Finalization and Validation of Questionnaire and Algorithm of SPUR, a New Adherence Profiling Tool

Elodie de Bock,1 Kevin Dolgin,2 Léa Kombargi,2 Benoit Arnould,1 Tanguy Vilcot,1 Guillaume Hubert,2 Marie-Eve Laporte,3 Lydiane Nabec,4 Gérard Reach5 1Patient Centred Outcomes, ICON plc, Lyon, France; 2Observia, Paris, France; 3IAE Paris - Sorbonne Business School, Université Paris 1 Panthéon-Sorbonne, Paris, France; 4Université Paris-Saclay, RITM (Réseaux, Innovation, Territoire et Mondialisation), Paris, France; 5Health Education and Practices Laboratory (LEPS), Sorbonne Paris-Nord University, Bobigny, FranceCorrespondence: Elodie de Bock, ICON plc, 27 rue de la Villette, Lyon, 69003, France, Tel + 33 472 13 59 81, Email Elodie.deBock@iconplc.comPurpose: The SPUR (Social, Psychological, Usage and Rational) Adherence Profiling Tool is a recently developed adaptive instrument for assessing key patient-level drivers for non-adherence. This study describes the SPUR questionnaire’s finalization and psychometric evaluation.Patients and Methods: Data were collected through an online survey among patients with type 2 diabetes included by general practitioners and diabetologists in France. The survey included four questionnaires, SPUR and three validated adherence measures: BMQ, MARS and ACCEPT. Item-level analysis and a partial credit model (PCM) were performed to refine the response option coding of SPUR items. The final item selection of SPUR was defined using a PCM and a principal component analysis (PCA). Construct validity, concurrent validity and known-groups validity were assessed on the final SPUR questionnaire.Results: A total of 245 patients (55% men, mean age of 63 years) completed the survey remotely and were included in this analysis. Refining response option coding allowed a better discrimination of patients on the latent trait. After item selection, a short, an intermediate, and a long form composed the final SPUR questionnaire. The short form will be used to screen patients for risk and then the other forms will allow the collection of further information to refine the risk assessment and decide the best levers for action. Results obtained were supportive of the construct validity of the forms. Their concurrent validity was demonstrated: moderate to high significant correlations were obtained with BMQ, MARS and ACCEPT scores. Their known-groups validity were shown with a logical pattern of higher scores obtained for patients considered non-adherent and significant differences between the scores obtained for patients considered adherent versus non-adherent.Conclusion: SPUR is a valid tool to evaluate the risk of non-adherence of patients, allowing effective intervention by providing insights into the respective individual reasons for lack of adherence.Keywords: digital questionnaire, non-adherence drivers, psychometric validation, Rasch modelling