Automatically identifying opioid use disorder in non-cancer patients on chronic opioid therapy.

Background: Using the International Classification of Diseases (ICD) codes alone to record opioid use disorder (OUD) may not completely document OUD in the electronic health record (EHR). We developed and evaluated natural language processing (NLP) approaches to identify OUD from the clinal note. We explored the concordance between ICD-coded and NLP-identified OUD.Methods: We studied EHRs from 13,654 (female: 8223; male: 5431) adult non-cancer patients who received chronic opioid therapy (COT) and had at least one clinical note between 2013 and 2018. Of eligible patients, we randomly selected 10,218 (75%) patients as the training set and the remaining 3436 patients (25%) as the test dataset for NLP approaches.Results: We generated 539 terms representing OUD mentions in clinical notes (e.g., "opioid use disorder," "opioid abuse," "opioid dependence," "opioid overdose") and 73 terms representing OUD medication treatments. By domain expert manual review for the test dataset, our NLP approach yielded high performance: 98.5% for precision, 100% for recall, and 99.2% for F-measure. The concordance of these NLP and ICD identified OUD was modest (Kappa = 0.63).Conclusions: Our NLP approach can accurately identify OUD patients from clinical notes. The combined use of ICD diagnostic code and NLP approach can improve OUD identification.

Validity of ICD9-CM codes to diagnose chronic obstructive pulmonary disease from National Health Insurance claim data in Taiwan.

PURPOSE: Claim data from Taiwan's National Health Insurance (NHI) database have previously been utilized in the study of COPD. However, there are limited data on the positive predictive value of claim data for COPD diagnosis. Therefore, this study aimed to characterize and validate the COPD cohort identified from the NHI research database. METHODS: This cross-sectional study compared records from claim data with those from a medical center. From 2007 to 2014, a COPD cohort was constructed from claim data using ICD9-CM codes for COPD. The diagnostic positive predictive value of these data was assessed with reference to physician-verified COPD. In addition, a multivariate logistic regression model was built to identify independent factors associated with the positive predictive value of COPD diagnosis by claim data. RESULTS: During the 8-year study period, a total of 12,127 subjects met the criterion of having two or more outpatient codes in 1 year or one or more inpatient COPD codes in their claim data. Of this total, the diagnosis of COPD was verified by physicians in 7,701 (63.5%) subjects. Applying a more stringent criterion - three or more outpatient codes or two or more inpatient codes - improved the diagnostic positive predictive value to 72.2%. Age ≥65 years and a claim for spirometry were the two most important factors associated with the positive predictive value of claim-data-defined COPD. Adding spirometry testing to diagnostic ICD9-CM codes for COPD increased the positive predictive value to 84.6%. CONCLUSION: This study emphasizes the importance of validation of disease-specific diagnosis prior to applying an administrative database in clinical studies. It also indicates the limitation of ICD9-CM codes alone in recognizing COPD patients within the NHI research database.