ELECTRONIC MEDICAL RECORD AGGREGATION VERSUS CLINICIAN DATA ENTRY: COMPARISON OF ENROLLMENT AND DATA COMPLETENESS FROM A REAL-WORLD MYASTHENIA GRAVIS REGISTRY
Author(s)
Fatemeh Amini, MScR1, Emma Bagshaw, BA1, Alasdair Fellows, MSc1, Carmen Petitjean, MPhil1, Kelly Graham Gwathmey, MD2, Mark Larkin, PhD1.
1Vitaccess, London, United Kingdom, 2Department of Neurology, Columbia University, New York, NY, USA.
1Vitaccess, London, United Kingdom, 2Department of Neurology, Columbia University, New York, NY, USA.
OBJECTIVES: Capture of clinical data using electronic medical record (EMR) aggregation enables “direct-to-patient” recruitment in real-world research, without clinician involvement. However, data quality and completeness may differ relative to clinician-entered (eCRF) data. We compared EMR and eCRF data capture in an international registry that uses both approaches.
METHODS: Between 2 August 2024 - 6 January 2026, EMR and eCRF data capture methodologies were compared by assessing enrolment and data completeness among participants in the Vitaccess Real MG myasthenia gravis (MG) registry.
RESULTS: By 06 January 2026, 140 participants were registered. Enrolment was quicker with EMR vs eCRF (time to first-patient-in 2 versus 47 days), higher (84 versus 56 participants), and geographic distribution was broader (25 versus 5 US states). However, owing to EMR limitations at the time of registry setup, this approach was US only. Proportion of participants with ≥1 recorded MG treatment: eCRF 83.9%; aggregator 86.9%. Proportion with ≥1 routine treatment: eCRF 83.9%; aggregator 85.7%. Proportion with ≥1 rescue treatment: eCRF 12.5%; aggregator 69.1%. As routine and rescue treatments were not differentiated in EMR data, assumptions had to be made to assign treatment categories. EMR aggregation enabled data extraction at a frequent, researcher-defined schedule, minimizing data accrual lags, whereas eCRF schedule was restricted by clinician workload considerations. EMR data capture relied on successful connection of medical records. Older records were sometimes unavailable, limiting visibility of early-life data, such as MG diagnosis date. While clinician burden was greater for eCRF, greater analytic effort was required to curate EMR data, and quality was limited by assumptions and missing data. Additional information could be inputted by clinicians into the eCRF where necessary, but EMRs contained a breadth of extra data that could be used for future research.
CONCLUSIONS: Incorporation of EMR aggregator-derived data in registries may be a valuable alternative or complement to clinician-entered eCRF data.
METHODS: Between 2 August 2024 - 6 January 2026, EMR and eCRF data capture methodologies were compared by assessing enrolment and data completeness among participants in the Vitaccess Real MG myasthenia gravis (MG) registry.
RESULTS: By 06 January 2026, 140 participants were registered. Enrolment was quicker with EMR vs eCRF (time to first-patient-in 2 versus 47 days), higher (84 versus 56 participants), and geographic distribution was broader (25 versus 5 US states). However, owing to EMR limitations at the time of registry setup, this approach was US only. Proportion of participants with ≥1 recorded MG treatment: eCRF 83.9%; aggregator 86.9%. Proportion with ≥1 routine treatment: eCRF 83.9%; aggregator 85.7%. Proportion with ≥1 rescue treatment: eCRF 12.5%; aggregator 69.1%. As routine and rescue treatments were not differentiated in EMR data, assumptions had to be made to assign treatment categories. EMR aggregation enabled data extraction at a frequent, researcher-defined schedule, minimizing data accrual lags, whereas eCRF schedule was restricted by clinician workload considerations. EMR data capture relied on successful connection of medical records. Older records were sometimes unavailable, limiting visibility of early-life data, such as MG diagnosis date. While clinician burden was greater for eCRF, greater analytic effort was required to curate EMR data, and quality was limited by assumptions and missing data. Additional information could be inputted by clinicians into the eCRF where necessary, but EMRs contained a breadth of extra data that could be used for future research.
CONCLUSIONS: Incorporation of EMR aggregator-derived data in registries may be a valuable alternative or complement to clinician-entered eCRF data.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
CO112
Topic
Clinical Outcomes, Methodological & Statistical Research, Study Approaches
Topic Subcategory
Clinician Reported Outcomes
Disease
Neurological Disorders