Long COVID as a Tracer of Care Fragmentation: Reconstructing Patient Care Trajectories from Belgian Health System Metadata

Background: Health systems increasingly face the challenge of managing conditions character- ized by diagnostic uncertainty, fluctuating symptoms, and prolonged functional impairment. In such contexts, care organization depends not only on individual clinical encounters but also on how patients move across institutions and specialties over time. Long COVID provides a con- temporary case through which to examine how healthcare systems respond to sustained clinical uncertainty in the absence of clearly structured care pathways.
Methods: We conducted an exploratory case-based study in Belgian primary care using routinely generated healthcare metadata combined with clinical records. Three cases were pur- posefully selected from a cohort of more than 400 patients with long COVID followed since 2021. Care trajectories were reconstructed through manual longitudinal review of primary care docu- mentation and hospital-derived metadata accessible through the Belgian eHealth infrastructure. Descriptive analyses examined institutional dispersion, referral patterns, diagnostic utilization, and travel distance across healthcare institutions.
Results: Care trajectories extended over 39–68 months and involved 7–12 healthcare insti- tutions, 10–21 medical specialties, and 25–81 individual providers per case. Patients underwent repeated cross-specialty referrals and extensive diagnostic investigations without evidence of a stable or coordinated care pathway. Routinely generated metadata enabled the reconstruction and visualization of these trajectories, revealing patterns of institutional dispersion, repeated investigations, and fragmented clinical responsibility across levels of care.
Discussion: In a decentralized Bismarckian health system lacking formal integration mech- anisms for long COVID, prolonged diagnostic uncertainty appears to be managed through suc- cessive and dispersed care episodes. Reconstruction of trajectories from routinely generated healthcare metadata provides a scalable method for making such organizational patterns visible and for examining how health systems respond to emerging chronic conditions.