358 - Demographic, clinical and research environment factors associated with research loss to follow-up completion in a pan-Canadian pediatric emergency department study.

Poster Number: 358
Publication Number: 358.95

Background: Loss to follow-up (LTFU) is an important problem in research, affecting the internal and external validity of research findings. The socio-demographic and clinical factors that influence LTFU in pediatric or emergency research settings are not well established.

Objective: To determine the associations between socio-demographic, clinical and research environment factors, and LTFU at 14-days and 12-months among children enrolled in a pan-Canadian emergency department (ED) research study.

Design/Methods: This was a secondary analysis of a prospective cohort study enrolling children tested for SARS-CoV-2 in 14 Canadian pediatric EDs between August 2020 and February 2022. Socio-demographic (participant age, sex, and self-reported race), clinical (covid test positivity, hospitalisation) and research environment (recruitment site and recruitment quartile) characteristics were assessed at the index visit. The primary outcome was LTFU, after at least one contact attempt, 14-days after index visit. The secondary outcome was LTFU at 12-months. Due to the evolving nature of the pandemic, the 12-months follow-up was incorporated in the protocol later in the study, after the follow-up window had closed for some participants. Research teams were instructed to contact participants by phone or email (per preference), up to 5 attempts. We evaluated the association between factors and LTFU using a multivariable logistic regression model.

Results: 6869 children were enrolled in the parent study. All participants were eligible for follow-up at 14-days, and 2829 (41.2%) were eligible for the 12-months follow-up. LTFU rates at 14-days and 12-months were 12.3% (n=848) and 8.0% (n= 235), respectively. Age, sex, self-reported race, test positivity and hospitalizations were not independently associated with LTFU at either time point. At 14-days, LTFU differed between recruitment sites, with adjusted odds ratios (aOR) ranging from 0.3 (95%CI 0.2-0.5) to 15.4 (95%CI 9.7-24.4). LTFU also increased with each recruitment quartile (aOR 1.27 [95%CI 0.9-1.8], aOR 1.66 [95% CI 1.2-2.4] and aOR 2.44 [95%CI 1.7-3.5] for Q2, Q3 and Q4 respectively). Similar findings were obtained for the 12-months follow-up.

Conclusion(s): Socio-demographic and clinical characteristics were not independently associated with LTFU in this pan-Canadian pediatric ED study. Recruitment site and timing of recruitment impacted LTFU, suggesting a need for better standardization of follow-up procedures across ED sites and strategies to mitigate study or follow-up “fatigue” among study teams.