Optimizing identification of Lyme disease diagnoses in commercial insurance claims data, United States, 2016-2019.

backgroundCommercial insurance claims data are a stable and consistent source of information on Lyme disease diagnoses in the United States and can contribute to our understanding of overall disease burden and the tracking of epidemiological trends. Algorithms consisting of diagnosis codes and antimicrobial treatment information have been used to identify Lyme disease diagnoses in claims data, but there might be opportunity to improve their accuracy.

methodsWe developed three modified versions of our existing claims-based Lyme disease algorithm; each reflected refined criteria regarding antimicrobials prescribed and/or maximum days between diagnosis code and qualifying prescription claim. We applied each to a large national commercial claims database to identify Lyme disease diagnoses during 2016-2019. We then compared characteristics of Lyme disease diagnoses identified by each of the modified algorithms to those identified by our original algorithm to assess differences from expected trends in demographics, seasonality, and geography.

resultsObserved differences in characteristics of patients with diagnoses identified by the three modified algorithms and our original algorithm were minimal, and differences in age and sex, in particular, were small enough that they could have been due to chance. However, one modified algorithm resulted in proportionally more diagnoses in men, during peak summer months, and in high-incidence jurisdictions, more closely reflecting epidemiological trends documented through public health surveillance. This algorithm limited treatment to only first-line recommended antimicrobials and shortened the timeframe between a Lyme disease diagnosis code and qualifying prescription claim.

conclusionsAs compared to our original algorithm, a modified algorithm that limits the antimicrobials prescribed and shortens the timeframe between a diagnosis code and a qualifying prescription claim might more accurately identify Lyme disease diagnoses when utilizing insurance claims data for supplementary, routine identification and monitoring of Lyme disease diagnoses.