Evaluation of autoimmune neurology testing clinical utility and stewardship initiatives

Abstract Testing for neural autoantibodies associated with autoimmune neurological diseases has an important role in clinical care. However, as they are niche tests for relatively rare diseases, they often have longer turn-around-times and more expensive costs than many other laboratory tests. In addition, due to the low yield of positive results, they often appear as attractive stewardship targets. Our objective was to review current utilization and ordering patterns and evaluate the impact of ongoing stewardship efforts to inform next steps regarding stewardship of these tests. Methods: Data on utilization and ordering patterns of autoimmune and paraneoplastic neural autoantibody panels from January 1 to December 31, 2023, was obtained from reference laboratory and electronic medical record (EMR) data. Additional data on best practice advisories from the most recent six months was obtained from the EMR. Results: Overall, we observed that providers tested 2270 patients both from our defined test menu (five total tests, ~89% of total orders) and through the miscellaneous route (12 total tests, ~11% of total orders), which allows providers to place orders for tests not currently on our menu. Just under one third of the orders were in cerebrospinal fluid (CSF) (n=725), while the remaining were for serum (n=1545). The percent positivity was generally higher in serum specimens, ranging from 0 to 50% per test (n=227 total positive specimens). The percent positivity in CSF specimens ranged from 0 to 40% per test (n=29 total positive specimens), which is generally consistent with previously published literature. Most of the orders were placed by neurology providers (56%). The best practice advisories (BPAs) currently in use are soft stops that target repeat tests in the same specimen type within thirty days of a previous order. Of note, these BPAs can only catch orders placed through the defined process. In six months, the BPA for repeat serum orders fired 104 times for 42 unique patients; the BPA for repeat CSF orders fired 96 times for 38 unique patients. The serum BPA resulted in elimination of the repeat order for 43% of patients (n=18), while the CSF BPA resulted in elimination of the repeat order for 24% of patients (n=9). We observed some situations in which the BPA appeared to be firing when previous orders were incorrectly collected and so not able to be sent out. Conclusion: Based on these data, we plan three main steps to continue to refine autoimmune neurology testing: 1) update our test menu to offer the most clinically useful orders through the defined route, 2) pilot review of tests ordered through the miscellaneous route, and 3) optimize BPAs to target situations with inappropriate utilization rather than technical issues.