Author(s): Pacheco JA, Avila PC, Thompson JA, Law M, Quraishi JA,
Abstract Share this page
Abstract Our aim was to identify asthmatic patients as cases, and healthy patients as controls, for genome-wide association studies (GWAS), using readily available data from electronic medical records. For GWAS, high specificity is required to accurately identify genotype-phenotype correlations. We developed two algorithms using a combination of diagnoses, medications, and smoking history. By applying stringent criteria for source and specificity of the data we achieved a 95\% positive predictive value and 96\% negative predictive value for identification of asthma cases and controls compared against clinician review. We achieved a high specificity but at the loss of approximately 24\% of the initial number of potential asthma cases we found. However, by standardizing and applying our algorithm across multiple sites, the high number of cases needed for a GWAS could be achieved.
This article was published in AMIA Annu Symp Proc
and referenced in International Journal of Biomedical Data Mining