Use of an oncology informatics knowledgebase to improve influenza vaccination rate in pediatric oncology patients

Abstract Introduction: It is well known that children with cancer are at an increased risk of severe influenza infections due to immunosuppression and co-morbidities. Serious influenza infections can lead to increased hospitalizations, outpatient visits, and delays in treatment. Influenza vaccination is an important strategy to decrease influenza-associated morbidity and mortality and has been a metric tracked for the US News and World Report in pediatric oncology. Electronic health record (EHR) based tools may be used to improve rates of vaccination. An oncology knowledgebase is an internal web portal that integrates relevant diagnostic and clinical data from the EHR to support clinical care, report disease-related outcomes and adverse events, and serves as a tool for numerous quality initiatives and research. Aspects of this data are manually curated, improving accuracy when compared to typical data extraction methods. The primary aim of this quality improvement (QI) initiative was to utilize the oncology knowledgebase to improve influenza vaccination completion rate in pediatric oncology patients in the 2024 influenza season to 88%. Methods: In collaboration with our Epic™ Clinical Applications and Bioinformatics Teams, an oncology knowledgebase was created that allows integration of diagnostic and clinical data with the EHR, namely through a customized oncology Epic™ Electronic SmartForm (ESF) tool. Primary drivers for this QI project included use of the knowledgebase to accurately identify the pediatric oncology population, detect vaccine eligible patients, and report associated vaccine data. Vaccine eligible patients were defined as patients aged 6 months or older who received chemotherapy after April 1, 2024, and were seen in our cancer program between October 1, 2024 and December 31, 2024. Using the knowledgebase, the 2023 influenza vaccination completion rate of 84% was used as the baseline data. The oncology knowledgebase created weekly automated reporting of unvaccinated eligible patients from October 2024 through December 31, 2024. Using this report, reminders via the EHR patient portal were sent to patients and pediatricians. The primary oncology team was also notified about which patients remained eligible for influenza vaccination. The data was compiled and analyzed in a run chart, with significance defined as a trend in 5 or more consecutive weekly datapoints going up or down. Knowledgebase data was directly compared to hospital-facilitated data extraction, which was also reported out weekly. Results: Between October 1, 2024 and December 31, 2024, 103 pediatric oncology patients were eligible for the influenza vaccine, including 62 patients with hematologic malignancies. Of the total eligible, 91patients were vaccinated, 54 of those vaccinated had a hematologic malignancy diagnosis. Reasons for declining influenza vaccination included religious beliefs, family perception of necessity post completion of therapy, and extended family influence. Influenza vaccination rates increased from a previous baseline of 84% to 88.34% within the designated reporting period of October 2024 through December 2024 in the pediatric oncology vaccine eligible patients. There was a consecutive upward trend in weekly vaccination rates in all but one reporting period. Discrepancies between the weekly vaccination rates from the oncology knowledgebase data compared to the hospital-facilitated data differed by as much as 38% at certain monthly timepoints. These differences can be attributed to the hospital-facilitated data excluding 17 patients that were eligible due to inaccurate logic for determining on and off therapy. Conclusion: The described oncology knowledgebase and similarly curated data tools allows for efficient and accurate data extraction, enabling QI efforts such as improving influenza vaccination rates. This QI project demonstrates how oncology knowledgebase data is more reliable than other EHR extraction methods, as it can more accurately curate patient populations. Future directions include comparing utilization and data retrieval from the oncology knowledgebase compared to other EHR data extraction methods for other QI initiatives, continued expansion of the cohort builder within the knowledgebase to further identify patient cohorts and assist in accurate data extraction, and implementation of streamlined data verification process and clinical decision support tools to enhance patient care and outcomes.