Byron Jaeger

Hello, and thanks for visiting!

I’m the director of data analytics at Perisphere Real World Evidence, where I lead our team of biostatisticians and develop the pericircle family of R packages. Our goal is to move research beyond publications and positively impact health outcomes using real world evidence.

If you’d like to learn more about my background in statistical computing, you may find a few software projects I’ve developed helpful:

• aorsf, an R package with fast routines to fit oblique random forests for classification, regression, and survival outcomes.

• Cardiometabolic statistics for US adults, a web application to explore trends in risk factors for cardiovascular disease among US adults from 1999 to 2020. Online tutorials are available as well as a publication describing design and rationale.

• r2glmm, an R package to compute \(R^2\) statistics for fixed effects in the generalized linear mixed model. A tutorial is available on GitHub and a publication is available from the Journal of Applied Statistics.

What I’ve been working on more recently

• pericircle, a slice of Perisphere with R packages that help write clear, effective code across our projects. The family includes perinary (data dictionaries), periglue (rounding and formatting), peripal (color palettes), pericircumference (project templates built on targets), and peridefs (validated condition definitions for claims data). Together, they take an analysis from raw data to publication-ready tables, figures, and text. See them work together →

Before Perisphere

Before joining Perisphere, I was a faculty biostatistician at Wake Forest University School of Medicine, and before that at the University of Alabama at Birmingham.

• Oblique random forests. I introduced oblique random survival forests and later developed aorsf, a fast and interpretable implementation for survival, classification, and regression (JCGS, 2023). The oblique random survival forests page collects the papers, software, and talks.

• Blood pressure and cardiovascular epidemiology. With colleagues, I studied national trends in blood pressure control using NHANES, including papers in JAMA and Hypertension, and the cardioStatsUSA web application.

• Clinical trials. I worked on data coordination, reporting, simulation, and analysis for trials including SPRINT, LEAP, and MoTrPAC. I led a study of longer-term mortality after intensive blood pressure control in SPRINT, published in JAMA Cardiology.

• Prediction and dementia. I worked with cognitive science experts to build prediction models that could help direct cognitive screening to patients at risk for cognitive impairment or dementia, and I led a workshop on oblique and causal random forests for the MELODEM consortium.

• Teaching. I taught introductory biostatistics and Introduction to R, and I still enjoy teaching R through talks and workshops.

See Selected publications for highlights.

Working together

I enjoy helping research teams turn complex data into clear, reproducible results. Some of the ways I can help:

• Designing and running analyses of real-world data, including claims and electronic health records
• Risk prediction and machine learning for time-to-event outcomes
• Building reproducible pipelines that generate publication-ready tables, figures, reports, and slides
• Developing R packages and tools that can support teams and simplify complex statistical processes

If you’d like to discuss working together, please reach out through Perisphere Real World Evidence or GitHub.