Skip to main content
Psych Congress  

Using Electronic Health Records to Estimate Disease Severity and Prevalence of Agitation in Alzheimer's Disease/Dementia: Challenges and Opportunities

Rachel Halpern, PhD; Jerry Seare, MD; Junliang Tong, MS, MA; Ann Hartry, PhD; Anthony Olaoye, MBA, MS; Myrlene Aigbogun, MPH
Otsuka Pharmaceutical Development & Commercialization, Inc. and Lundbeck

This poster was presented at the 30th annual  Psych Congress, held Sept. 16-19, 2017, in New Orleans, Louisiana.

Background: Neuropsychiatric symptoms such as agitation contribute to disability, institutionalization, and diminished quality of life for patients with Alzheimer's dementia (AD). Agitation prevalence estimates vary widely and little is published on prevalence by AD severity. Electronic health records (EHRs) provide clinical insight by leveraging information from patients' medical records. We sought to estimate the prevalence of agitation symptoms overall and by disease severity.

Methods: This retrospective cross-sectional study examined patients with dementia, including AD, from the Optum linked EHR and claims database. Patients with AD/dementia diagnoses or abstracted notes records produced with natural language processing (NLP) and indicating AD/dementia were identified from January 2008- June 2015. AD severity was based on notes records with Mini-mental State Examination scores or explicit notations of mild, mild-to-moderate, moderate, moderate-to-severe, or severe; patients without information were "unknown". Agitation symptoms were identified per the 2015 International Psychogeriatric Association definition.

Results: The sample comprised 511,994 patients; 44% had agitation during the two-year observation period. Top agitation categories were agitation (30%), restlessness (20%), irritability (19%), and behavior (17%). AD/dementia severity was unknown for 82% of patients. Of the 18% with staging, 55% were mild, 22% severe, 10% moderate, 8% mild-to-moderate, and 5% moderate-to-severe.

Conclusions: The prevalence of agitation in this study is consistent with previous estimates. However, it was difficult to estimate agitation prevalence by severity given the preponderance of unknown staging. EMR notes are not sufficient to evaluate severity perhaps attributable to physician charting practices and/or information the current NLP does not capture.

Back to Top