Andrew Tredennick, Ph.D.

Senior Scientist II

Andrew joined MetrumRG in September 2022 as a Research Scientist. He earned his PhD in Ecology from Colorado State University, followed by a National Science Foundation Postdoctoral Research Fellowship in Biology and Mathematics at Utah State University. Nearly all of Andrew’s work relies on confronting mechanistic models with data to make predictions and inference. His interest in fitting dynamical models to data led to a postdoctoral appointment at the University of Georgia’s Center for the Ecology of Infectious Diseases. Andrew then worked for four years as a statistical consultant in the environmental space before joining MetrumRG. Andrew has completed projects on early warning signals for disease re-emergence and forecasting models of emerging infectious diseases, like COVID-19.

Recent publications by this scientist

A MODEL INFORMED DRUG DEVELOPMENT (MIDD)-BASED QUANTITATIVE DECISION FRAMEWORK (QDF) FOR IMPROVING R&D PRODUCTIVITY: PROOF OF CONCEPT FOR ATOPIC DERMATITIS (AD)

March 18, 2026

A MODEL INFORMED DRUG DEVELOPMENT (MIDD)-BASED QUANTITATIVE DECISION FRAMEWORK (QDF) FOR IMPROVING R&D PRODUCTIVITY: PROOF OF CONCEPT FOR ATOPIC DERMATITIS (AD)
E. Anderson¹, BW. Corrigan¹, M. Cala Pane¹, A. Tredennick¹, T. Dunlap¹, L. Lomeli¹, B. Davis¹, MR.Gastonguay¹1Metrum Research Group, Boston, MA

Project Rationale

QDF Components QDF Components
Competitive Landscape
MIDD Enhanced Valuations

Rising costs, uncertain reimbursement, competition, and declining success rates have
reduced drug R&D productivity and investment over the last decade.
Proposed strategies to improve R&D productivity include four key factors: 1) leveraging all
data sources; 2) utilizing quantitative models; 3) elimination of information silos across R&D
and commercial organizations; and 4) application of decision frameworks to reduce
cognitive bias and improve decision making.1

A QDF for a drug development program in atopic dermatitis (AD) was developed to: 1) link
MIDD models aligned with a target product profile (TPP) to risk-adjusted net present value
(rNPV); and 2) integrate context-sensitive large language models (LLMs) to incorporate
non-structured data from novel sources into the decision-making framework in a responsible
manner.

Download PDF

Gompertz Cure Rate Survival Models with Stan and brms.

December 6, 2024

Presented at ACoP 2024. A Gompertz distribution is implemented in Stan and brms to enable cure rate survival modeling. The Gompertz brms family can be used to model exposure-response data in a Bayesian framework where a proportion of the population never experiences the event of interest.

Download PDF