note, templated or non-templated, is significantly associated with improved outcomes of death within 24 hours and death within 30 days of admission compared to no note. Given these outcomes were unadjusted, we expanded our analysis and adjusted for patient complexity using Vizient data.
Methods: OSH transfer data was obtained through retrospective chart review and automated data export. Physician utilization of any transfer note, templated or non-templated, and objective clinical outcomes of code, ICU transfer, and death within 24 hours of admission, as well as death within 30 days of admission were similarly extracted. Analyzed OSH transfers included 1,784 transfers to a DOM service between 4/01/2022-3/31/2023. This OSH dataset was cross-referenced with Vizient data; patients were matched if the medical record number was the same in both datasets and, due to data discrepancies between EHR data and Vizient-coded data, if the admit dates were within 10 days. There was a 1,567/1,784 (88%) match. Length of stay (LOS) index (LOSi) was calculated by dividing Vizient LOS by Vizient expected LOS. Similarly, 30-day adjusted mortality and, a novel metric, of adjusted 24-hour mortality were calculated by dividing EHR-extracted 30-day and 24-hour mortality by Vizient 30-day expected mortality, respectively. For comparison ease, these ratios were normalized to an observed/expected of 1 for the no-note cohort.
Results: Templated notes (not any note), relative to no-note, are associated with significant decreases in 24-hour unadjusted (0.0% v 1.3%, p=0.0078) and adjusted mortality (0.0 v 1.0, p=0.047). Templated notes are associated with a significant decrease in unadjusted 30-day mortality (6.1% v 9.7%, p=0.018) but not adjusted (0.81 v 1.0, p=0.62). There continue to be promising, non-significant, trends in decreased RRTs and codes within 24 hours of arrival (0.0% v 0.13% p=0.39; 0.36% v 0.75%, p=0.36, respectively). Templated notes are associated with significant and non-significant trends toward increased telemetry and ICU upgrade, respectively (9.1% v 4.2%, p=0.0002; 2.7% v 2.0%, p=0.41). There was no difference in LOSi (1.15 v 1.19, p=0.48).
Conclusions: Accounting for patient complexity using Vizient data produced adjusted outcome findings that reinforce the positive impact of our transfer process and communication tool on patient outcomes but not on LOSi. The crossed-referenced data demonstrates both significant and nonsignificant decreases in mortality within 24-hours and mortality within 30 days, respectively. Our institution is working to power this study with increased template utilization.