Cohort design and natural language processing to reduce bias in electronic health records research.
Shaan KhurshidChristopher ReederLia X HarringtonPulkit SinghGopal SarmaSamuel F FriedmanPaolo Di AchilleNathaniel DiamantJonathan W CunninghamAshby C TurnerEmily S LauJulian S HaimovichMostafa A Al-AlusiXin WangMarcus D R KlarqvistJeffrey M AshburnerChristian DiedrichMercedeh GhadessiJohanna MielkeHanna M EilkenAlice McElhinneyAndrea DerixSteven J AtlasPatrick T EllinorAnthony A PhilippakisChristopher D AndersonJennifer E HoPuneet BatraSteven A LubitzPublished in: NPJ digital medicine (2022)
Electronic health record (EHR) datasets are statistically powerful but are subject to ascertainment bias and missingness. Using the Mass General Brigham multi-institutional EHR, we approximated a community-based cohort by sampling patients receiving longitudinal primary care between 2001-2018 (Community Care Cohort Project [C3PO], n = 520,868). We utilized natural language processing (NLP) to recover vital signs from unstructured notes. We assessed the validity of C3PO by deploying established risk models for myocardial infarction/stroke and atrial fibrillation. We then compared C3PO to Convenience Samples including all individuals from the same EHR with complete data, but without a longitudinal primary care requirement. NLP reduced the missingness of vital signs by 31%. NLP-recovered vital signs were highly correlated with values derived from structured fields (Pearson r range 0.95-0.99). Atrial fibrillation and myocardial infarction/stroke incidence were lower and risk models were better calibrated in C3PO as opposed to the Convenience Samples (calibration error range for myocardial infarction/stroke: 0.012-0.030 in C3PO vs. 0.028-0.046 in Convenience Samples; calibration error for atrial fibrillation 0.028 in C3PO vs. 0.036 in Convenience Samples). Sampling patients receiving regular primary care and using NLP to recover missing data may reduce bias and maximize generalizability of EHR research.
Keyphrases
- electronic health record
- atrial fibrillation
- primary care
- heart failure
- clinical decision support
- oral anticoagulants
- catheter ablation
- left atrial
- left atrial appendage
- visible light
- direct oral anticoagulants
- healthcare
- adverse drug
- left ventricular
- percutaneous coronary intervention
- general practice
- palliative care
- mental health
- machine learning
- pain management
- cross sectional
- mitral valve
- single cell