Development of a natural language processing algorithm to detect chronic cough in electronic health records

Abstract

Background Chronic cough (CC) is difficult to identify in electronic health records (EHRs) due to the lack of specific diagnostic codes. We developed a natural language processing (NLP) model to identify cough in free-text provider notes in EHRs from multiple health care providers with the objective of using the model in a rules-based CC algorithm to identify individuals with CC from EHRs and to describe the demographic and clinical characteristics of individuals with CC. Methods This was a retrospective observational study of enrollees in Optum’s Integrated Clinical + Claims Database. Participants were 18–85 years of age with medical and pharmacy health insurance coverage between January 2016 and March 2017. A labeled reference standard data set was constructed by manually annotating 1000 randomly selected provider notes from the EHRs of enrollees with ≥ 1 cough mention. An NLP model was developed to extract positive or negated cough contexts. NLP, cough diagnosis and medications identified cough encounters. Patients with ≥ 3 encounters spanning at least 56 days within 120 days were defined as having CC. Results The positive predictive value and sensitivity of the NLP algorithm were 0.96 and 0.68, respectively, for positive cough contexts, and 0.96 and 0.84, respectively, for negated cough contexts. Among the 4818 individuals identified as having CC, 37% were identified using NLP-identified cough mentions in provider notes alone, 16% by diagnosis codes and/or written medication orders, and 47% through a combination of provider notes and diagnosis codes/medications. Chronic cough patients were, on average, 61.0 years and 67.0% were female. The most prevalent comorbidities were respiratory infections (75%) and other lower respiratory disease (82%). Conclusions Our EHR-based algorithm integrating NLP methodology with structured fields was able to identify a CC population. Machine learning based approaches can therefore aid in patient selection for future CC research studies.

Epidemiology and Healthcare Service Utilization among Adults with Chronic Cough

“…The demographic characteristics of our CC cohort including older age and female predominance, were similar to those reported in previous studies [19,22] and to real-world data analyses [5,[23][24][25]. In line with the results of previous epidemiological studies [2,19], GERD, asthma, and COPD were substantially more common among PwCC compared to people without cough [26].…”

Patterns of Cough Medication Prescribing among Patients with Chronic Cough in Florida: 2012–2021

“…(6) In addition, our study was limited to using structured data fields in the OneFlorida data; thus, information documented in unstructured clinical data (e.g., duration of cough symptoms and severity) was not captured. Although previous studies demonstrated that the use of natural language processing (NLP) in unstructured data significantly improved sensitivity in identifying CC patients in EHR data [25,26], using an NLP-integrated algorithm to identify CC is not widely used because it can be time-consuming and computationally intensive in clinical practice. In sum, the CC prevalence in our analysis is likely to be substantially underestimated, although there is a possibility that we may have misclassified uncontrolled COPD patients as CC patients, based on the fact that CC patients had COPD as the most common pre-index respiratory comorbidity.…”

“…We excluded patients who were (1) aged < 18 years (measured on 30 June of each calendar year), (2) had any malignant cancer diagnoses or any respiratory tumor diagnoses, and (3) had <2 medical encounters in each calendar year. We used an existing algorithm that was used in previous studies to identify patients with CC based on the presence of any 3 clinical cough events occurring within a 120-day period and were separated from each other by at least 21 days (Figure S1) [24][25][26][27]. These cough events included a diagnosis of cough [ICD-9-CM: 786.2 or ICD-10-CM: R05] or a CM prescription order (opioid antitussives, benzonatate, or dextromethorphan-containing products).…”