Background: Pediatric influenza vaccination rates remain <50% in the United States. Children with chronic medical conditions are at higher risk of morbidity and mortality from influenza, yet most experience missed opportunities for immunization in outpatient settings. In an adult cohort study, 74% of patients who had not received the influenza vaccine before or during hospitalization remained unvaccinated through the rest of the season. Thus, inpatient settings represent another important opportunity for vaccinating an especially susceptible population. In addition, 4 published studies have shown promise in improving inpatient pediatric influenza vaccination. However, these studies had limited effect sizes and included interventions requiring ongoing maintenance with dedicated staff. In this study, we hypothesized that a clinical decision support (CDS) intervention designed with user-centered design principles would increase inpatient influenza vaccine administration rates in the 2019–2020 influenza season. Methods: We performed a workflow analysis of different care settings to determine optimal timing of influenza vaccine decision support. Through formative usability testing with frontline clinicians, we developed electronic health record (EHR) prototypes of an order set module containing a default influenza vaccine order. This module was dynamically incorporated into order sets for patients meeting the following criteria: ≥6 months old, no prior influenza vaccine in the current season in our medical system or the state immunization registry, and no prior anaphylaxis to the vaccine. We implemented the CDS into select order sets based on operational leader support. We compared the proportion of eligible hospitalized patients in which the influenza vaccine was administered between our intervention period and the 2018–2019 season (historical controls). To account for secular trends, we also compared the vaccination rates for hospitalized patients exposed to our CDS to those that were not exposed to the CDS during the intervention period (concurrent controls). Results: During the intervention period (September 5, 2019–November 1, 2019), influenza vaccine was administered to 762 of 3,242 (24%) of eligible patients, compared to 360 of 2,875 (13%) among historical controls (P < .0001). Among the 42% of patients exposed to the CDS, vaccination rates were 33% compared to 9% for concurrent controls (p < .0001). Our intervention was limited by end-user uptake, with some physicians or nurses discontinuing the default vaccine order. In addition, early in the intervention, some vaccines were ordered but not administered, leading to vaccine waste. Conclusions: CDS targeting eligible hospitalized patients for influenza vaccination incorporated early into the workflow of nurses and ordering clinicians can substantially improve influenza vaccination rates among this susceptible and hard-to-reach population.Funding: NoneDisclosures: None
Use of Clinical Decision Support to Increase Influenza Vaccination: Multi-year Evolution of the System
