70 likes | 182 Views
Mapping CEMs to SHARPn Phenotyping Algorithms. Tom Oniki, Ph.D ., Medical Informaticist , Intermountain Healthcare Hongfang Liu, Ph.D., SHARPn Data Normalization Team Leader, Mayo Clinic Susan Rea Welch, Ph.D., B.S.N., Informatics Research Fellow, Intermountain Healthcare.
E N D
Mapping CEMs to SHARPn Phenotyping Algorithms Tom Oniki, Ph.D., Medical Informaticist, Intermountain Healthcare Hongfang Liu, Ph.D., SHARPn Data Normalization Team Leader, Mayo Clinic Susan Rea Welch, Ph.D., B.S.N., Informatics Research Fellow, Intermountain Healthcare
Normalized, standardized data definitions Enable shared data across EHRs Enable use of text-derived and/or structured data Computable data input to high throughput processes Clinical conceptualization of data requirements Logical, not physical models: Flexible to physical implementations Optionality in contents used Why Map CEMs to Phenotyping Algorithms
Specific EHR structure/content used Selection criteria of cases for input Temporality of data Non-computable algorithm input, as written Aggregations, transformations, expressions and functions on EHR data in algorithm input specs: Compute Body Mass Index from Height and Weight data MAX (BODY MASS INDEX) GLUCOSE RESULT not coincident with PREGNANCY Not Solved by CEM to Algorithm Data Mapping
Snippet of Medicinally Managed Diabetes Algorithm Patient meets AT LEAST One of the following Criteria: At least 2 face-to-faceoutpatient visits [1/1/09-12/31/10] with a diabetes ICD-9-CM code [list of codes given] OR Any medications ordered in [list of brand and generic medication names] OR At least 2 face-to-face outpatient visits with a capillary glucose result (glucometer) in the measurement period OR with an abnormal blood glucose or HbA1c result [thresholds given], not a glucose tolerance test and not part of a pregnancy screen.