Referral bias in ALS epidemiological studies

Giancarlo Logroscino (1) , Benoît Marin (2, 3, 1, 4, 5, 6) , Marco Piccininni (1) , Simona Arcuti (1) , Adriano Chiò (7, 8) , Orla Hardiman (9) , James Rooney (9) , Stefano Zoccolella (1) , Philippe Couratier (10, 2, 4) , Pierre-Marie Preux (2, 1, 4) , Ettore Beghi (6)

1 UNIBA - Università degli studi di Bari Aldo Moro = University of Bari Aldo Moro
2 NET - Neuroépidémiologie Tropicale
3 Laboratoire de Biostatistique et d'Informatique Médicale
4 GEIST FR CNRS 3503 - Génomique, Environnement, Immunité, Santé, Thérapeutique
5 Centre d'Epidémiologie de Biostatistique et de Méthodologie de la Recherche
6 IRCCS - Istituto di Ricerche Farmacologiche "Mario Negri" [Milan, Italy]
7 UNITO - Università degli studi di Torino = University of Turin
8 ICST-CNR - Institute of Cognitive Sciences and Technologies
9 Trinity College Dublin
10 Service de Neurologie [CHU Limoges]

Despite concerns about the representativeness of patients from ALS tertiary centers as compared to the ALS general population, the extent of referral bias in clinical studies remains largely unknown. Using data from EURALS consortium we aimed to assess nature, extent and impact of referral bias. METHODS: Four European ALS population-based registries located in Ireland, Piedmont, Puglia, Italy, and Limousin, France, covering 50 million person-years, participated. Demographic and clinic characteristics of ALS patients diagnosed in tertiary referral centers were contrasted with the whole ALS populations enrolled in registries in the same geographical areas. RESULTS: Patients referred to ALS centers were younger (with difference ranging from 1.1 years to 2.4 years), less likely to present a bulbar onset, with a higher proportion of familial antecedents and a longer survival (ranging from 11% to 15%) when compared to the entire ALS population in the same geographic area. CONCLUSIONS: A trend for referral bias is present in cohorts drawn from ALS referral centers. The magnitude of the possible referral bias in a particular tertiary center can be estimated through a comparison with ALS patients drawn from registry in the same geographic area. Studies based on clinical cohorts should be cautiously interpreted. The presence of a registry in the same area may improve the complete ascertainment in the referral center.

Domaines: Santé publique et épidémiologie

Giancarlo Logroscino, Benoît Marin, Marco Piccininni, Simona Arcuti, Adriano Chiò, et al.. Referral bias in ALS epidemiological studies. PLoS ONE, 2018, 13 (4), pp.e0195821. ⟨10.1371/journal.pone.0195821⟩. ⟨hal-01773975⟩