3086 — NoteAid: A System that Helps Veterans Comprehend their own EHR Notes

Yu H, VA Central Western Massachusetts, Northampton, MA; Ramesh B, VA Central Western Massachusetts, Northampton, MA; Brandt C, VA Connecticut Health Care System, West Haven, CT; Zulman D, VA Palo Alto Health Care System, Menlo Park, CA; Houston T, Bedford VA Medical Center, Bedford, MA;

Objectives:
In January 2013, the VHA began to make clinical notes in the electronic health record (EHR) available through the My HealtheVet portal. However, studies have shown that patients are confused by EHR notes, especially patients in vulnerable groups (e.g., low literacy, low income). Such confusion may result in unintended increases in service utilization and lead to changes in perceptions that may disrupt patient-provider relationships. We are developing NoteAid, a multi-component, natural language processing (NLP) system that translates medical jargon into lay terms and provides definitions and links to related educational material from trusted resources. Here we report the development and evaluation of NoteAid.

Methods:
NoteAid has two main components: a knowledge resource comprised of patient education materials and the NLP system that links EHR notes to the knowledge resource. The knowledge resource integrates the medical jargon-lay term-definitions from the Unified Medical Language System and the MedlinePlus. The NLP system links the complex medical jargon that appear in the note to simple consumer oriented definitions and explanations from the knowledge resource. We evaluated NoteAid using 40 de-identified EHR notes (20 progress notes and 20 discharge summaries). We recruited 64 subjects. Each subject reads an assigned EHR note before and after NoteAid and self-scores comprehension (on a scale of 1 to 5, with 1 the poorest and 5 the best comprehension). Each subject completes the evaluation of either 20 PGNs or 20 DSs. We used Flesch-Kincaid grade level (FKGL) to score readability of each note.

Results:
Our results show a negative association between FKGL of a note and its subject's self-reported comprehension score (Spearman rho coefficient, -0.807; p < 0.0001). NoteAid improves self-reported EHR comprehension in both PGNs and DSs and the improvement is statistically significant (Wilcoxon signed-rank test, p < 0.05).

Implications:
NoteAid improves EHR note comprehension.

Impacts:
We are preparing to implement and evaluate NoteAid within the VA. NoteAid has the potential to facilitate translation of clinical records to patients, which could enhance patient engagement and lead to improved self-management and clinical outcomes.