Biostatistician | Founder of Evidence in the Wild and Zetyra

I build statistical design tools and advise clinical trial teams on how to make better decisions before they commit years of time and tens of millions of dollars.

I spent a decade as a biostatistician at Fred Hutchinson Cancer Center and SWOG, focused on oncology trial design and Phase II adaptive methods.

Over that time, I kept seeing the same pattern: protocols finalized before assumptions were stress-tested, interim decisions made on instinct instead of simulation, and trials continuing long after the math suggested they should stop. I call these Zombie Trials.

I founded Evidence in the Wild and Zetyra around a simple belief: the most expensive experiment is the one that was doomed from the start.

What drives my work:

• Building Zetyra: the platform I wanted as a practicing biostatistician, combining validated methods with a modern, usable interface
• Providing consulting through Evidence in the Wild so teams can pressure-test trial designs before major investment
• Expanding access to statistical tools that are too often locked behind expensive enterprise software

Beyond work, I train for Ironman and marathon competitions. Pacing matters in both endurance sports and trial design.

Core philosophy:

• Web-first: shareable, browser-based design without installation barriers
• Collaborative: scenario comparison, project workflows, and professional reporting
• Transparent: visible methodology grounded in peer-reviewed methods and standard references

Evidence in the Wild provides strategic biostatistics consulting for clinical trial teams making high-stakes design decisions before protocol finalization, funding, or major operational spend.

Protocol Statistical Review

• Review draft protocols for design vulnerabilities before they become expensive execution problems
• Stress-test assumptions around sample size, endpoints, estimands, interim looks, and decision rules
• Identify where the design is overbuilt, underpowered, or operationally misaligned
• Typical turnaround: 1-2 weeks

Sample Size & Power Strategy

• Build defensible sample size justifications for protocols, grants, and regulatory documents
• Evaluate covariate adjustment and efficiency strategies such as CUPED and ANCOVA
• Quantify tradeoffs across power, timeline, enrollment burden, and budget
• Provide simulation-based operating characteristics when closed-form methods are not enough

Adaptive Design Feasibility

• Assess whether an adaptive design adds real decision value or just extra complexity
• Compare fixed, group sequential, Bayesian, and sample size re-estimation options
• Outline what would be required statistically and operationally to implement the design well

Interim Analysis & Futility Planning

• Design interim looks, stopping criteria, and futility frameworks before first patient in
• Support Bayesian predictive methods and frequentist alpha-spending approaches
• Translate statistical monitoring strategy into a design that stakeholders can defend

• Biotech teams preparing Phase I/II or Phase II/III designs
• Academic investigators developing grant-funded or cooperative-group studies
• CROs and sponsors needing independent design review
• Medical device teams building more rigorous trial design infrastructure

I focus on trial design strategy and statistical decision-making. I do not offer routine SAS programming, data cleaning, table/listing production, or general statistics tutoring.

→ Learn more about consulting services

Zetyra is a web-based statistical software platform with 13 calculators (11 validated pro calculators) spanning frequentist and Bayesian methodologies for clinical trial design.

Phase III oncology trials average $50-100M and take 4-6 years to complete. Traditional tools (PASS, nQuery) cost $5,000-$15,000/year per seat with no validation transparency. Zetyra provides the same rigor at a fraction of the cost, validated with 499 automated tests benchmarked against gsDesign, rpact, conjugate analytical solutions, and published clinical trials.

Frequentist Calculators:

• CUPED - ~30% sample size reduction using baseline-outcome correlations (FDA-endorsed covariate adjustment)
• Group Sequential Design - Interim monitoring with O'Brien-Fleming/Pocock boundaries (validated against gsDesign R package)
• Survival Analysis - Time-to-event sample sizing via Schoenfeld, Freedman, and Lakatos formulas
• Sample Size Re-estimation - Blinded and unblinded adaptive designs with promising zone methodology

Bayesian Toolkit (6 calculators, add-on aligned with FDA January 2026 draft guidance):

• Prior Elicitation - Quantile matching, ESS-based, and historical data methods
• External Data Borrowing - MAP prior, power prior, and commensurate prior approaches
• One-Arm & Two-Arm Design - Bayesian sample sizing for binary and continuous endpoints
• Sequential Monitoring - Bayesian interim analysis with posterior probability stopping rules
• Predictive Power (PPoS) - Probability of trial success given interim data for decision guidance at planned or ad-hoc interim analyses

Features: Free tier with basic calculators, pro tier with advanced frequentist calculators + project management and PDF export, Bayesian Toolkit available as add-on

→ Try Zetyra | Platform Whitepaper | Bayesian Toolkit Whitepaper

Building tools and providing expertise to make rigorous statistical design accessible to everyone running clinical trials.