Presentation (CTI)

OBJECTIVES: Longitudinal registries depend on sustained patient engagement, representative participation, and consistent data capture of patient-reported outcomes (PRO) over prolonged periods to generate reliable real-world evidence (RWE). Reduced engagement, attrition, and missing data can accumulate silently and may be detected too late for effective mitigation, resulting in incomplete or biased evidence. We describe a dashboard-enabled approach for near real-time monitoring for early identification of emerging data quality and bias risks in the VRMG myasthenia gravis registry.
METHODS: Patient-reported data were collected and ingested into a digital registry platform via automated cloud-based daily data pipelines and presented through daily-refreshing interactive dashboards. Pre-specified monitoring indicators included: numbers of active/inactive participants, recruitment pathway split, survey completion rates at each timepoint aggregated by type and patient-level drilldowns, and demographic and PRO score distributions. We summarised monitoring outputs descriptively.
RESULTS: The dashboards provided centralised, daily snapshots of data collection. Application enabled near real-time monitoring of 241 enrolled patients and four PRO measures over 24 months, where PRO completion was ≥ 95% at baseline and ≥ 69% across all timepoints. There were no concerning distribution drifts in PRO scores or imbalance in demographic distributions. Dashboarding enabled early identification of patients with reduced engagement or incomplete onboarding, and unexpected patterns that could indicate data quality concerns or registry population shifts for targeted follow-up before issues became embedded in the registry dataset.
CONCLUSIONS: Dashboard-enabled monitoring operationalises evidence governance. It can strengthen real-world registry oversight by converting routinely collected digital registry data into actionable indicators of engagement and data quality. Unlike traditional approaches that rely on periodic data cuts or retrospective reviews, this approach supports early and proactive management of bias, missing data and anomalies during the ongoing conduct of the registry. These can improve the reliability and completeness of RWE generated from longitudinal registry data.