Loss to follow-up in registries of rheumatic patients treated with biologics: a potential information bias in assessing pharmacovigilance and efficacy outcomes

• Main Authors: Ana Valido, Joana Silva-Dinis, Ana Rita Cruz-Machado, Maria João Gonçalves, Vasco C Romão, Maria João Saavedra, João Eurico Fonseca
• Format: Article
• Language: English
• Published: Sociedade Portuguesa de Reumatologia 2019-10-01
• Series: Acta Reumatológica Portuguesa
• Subjects: biologic therapy
• Online Access: http://www.actareumatologica.pt/files/article/1278_loss_to_follow_up_in_registries_of_rheumatic_patients_treated_with_biologics_a_potential_information_bias_in_assessing_pharmacovigilance_and_efficacy_outcomes_file.pdf

Background: The information associated with loss to follow-up (LFU) patients may affect real-world data evaluation of the use of biologics that is not being adequately captured in registries. Methods: We identified all patients (Pts) treated with biologics in our center who had no visits registered for more than 6 months, in the Rheumatic Diseases Portuguese Register, Reuma.pt. We retrieved baseline information from Reuma.pt and from the hospital electronic clinical record. We then performed a telephonic interview to characterize the reasons for LFU at our day care unit. For Pts unable to be contacted by telephone a letter of invitation to an appointment at the hospital was sent. Results: From a total of 794 Pts registered in Reuma.pt at our center with active biologic therapy 227 did not have any information registered in the last 6 months. Of this, 36 Pts were on biologic therapy prescribed by other departments and maintained follow-up in these departments. 102 Pts had suspended biologic administration by medical indication and this information was registered in the hospital electronic clinical records but not updated in Reuma.pt. For 89 Pts no information could be retrieved from either the hospital electronic clinical record or Reuma.pt and we classified these Pts as true LFU. 26 of these LFU Pts were being followed up in another Rheumatology center. 26 of the LFU Pts died. 11 Pts had an adverse effect. 4 Pts of the LFU were considering to be in remission. We were not able to contact 15 of the LFU pts. Conclusion: Identifying LFU Pts and clarifying the reason for the loss of data in a register contributes to a better knowledge on strategies to discontinue biologics in stable pts, to a better pharmacovigilance of adverse effects and to more efficiency in data capture by registries. Due to data protection reasons it was impossible to have access to the Pts's death certificates