RWE Expert UCB Pharma, Raleigh, USA West Chester, United States
Background: Amyotrophic lateral sclerosis (ALS) is a rare, progressive, neurodegenerative motor neuron disease (MND), which often leads to respiratory failure and death. There is an increasing need to better understand the ALS patient journey from diagnosis to treatment and outcomes using real-world data (RWD). However, identifying ALS cases using RWD presents several challenges due to the rarity and heterogeneity of ALS and the differences in database coding systems.
Objectives: To develop a robust algorithm to identify patients with ALS in US claims and UK primary care databases and to estimate the prevalence of ALS in each.
Methods: MarketScan claims, and the UK Clinical Practice Research Datalink (CPRD) databases were searched for diagnosis codes for ALS or MND, two therapies (riluzole and edaravone) and neurology/neurosurgery visits with 12-month enrollment prior to 1 January 2011. The study period was 1 January 2011 until 31 December 2020. An existing algorithm was expanded to identify unspecific ALS cases (MND with an ALS prescription). For MarketScan, the main algorithm required: ≥1 ALS diagnosis code (ICD-9/10, 335.20/G12.21) and a treatment prescription; ≥2 ALS codes and ≥1 visits; or ≥1 ALS code and ≥3 visits. Additionally, ≥1 MND code and an ALS prescription was used to identify patients with ALS among those with MND codes only (proxy cases). For CPRD, SNOMED/Read coding systems were used. Age-specific prevalence was examined in both databases.
Results: We identified 16,246 patients with ≥1 ALS code in Marketscan (denominator n=85,279,619), yet only 184 were found in the CPRD (denominator n=21,318,589). Using the main algorithm 9,433 ALS patients were included in MarketScan; ALS prevalence was in a range of 4.5 per 100,000 in 2019 (mean age 59 years; 59.5% male) to 6.2 in 2015. During the study period 3,658 (4.3 per 100,000 person-year (PY)) had ≥1 MND code with ALS therapy (proxy cases); large overlapping (98.9%) between “proxy cases” and “case definition specific to ALS” was observed in MarketScan. Most patients in the CPRD did not meet the main algorithm criteria. Of 2,785 with an unspecific MND code, 47.9% (n=1334) had a prescription for riluzole (6.3 per 100,000 PY; 60% male; mean age 65.6 years in 2020).
Conclusions: The expanded algorithm enabled the identification of a large population with ALS, or proxy cases, in MarketScan and CPRD. ALS prevalence in Marketscan and CPRD was similar over the study period and consistent with estimates from the US National ALS Registry.
