Utilization of multiple imperfect assessments of the dependent variable in a logistic regression analysis

• Published in: Statistics in medicine Vol. 19; no. 1; pp. 99 - 111
• Main Authors: Magder, Laurence S., Sloan, Michael A., Duh, Show-Hong, Abate, Joseph F., Kittner, Steven J.
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 15.01.2000; Wiley
• Subjects: Adult; Aged; Algorithms; Biological and medical sciences; Case-Control Studies; Cocaine - adverse effects; Cocaine-Related Disorders - complications; Computerized, statistical medical data processing and models in biomedicine; Humans; Hypertension - complications; Likelihood Functions; Logistic Models; Medical sciences; Medical statistics; Middle Aged; Risk Factors; Sensitivity and Specificity; Smoking - adverse effects; Stroke - chemically induced; Stroke - etiology; Human; Nervous system diseases; Stroke; Correlation; CNS stimulant; Psychotropic; Cardiovascular disease; Regression analysis; Cerebral disorder; Local anesthetic; Vascular disease; Statistical method; Missing data; Ester; Logistic regression; Information source; Multiple system; Central nervous system disease; Classification; Imperfect information; Drug of abuse; Cocaine; Maximum likelihood; Cerebrovascular disease
• ISSN: 0277-6715; 1097-0258
• DOI: 10.1002/(SICI)1097-0258(20000115)19:1<99::AID-SIM327>3.0.CO;2-O

Often, in biomedical research, there are multiple sources of imperfect information regarding a dichotomous variable of interest. For example, in a study we are conducting on the relationship between cocaine use and stroke risk, information on the cocaine use of each study patient is available from three fallible sources: patient interviews; urine toxicology testing, and medical record review. Regression analyses based on a rule for classifying patients from this information can result in biased estimation of associations and variances due to the misclassification of some subjects and to the assumption of certainty. We describe a likelihood‐based method that directly incorporates multiple sources of information regarding an outcome variable into a regression analysis and takes into account the uncertainty in the classification. The method can be applied when some sources of information are missing for some subjects. We show how the availability of multiple sources can be exploited to generate estimates of the quality (for example, sensitivity and specificity) of each source and to model the degree to which missing data are informative. A fitting algorithm and issues of identifiability are discussed. We illustrate the method using data from our study. Copyright © 2000 John Wiley & Sons, Ltd.