2012 HSR&D/QUERI National Conference Abstract
3110 — Clinical Impact of Clinically Specific Interventionist ADE Alerts Served to Resident Physicians
Nebeker JR, Informatics Decision Enhancement And Surveillance (IDEAS) Center; Weir CR, IDEAS; Hoffman J, IDEAS; Nelson R, IDEAS; Stoddard G, IDEAS; Long BM, IDEAS; Sauer B, IDEAS;
Centers for Medicare and Medicaid Services are encouraging use of surveillance systems for adverse events. Systems for adverse drug events (ADE) are best studied. Most ADE surveillance systems have been designed to detect completed events with the goal of accounting for approximate rates of events. These surveillance systems are characterized by simple logical rules that lack clinical specificity. The goal of this study is to determine the impact of more complicated, clinically specific surveillance rules that are presented to inpatient providers with the goal of promoting measures to prevent or mitigate ADEs.
At a 110-bed teaching VAMC, we created a VistA modification to randomize patients on admission to the hospital to the intervention or control group. We created logic for 9 different ADE conditions. The physician teams of the intervention group received notification of a possible ADE through an un-signable progress note that provided other clinical information that supported the suspicion of an impending or progressing ADE. We analyzed time to the first clinically relevant response or resolution of ADE condition using Cox regression analysis. Responses were defined as starting a treatment specific to type of ADE, decrease/discontinuation of the drugs targeted by an alert, or either response. We also collected anonymous, qualitative data from internal medicine teams.
Over 1 year, 573 triggers fired during 126 unique ADE conditions. For any of the ADE surveillance rules, there was no significant difference in time to resolution or time to any intervention. Qualitative analysis revealed a heterogeneous impression of the rules. A roughly equal number of providers liked the alerts and found them useful as those who found them bothersome. Some providers did not report noticing any alerts.
Serving ADE surveillance alerts to inpatient residents did not change measureable responses to ADE conditions and did not change time in ADE state.
Future alerts should not target inpatient residents. Further study is warranted to determine whether ADE alerts addressed to clinical pharmacists may improve treatment of ADE conditions.