Development of cohort definitions and algorithms to identify patients with Lennox-Gastaut syndrome or Dravet syndrome from real-world administrative healthcare databases

To develop cohort definitions and algorithms that can be applied to a range of real-world retrospective data sources in the United States, France, Germany, Italy, Spain, United Kingdom, China, and Japan to identify patients with Lennox-Gastaut syndrome (LGS) or Dravet syndrome (DS) to aid future research. The study was conducted in 3 stages. A targeted literature review was used to identify retrospective healthcare database studies identifying LGS or DS populations and develop overall draft cohort definitions and algorithms. Country-specific research explored the diagnosis codes and antiseizure medications (ASMs) with available indications for LGS or DS, with the findings used to adapt the cohort definitions and algorithms to country-specific settings. Physician interviews were conducted to validate and refine the draft country-specific cohort definitions and algorithms, and better understand the diagnosis and treatment of patients with LGS and DS in each country. Forty-eight publications were reviewed; 25 focused on LGS only and 14 on DS only. LGS-specific and DS-specific diagnosis codes were identified in the United States, Spain, and United Kingdom; local codes in France, China, and Japan; and no LGS-specific or DS-specific diagnosis codes in Germany or Italy. ASMs only indicated for LGS or DS were available across all countries except China. Multiple seizure types, developmental delay, and intellectual disabilities were considered by the physicians to be key diagnostic features. The country-specific definitions for all 3 approaches (specific diagnostic codes, ASM indicated for LGS or DS only, and broader epilepsy diagnostic codes) were refined further using consensus from the physician interviews. To our knowledge, this is the first study to comprehensively develop country-specific definitions and algorithms for LGS and DS across several countries that provide a solid foundation toward identifying such patients from real-world retrospective databases. However, additional research and validation using real-world data sources are warranted.