Study design, analysis, and the science behind both.
STEP-CM Analytics helps researchers and innovators design sound studies and build the evidence behind their claims — pairing statistical rigor with biomedical expertise rooted in veterinary medicine.
Built for teams where the data has to be right.
From single-investigator studies to biotech programs, STEP-CM Analytics brings statistical rigor and modern, interactive reporting to every stage of the work.
Veterinary researchers
Study design, sample-size and power analysis, and statistical modeling that stands up to peer review and regulators.
Veterinary biotech
Trial design, endpoint strategy, and analysis for companies developing therapeutics and diagnostics in animal health — plus the domain insight to know how the field will receive them.
Human biotech
Statistical support for life-science teams — from early exploratory analysis to rigorous, decision-grade study readouts.
An independent partner for rigorous analysis.
STEP-CM Analytics is built on a simple conviction: good data improves lives — animal and human alike — and research decisions are only ever as strong as the design and analysis behind them. The name comes from the three disciplines behind the work: Statistics, Epidemiology, and Comparative Medicine.
Ben Ouyang PhD, VMD, MSE, MS
I'm trained in veterinary medicine and applied mathematics, and I've been analyzing research data since 2004 — including five years in biotech as a senior biostatistician. STEP-CM Analytics is where those threads come together: I help researchers and life-science teams design studies that can genuinely answer their questions, and turn the resulting data into something people can act on.
My path through veterinary medicine ran through research and infection control rather than clinical practice — a deliberate choice. It kept me close enough to the clinic to understand how the field actually works, while my focus stayed on the science and the evidence behind it.
On the quantitative side, my graduate work in applied mathematics and statistics sat within a school of engineering, and ranged well beyond conventional biostatistics — into epidemiology, game theory, and graph theory. That breadth is why I tend to reach for the method that fits the problem rather than the one that's most familiar.
What drives me is a conviction that good data improves lives — animal and human alike. My work in veterinary medicine has always been One Health oriented, holding animals, people, and the environment in the same frame rather than treating them as separate problems. That perspective shapes how I approach a study: I'm looking for sustainable, future-proof answers, not just a result that clears a threshold.
I also care a great deal about how findings are communicated. An analysis only a statistician can interpret hasn't finished its job. I work to present results in a way that fits the audience — whether that's a review panel, a leadership team, or a clinician deciding what to do next.
Most of all, I try to listen well. Staying aligned with what a client actually needs is an ongoing process, not a box checked at a kickoff meeting. It's something I work at throughout a project, and never take for granted.
Rigor you can defend
Methods chosen to fit the question and the data — reproducible, documented, and ready for anyone who checks the assumptions.
A One Health lens
Animals, people, and the environment held in the same frame — so the answers point toward sustainable, future-proof solutions.
Alignment with your goals
Listening well is ongoing work, not a kickoff formality. I stay aligned with what you actually need throughout the project.
Statistical support, start to finish.
Engage for a single analysis or as an ongoing partner across the life of a study — from the first hypothesis to the final decision-ready report, plus the domain advice that helps a product find its footing in the field.
Scientific advisory & veterinary insight
Not every question is a statistical one. Trained in veterinary medicine and working across research and industry, I help teams work out what evidence their product actually needs, connect with the researchers and clinicians who can help build it, and understand how veterinarians in the field are likely to receive it. For startups especially, that domain perspective opens doors that data alone cannot.
Study & experimental design
The most valuable statistics happen before any data is collected. I help you frame testable hypotheses, choose endpoints, plan randomization and controls, and run sample-size and power calculations — so your study is set up to give an interpretable answer.
Statistical analysis & modeling
Analysis matched to your design and defensible under review — mixed-effects models for repeated measures, survival analysis for time-to-event data, longitudinal and Bayesian methods where they fit. I check assumptions, quantify uncertainty honestly, and document every step.
Interactive data visualization
Results are only useful if people can see them clearly. I build interactive dashboards and visual reports that let collaborators, reviewers, and stakeholders explore the data themselves — combining a strong R analysis foundation with modern, interactive presentation.
Reproducible reporting
Every analysis is delivered as a transparent, well-documented pipeline you can rerun and audit. That means fewer surprises, easier revisions, and reports you can confidently hand to a journal, a regulator, or an investor.
A simple, collaborative process.
Most projects follow the same clear arc — so you always know what happens next and what you'll receive.
- Discovery — we talk through your question, data, and constraints
- Plan — a written analysis or design plan you approve
- Execution — analysis with checkpoints along the way
- Delivery — reproducible report, visuals, and a plain-language walkthrough
Selected work & interactive analyses.
A showcase of study designs and analyses — reimagined as interactive, explorable visuals.
Longitudinal outcomes
Mixed-effects modeling of repeated measures across a treatment cohort.
Survival analysis
Time-to-event estimates with clear, interactive confidence bands.
Diagnostic accuracy
Sensitivity, specificity, and ROC exploration for a screening assay.
Let's talk about your study.
Tell me what you're working on.
Whether you're planning a study, sitting on data you need analyzed, or building an analytics foundation for your program — send a few details and I'll follow up to talk it through.