Validity of type 2 diabetes diagnosis in a population-based electronic health record database

Abstract

BackgroundThe increasing burden of type 2 diabetes mellitus makes the continuous surveillance of its prevalence and incidence advisable. Electronic health records (EHRs) have great potential for research and surveillance purposes; however the quality of their data must first be evaluated for fitness for use. The aim of this study was to assess the validity of type 2 diabetes diagnosis in a primary care EHR database covering more than half a million inhabitants, 97% of the population in Navarra, Spain.MethodsIn the Navarra EPIC-InterAct study, the validity of the T90 code from the International Classification of Primary Care, Second Edition was studied in a primary care EHR database to identify incident cases of type 2 diabetes, using a multi-source approach as the gold standard. The sensitivity, specificity, positive predictive value, negative predictive value and the kappa index were calculated. Additionally, type 2 diabetes prevalence from the EHR database was compared with estimations from a health survey.ResultsThe sensitivity, specificity, positive predictive value and negative predictive value of incident type 2 diabetes recorded in the EHRs were 98.2, 99.3, 92.2 and 99.8%, respectively, and the kappa index was 0.946. Overall prevalence of type 2 diabetes diagnosed in the EHRs among adults (35–84 years of age) was 7.2% (95% confidence interval [CI] 7.2–7.3) in men and 5.9% (95% CI 5.8–5.9) in women, which was similar to the prevalence estimated from the health survey: 8.5% (95% CI 7.1–9.8) and 5.5% (95% CI 4.4–6.6) in men and women, respectively.ConclusionsThe high sensitivity and specificity of type 2 diabetes diagnosis found in the primary care EHRs make this database a good source for population-based surveillance of incident and prevalent type 2 diabetes, as well as for monitoring quality of care and health outcomes in diabetic patients.

Cross-national health care database utilization between Spain and France: results from the EPICHRONIC study assessing the prevalence of type 2 diabetes mellitus

“…The validity of the information contained in these data sets has recently begun to be assessed. Results suggest that they are valid to assess the prevalence of cardiovascular risk factors, such as type 2 diabetes mellitus, 18 , 21 although more studies are required to assess validity to conduct other types of population-based surveillance studies. Another database with outpatient information is the outpatient drug-dispensing database, with information on drugs dispensed at retail pharmacies.…”

“…This procedure is similar to that developed by Roberto et al 7 for databases from a primary care setting and is similar to that used by Vinagre et al 29 in the Primary Care Electronic Medical System of Catalonia. The codes used in our study to identify type 2 diabetes mellitus have been validated in the Navarre database, 18 and also in other regions, such as Madrid. 30 In the Basque Country, the quality of the codification of diagnoses in the electronic health care records for primary care is also high.…”

“…A recent study on type 2 diabetes mellitus ICPC-2 codes in Navarre showed a sensitivity of 98%, a specificity of 99%, and a positive predictive value of 92%. 18 …”

Comparative Retinopathy Risk in People with Type 2 Diabetes Treated with Post Metformin Second-line Incretin Therapies: Study Based on US Electronic Medical Records

“…All individuals with a diagnosis of T2DM (excluding type 1 and gestational diabetes) were included in this study with the conditions of (1) no missing data for age and sex; (2) age ≥ 18 and <80 years at diagnosis of T2DM; (3) initiated therapy with metformin, and (4) received a second-line ADD for at least 3 months from 2005 to 2016. The clinically driven machine-learning-based algorithms to identify patients with T2DM from EMRs have been described previously [22,23]. The second line ADDs were DPP-4i, GLP-1RA, INS, TZD, or SU.…”

Evaluation of the obesity paradox in diabetes: a longitudinal case control study

“…The potential for misdiagnosis, misclassification, and miscoding of diagnostic codes in EMRs such as the THIN database cannot be understated [130][131][132][133] . However, according to a number of studies, based on extensive data mining and quality assessments, most of the diagnosis codes in the UK primary care database are well recorded [134][135][136] . The Read code system allows easy recording of clinical information on a computer by the GP, without advance knowledge of coding and classification.…”