Patients presenting to a tertiary care GI clinic report poorer sleep than healthy controls. In general, patients with IBS report the highest rates of sleep difficulties compared to patients with other diagnoses.

No abstract

No abstract

Big data analysis of asthma provides a powerful tool to understand this highly morbid disease but depends on an accurate identification of a computable asthma phenotype in the electronic health record (EHR). METHODS: All children 5-17 years old living in Durham County and seen at Duke University Health System from 2007-2017 were included in an asthma cohort if they had an asthma medication prescription and either (1) asthma on the problem list, (2) one inpatient asthma exacerbation, or (3) two outpatient asthma encounters. Among these, 124 were randomly selected for review. Each chart was reviewed by two physicians evaluating clinical encounters, exams, pulmonary function tests, co-morbid conditions, and studies to assess for the presence or absence of asthma. Discrepancies were reviewed by a third specialty-trained physician. The positive predictive value of physician confirmed asthma was calculated. Inter-observer reliability was measured by Fleiss' kappa and differences in agreement between groups was tested by McNemar's test. Analyses were stratified by age, reviewer training level, and clinical characteristics. RESULTS: The positive predictive value of a computable phenotype including provider diagnosis and asthma medication was 97%. There was 60% agreement among reviewers on the diagnosis of asthma with a kappa of-0.08, indicating poor agreement. CONCLUSIONS: Asthma cases identified within an EHR using problem list or encounter based asthma diagnosis with accompanying asthma drug prescription has high concordance with expert review. Inter-rater reliability of pediatric asthma diagnosis remains a challenge.

No abstract