Ascertainment of chronic diseases using population health data: a comparison of health administrative data and patient self-report

• DOI: 10.1186/1471-2458-13-16
• Publication Date: 2013-01-09T20:32:37Z
• Authors: Elizabeth Muggah; Erin Graves; Carol Bennett; Douglas G Manuel
• Keywords: Stroke; Concordance; Biostatistics; Cohen's kappa

Health administrative data is increasingly being used for chronic disease surveillance. This study explored agreement between and survey ascertainment of seven key diseases, using individually linked from a large population individuals in Ontario, Canada. All adults who completed any one three cycles the Canadian Community Survey (2001, 2003 or 2005) agreed to have their responses provincial health were included. The sample included 85,549 persons. Previously validated case definitions myocardial infarction, asthma, diabetes, lung disease, stroke, hypertension congestive heart failure based on hospital physician billing codes identify cases these compared with self-report each survey. Concordance was measured Kappa statistic, percent positive negative prevalence estimates. Agreement statistic good very (kappa range: 0.66-0.80) diabetes hypertension, moderate infarction asthma poor fair 0.29-0.36) COPD. Prevalence higher all diseases except stroke infarction. Utilities Index scores identified by self-reported some (acute failure), suggesting that may pick up less severe cases. In general population, discordance many particularly low prevalence, differences not easily explained individual characteristics.