Examining National Representativeness of the Axon Registry
Andrew Wilson1,2, Kavita Nair3, Sarah Benish4, Gregory Esper5, Amanda Lien6, Aristotle Mante7, Karen Lundgren7, Becky Schierman7, John Ney8
1Greater Los Angeles VA Health System, 2University of California, Los Angeles, 3University of Colorado, 4University of Minnesota Physicians, 5Emory University, 6Verana Health, 7American Academy of Neurology, 8Edith Nourse Rogers VA Medical Center
Objective:

To determine external generalizability of the Axon Registry by comparison to two nationally representative, publicly available data sources, the National Ambulatory Medical Care Survey (NAMCS) and Medical Expenditure Panel Survey (MEPS).

Background:

The Axon Registry is AAN’s neurology-focused qualified clinical data registry that reports and analyzes electronic health record data from participating U.S. neurology providers.  It supports quality improvement within ambulatory neurology practices and promotes evidence-based clinical neurology. 

Design/Methods:
This is a cross-sectional, retrospective comparison of three national datasets: NAMCS (2012-2016), MEPS (2013-2017, 2019), and Axon Registry (2019).  We extracted patient demographics (age, gender, race, ethnicity), geographic location, neurological conditions (ICD codes grouped into eight clinical classifications for headache, epilepsy, cerebrovascular disease, multiple sclerosis, parkinsonism, dementia, spinal pain, and polyneuropathy), and healthcare utilization (neurology visits, MRI/CT neuroimaging studies and EEG/EMG neurophysiological studies).  

We performed descriptive analyses for pooled five-year data from the two public datasets compared to the Axon Registry.  We conducted Wald tests to evaluate the null hypothesis that comparable variables in Axon Registry and MEPS in 2019 were not different (alpha=0.05).
Results:
The Axon Registry recorded 645K people with 1.3M annual neurology visits (NAMCS, 14M weighted visits; MEPS, 9M weighted persons).  Compared to the pooled national surveys, Axon Registry has similar patient demographics, location, and neurological condition prevalence, and higher utilization of neuroimaging and neurophysiological studies.  Compared to MEPS 2019, Axon Registry represents a smaller proportion of children (6% vs. 9%, p=0.02), Asian, Pacific Islander, and indigenous persons (5% vs. 8%, p=0.001) and persons with epilepsy (10% vs. 13%, p=0.04), but more persons with dementia (8% vs. 6%, p=0.01), cerebrovascular disease (11% vs 8%, p<0.001), and persons in the Midwest (25% vs. 20%, p<0.001).
Conclusions:
The Axon Registry demonstrates high concordance with two nationally representative surveys, with implications for further recruitment of neurology practices into Axon and generalizability of Axon Registry data.