Kristian I Macdonald1, Shaun J Kilty1, Carl van Walraven2,3. 1. Department of Otolaryngology-Head and Neck Surgery, University of Ottawa, Ottawa, Ontario, Canada. 2. Department of Medicine, University of Ottawa, Ottawa, Ontario, Canada. 3. Ottawa Hospital Research Institute, Ottawa, Ontario, Canada.
Abstract
OBJECTIVES/HYPOTHESIS: Much of the epidemiological data on chronic rhinosinusitis (CRS) are based on large administrative databases and health surveys. The accuracy of CRS identification with these methods is unknown. METHODS: A systematic review was performed to identify studies that measured the accuracy of CRS diagnoses in large administrative databases or within health surveys. The Quality Assessment of Diagnostic Accuracy Studies 2 tool was used to assess study quality. RESULTS: Of 512 abstracts initially identified, 122 were selected for full-text review; only three studies (2.5%) measured the accuracy of CRS patient identification. In a single, large administrative database study with a CRS prevalence of 54.8%, a single International Classification of Diseases-9th Revision diagnostic code for CRS had a positive predictive value (PPV) of only 34%. A diagnostic code algorithm identified CRS patients with a PPV of 91.3% (95% confidence interval [CI], 85.3-95.1); in a population with a CRS prevalence of 5%, this algorithm had a PPV of 31%. In health survey studies having an estimated CRS prevalence of 25% to 46%, self-reported symptom-based CRS diagnosis had a PPV of 62% (95% CI, 50.2-72.1) when nasal endoscopy was the gold standard for CRS diagnosis, and 70% (95% CI, 57.4-80.8) when otolaryngologist-based CRS diagnosis (after interview and nasal endoscopy) was the gold standard. CONCLUSION: Most health administrative data and health surveys examining CRS did not consider the accuracy of case identification. For unselected populations, administrative data and health surveys using self-reported diagnoses inaccurately identify patients with CRS. Epidemiological results based on such data should be interpreted with these results in mind. Laryngoscope, 126:1303-1310, 2016.
OBJECTIVES/HYPOTHESIS: Much of the epidemiological data on chronic rhinosinusitis (CRS) are based on large administrative databases and health surveys. The accuracy of CRS identification with these methods is unknown. METHODS: A systematic review was performed to identify studies that measured the accuracy of CRS diagnoses in large administrative databases or within health surveys. The Quality Assessment of Diagnostic Accuracy Studies 2 tool was used to assess study quality. RESULTS: Of 512 abstracts initially identified, 122 were selected for full-text review; only three studies (2.5%) measured the accuracy of CRSpatient identification. In a single, large administrative database study with a CRS prevalence of 54.8%, a single International Classification of Diseases-9th Revision diagnostic code for CRS had a positive predictive value (PPV) of only 34%. A diagnostic code algorithm identified CRSpatients with a PPV of 91.3% (95% confidence interval [CI], 85.3-95.1); in a population with a CRS prevalence of 5%, this algorithm had a PPV of 31%. In health survey studies having an estimated CRS prevalence of 25% to 46%, self-reported symptom-based CRS diagnosis had a PPV of 62% (95% CI, 50.2-72.1) when nasal endoscopy was the gold standard for CRS diagnosis, and 70% (95% CI, 57.4-80.8) when otolaryngologist-based CRS diagnosis (after interview and nasal endoscopy) was the gold standard. CONCLUSION: Most health administrative data and health surveys examining CRS did not consider the accuracy of case identification. For unselected populations, administrative data and health surveys using self-reported diagnoses inaccurately identify patients with CRS. Epidemiological results based on such data should be interpreted with these results in mind. Laryngoscope, 126:1303-1310, 2016.
Authors: Nikita Chapurin; Melissa A Pynnonen; Rhonda Roberts; Kristine Schulz; Jennifer J Shin; David L Witsell; Kourosh Parham; Alan Langman; David Carpenter; Andrea Vambutas; Anh Nguyen-Huynh; Anne Wolfley; Walter T Lee Journal: Otolaryngol Head Neck Surg Date: 2017-02-14 Impact factor: 3.497