Deanna Schreiber-Gregory is a Lead Research Statistician and Data Manager on contract through the Henry M Jackson Foundation for the Advancement of Military Medicine to the Department of Defense in Bethesda, MD. She is also an Independent Consultant for Statistics, Research Methods, and Data Management in the private sector through Juxdapoze, LLC. Deanna has an MS in Health and Life Science Analytics, a BS in Statistics, and a BS in Psychology. Deanna has presented as a contributed and invited speaker at over 50 local, regional, national, and global SAS user group conferences since 2011.
Peter Flom is a retired independent statistical consultant who worked with graduate students and researchers in the social, medical and behavioral sciences. He has been using SAS for over 20 years and has given talks at SGF and many local and regional groups.
Patricia Ferido is a Senior Research Programmer at the Leonard D. Schaeffer Center for Health Policy and Economics where she analyzes medical data for research on dementia care and treatment. Prior to joining the Schaeffer Center, she worked as an economics litigation consultant specializing in the analysis of labor data for wage and hour litigation. She holds a BA in both Economics and International Development Studies from UCLA and is pursuing a Masters in Public Policy Data Science at USC.
Bruce Lund is a statistical modeling consultant and trainer. For 15 years he was a statistical and modeling consultant for OneMagnify of Detroit. Before OneMagnify, he was the customer database manager at Ford Motor Company and a mathematics professor at University of New Brunswick, Canada. He has a mathematics PhD from Stanford University. Bruce Lund has presented at SAS Global Forum, SAS AnalyticsX, ASA CSP, and at regional SAS user group conferences.
How Sick is my Cohort of Patients? A General Approach to Identifying Chronic Conditions
June 22, 2021 10:00 AM - 2:00 PM PacificInstructor: Patricia FeridoWith the COVID-19 pandemic, the need for evidence-based healthcare research has become increasingly apparent. Even before the recent health crisis, the volume of available data on healthcare had been growing exponentially. Claims data and electronic health records provide rich insight into the health status of patients and the care provided by health care systems. Successfully uncovering these insights, however, requires an understanding of the data, as well as standardized and validated methods of analysis. This class will provide an overview of best practices when working with claims data, specifically Medicare claims data. Topics covered will include: the general structure of claims data, how best to use that information to identify disease cohorts, different approaches for measuring the health status of patients (e.g., Charlson Comorbidity Index, the Elixhauser Comorbidity Index, Hierarchical Condition Category Coding, etc.), and a deep dive into the Chronic Condition Warehouse (CCW) algorithms. Finally, the class will conclude with the workshopping of a SAS Macro that applies CCW-like rules to any dataset that resembles insurance claims or electronic health records with a full-picture of diagnoses and procedures from patient medical visits. The macro package includes the CCW validated algorithms (the default option), but also has the flexibility for the user to apply the algorithm to a different set of diagnoses and procedures. The user can either implement variations of the CCW-definitions or identify entirely new conditions, so long as they can be implemented using diagnosis or procedure codes, claim types, and CCW-like rules. After taking the class, students will have an understanding of key factors to consider in disease cohort analysis and will have direct experience using this package to identify diseases in simulated data.
Elementary Logistic Regression with Predictive Modeling
June 25, 2021 10:00 AM - 2:00 PM PacificInstructor: Bruce LundThis class presents light theory, supported by simulations, for understanding binary logistic regression models using SAS®. This discussion of logistic regression begins at the beginning. No prior experience is assumed.Once the basics of logistic regression are introduced, the class focuses on using logistic models in predictive modeling on large datasets. Examples from credit risk and automotive marketing are given. The class will be less focused on explanatory models as would arise in the bio-sciences.Topics include: Logistic regression versus other methods; Likelihood function and maximum likelihood estimators; Statistics for predictor and overall model fit; Screening, binning, transforming of predictors (including weight of evidence coding); Discussion of multicollinearity; Predictor selection methods using PROC LOGISTIC, HPLOGISTIC, HPGENSELECT including best subsets, stepwise with sbc/aic, Lasso; Model validation and assessment including c statistic, R-squares classification error, and lift charts in the context of training, cross-validation, and validation samples.Class uses BASE SAS and SAS/STAT. No usage of Viya or Enterprise Miner.
Note that classes can be taken as a series or as individual classes.