Dataset Information

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

ABSTRACT: BACKGROUND: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. METHODS: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. RESULTS: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

SUBMITTER: Ke C

PROVIDER: S-EPMC7038546 | biostudies-literature | 2020 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

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

Ke Calvin C Stukel Thérèse A TA Luk Andrea A Shah Baiju R BR Jha Prabhat P Lau Eric E Ma Ronald C W RCW So Wing-Yee WY Kong Alice P AP Chow Elaine E Chan Juliana C N JCN

BMC medical research methodology 20200224 1

<h4>Background</h4>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.<h4>Methods</h4>We included all people with diabetes (age at diagnosis 1.5-100 ye ...[more]

PMID: 32093635

Similar Datasets

Project description:AIMS:To estimate the overall and cause-specific mortality in a population of African-Americans and white Americans with a low socio-economic status who had young-onset insulin-treated diabetes and had survived beyond the age of 40 years, and to examine whether any excess risk varied according to age at diabetes onset. METHODS:Using the Southern Community Cohort Study, we conducted a mortality follow-up of a cohort of mostly low-income participants aged 40-79 years (mean 50 years) at cohort entry with insulin-treated diabetes diagnosed before age 30 years (n=475) and without diabetes (n=62 266). Childhood onset was defined as diabetes diagnosed before age 20 years (n=162), while young-adulthood onset was defined as diabetes diagnosed between ages 20 and 29 years (n=313). Cause-specific mortality was based on both underlying and contributing causes of death, obtained from death certificates. Multivariable Cox analysis was performed. RESULTS:During follow-up (mean 9.5 years), 38.7% of those with and 12.9% of those without diabetes died. Compared with those without diabetes, increases in mortality rate were generally similar among those with childhood- and young-adulthood-onset diabetes for deaths from: all causes (childhood: hazard ratio 4.3, CI 3.3-5.5; young adulthood: hazard ratio 4.9, CI 4.0-5.8); ischaemic heart disease (childhood: hazard ratio 5.7, CI 3.5-9.4; young adulthood: hazard ratio 7.9, CI 5.6-11.0); heart failure (childhood: hazard ratio 7.3, CI 4.2-12.7; young adulthood: hazard ratio 5.4, CI 3.3-8.9); sepsis (childhood: hazard ratio 10.3, CI 6.1-17.3; young adulthood: hazard ratio 8.8, CI 5.7-13.5); renal failure (childhood: hazard ratio 15.1, CI 8.6-26.5; young adulthood: hazard ratio 18.2, CI 12.3-27.1); respiratory disorders (childhood: hazard ratio 3.9, CI 2.3-6.7; young adulthood: hazard ratio 5.3, CI 3.7-7.7); suicide/homicide/accidents (childhood: hazard ratio 2.3, CI 0.72-7.0; young adulthood: hazard ratio 5.8, CI 3.4-10.2); and cancer (childhood: hazard ratio 2.1, CI 0.98-4.4; young adulthood: hazard ratio 1.2, CI 0.55-2.5). CONCLUSIONS:We observed high excess long-term mortality for all-cause, renal failure, ischemic heart disease and heart failure mortality in African-American and white American people with early-onset insulin-treated diabetes.

Project description:ObjectiveTo discover diverse genotype-phenotype associations affiliated with Type 2 Diabetes Mellitus (T2DM) via genome-wide association study (GWAS) and phenome-wide association study (PheWAS), more cases (T2DM subjects) and controls (subjects without T2DM) are required to be identified (e.g., via Electronic Health Records (EHR)). However, existing expert based identification algorithms often suffer in a low recall rate and could miss a large number of valuable samples under conservative filtering standards. The goal of this work is to develop a semi-automated framework based on machine learning as a pilot study to liberalize filtering criteria to improve recall rate with a keeping of low false positive rate.Materials and methodsWe propose a data informed framework for identifying subjects with and without T2DM from EHR via feature engineering and machine learning. We evaluate and contrast the identification performance of widely-used machine learning models within our framework, including k-Nearest-Neighbors, Naïve Bayes, Decision Tree, Random Forest, Support Vector Machine and Logistic Regression. Our framework was conducted on 300 patient samples (161 cases, 60 controls and 79 unconfirmed subjects), randomly selected from 23,281 diabetes related cohort retrieved from a regional distributed EHR repository ranging from 2012 to 2014.ResultsWe apply top-performing machine learning algorithms on the engineered features. We benchmark and contrast the accuracy, precision, AUC, sensitivity and specificity of classification models against the state-of-the-art expert algorithm for identification of T2DM subjects. Our results indicate that the framework achieved high identification performances (∼0.98 in average AUC), which are much higher than the state-of-the-art algorithm (0.71 in AUC).DiscussionExpert algorithm-based identification of T2DM subjects from EHR is often hampered by the high missing rates due to their conservative selection criteria. Our framework leverages machine learning and feature engineering to loosen such selection criteria to achieve a high identification rate of cases and controls.ConclusionsOur proposed framework demonstrates a more accurate and efficient approach for identifying subjects with and without T2DM from EHR.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets