Development and validation of algorithms to classify type 1 and 2 diabetes according to age at diagnosis using electronic health records

• Published in: BMC medical research methodology 2020-02, Vol.20 (1), p.35-35, Article 35
• Main Authors: Ke, Calvin, Stukel, Thérèse A, Luk, Andrea, Shah, Baiju R, Jha, Prabhat, Lau, Eric, Ma, Ronald C W, So, Wing-Yee, Kong, Alice P, Chow, Elaine, Chan, Juliana C N
• Format: Article
• Language: English
• Subjects: Administrative data; Adolescent; Adult; Age; Aged; Algorithms; Asian People - statistics & numerical data; Child; Children; Chinese ethnicity; Classification; Codes; Cohort Studies; Databases, Factual - statistics & numerical data; Diabetes; Diabetes Mellitus, Type 1 - classification; Diabetes Mellitus, Type 1 - diagnosis; Diabetes Mellitus, Type 1 - ethnology; Diabetes Mellitus, Type 2 - classification; Diabetes Mellitus, Type 2 - diagnosis; Diabetes Mellitus, Type 2 - ethnology; Diabetics; Diagnosis; Disease age factors; Electronic health records; Electronic Health Records - statistics & numerical data; Electronic records; Female; Glucose; Hong Kong; Hospitals; Humans; Hyperglycemia; Identification and classification; Insulin; Insulin resistance; Laboratories; Male; Medical diagnosis; Medical records; Medical research; Methods; Middle Aged; Pediatrics; Population; Population-based studies; Prescriptions; Reproducibility of Results; Type 1 diabetes; Type 2 diabetes; Validation study; Young Adult
• ISSN: 1471-2288; 1471-2288
• DOI: 10.1186/s12874-020-00921-3

Validated algorithms to classify type 1 and 2 diabetes (T1D, T2D) are mostly limited to white pediatric populations. We conducted a large study in Hong Kong among children and adults with diabetes to develop and validate algorithms using electronic health records (EHRs) to classify diabetes type against clinical assessment as the reference standard, and to evaluate performance by age at diagnosis. We included all people with diabetes (age at diagnosis 1.5-100 years during 2002-15) in the Hong Kong Diabetes Register and randomized them to derivation and validation cohorts. We developed candidate algorithms to identify diabetes types using encounter codes, prescriptions, and combinations of these criteria ("combination algorithms"). We identified 3 algorithms with the highest sensitivity, positive predictive value (PPV), and kappa coefficient, and evaluated performance by age at diagnosis in the validation cohort. There were 10,196 (T1D n = 60, T2D n = 10,136) and 5101 (T1D n = 43, T2D n = 5058) people in the derivation and validation cohorts (mean age at diagnosis 22.7, 55.9 years; 53.3, 43.9% female; for T1D and T2D respectively). Algorithms using codes or prescriptions classified T1D well for age at diagnosis