Autor(es):
Georgiadis, Stylianos ; Di Giuseppe, Daniela ; Scherer, Almut ; Hetland, Merete Lund ; Jones, Gareth ; Glintborg, Bente ; Loft, Anne Gitte ; Wallman, Johan K ; Michelsen, Brigitte ; Kristianslund, Eirik Klami ; Yazici, Ayten ; Birlik, Merih ; Závada, Jakub ; Nissen, Eva Rames ; Ciurea, Adrian ; Gudbjornsson, Bjorn ; Palsson, Olafur ; Rotar, Ziga ; Tomšič, Matija ; Relas, Heikki ; Huhtakangas, Johanna ; Maria Rodrigues, Ana ; Santos, Maria Jose ; Castrejon, Isabel ; Díaz-González, Federico ; van de Sande, Marleen ; Hellamand, Pasoon ; Ørnbjerg, Lykke
Data: 2025
Identificador Persistente: http://hdl.handle.net/10362/185193
Origem: Repositório Institucional da UNL
Assunto(s): Missing data; Patient reported outcomes; Registry data; Single imputation; Spondyloarthritis; Rheumatology
Descrição
Funding Information: Open access funding provided by Copenhagen University. Open access funding provided by Copenhagen University. Open access funding provided by Copenhagen University. This work was supported by Novartis Pharma AG. Novartis had no influence on the data collection, statistical analyses, manuscript preparation or decision to submit the manuscript. Funding Information: The EuroSpA collaboration has been supported by Novartis Pharma AG since 2017 and UCB Biopharma SRL since 2022. This EuroSpA study was financially supported by UCB. No financial sponsors had any influence on the data collection, statistical analyses, manuscript preparation, or decision to submit. Publisher Copyright: © The Author(s) 2025.
Background: To investigate a patient-level single imputation approach for patient reported outcomes (PROs) that express similar contents or associated PROs, where a PRO whose value is missing at a particular timepoint is substituted by another PRO whose value is available at the same timepoint. Methods: We performed a simulation study on registry-based spondyloarthritis data to explore the potential interchangeability between the patient pain (PPA) and fatigue (PFA) assessment scores and relevant Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) individual questions, and between PPA, PFA and patient global assessment (PGA). Performance was assessed per imputation method in terms of relative bias and coverage. Sample size, level of missingness and missing data pattern were included as parameters in the simulations. Results: All applied scenarios to interchange PPA with BASDAI question 2 (axial pain), BASDAI question 3 (peripheral joint pain/swelling) or their average failed. Interchangeability between PFA and BASDAI question 1 (fatigue/tiredness) was acceptable for partially (up to 50%) missing data. When interchanging patient assessment scores (PPA, PFA and PGA), we observed inconsistent results in terms of performance. The performance of the applied methods depended on the sample size and the level of missingness, but not heavily on the underlying missing data pattern. Conclusions: Interchanging PFA and the BASDAI fatigue question was justified for partially missing data, while interchangeability between PPA, PFA and PGA, and between PPA and the BASDAI pain questions was not advised. Our findings suggest that registering patient assessment scores and BASDAI questions is recommended.
