Summit Epi Analytics · Colorado Springs

Independent causal inference and trial design for interconnected populations.

A small practice for trial consortia, NGOs, behavioral & infectious disease researchers, ministries, donors, and impact investors who need rigorous answers to hard questions about how interventions actually work — and how to evaluate them when interference, networks, and spillover break standard methods.

Request a scoping call Read the operating principles
$2MNIH R01 PI, network-informed HIV prevention trial design
3×Atlantic Causal Inference Conference data-analysis competition winner
treatSensAuthor, CRAN package for assessing sensitivity to unmeasured confounding
FDATargeted Learning webinar on causal inference with BART
What we do

Three engagement types. One practice.

Summit Epi Analytics offers a small number of carefully scoped engagements. Each is short enough to fit a working founder's calendar, deep enough to change a decision, and structured around a single methodological wedge: causal inference for interconnected populations where standard methods break.

Engagement · 3 weeks

Trial Design Review

$5,000 – $15,000 · fixed fee

A stress test of a planned cluster-randomized, stepped-wedge, or ring-vaccination trial for interference, contamination, and identifiability before data collection begins.

  • Methods memo with redesign options
  • 90-minute working session with PI and trial statistician
  • Optional analysis-plan co-authorship
View engagement details → Engagement · 6–8 weeks

Decision-Grade Modeling Brief

$30,000 – $60,000 · co-developed

A mechanistic, stochastic, or agent-based model that answers one bounded question, built in partnership with a named client-affiliated statistician.

  • Open, reusable code
  • In-country or on-site workshop deliverable
  • Co-authorship on resulting publications
View engagement details → Engagement · 2 weeks

Biosecurity Diligence Sprint

$10,000 – $20,000 · fixed fee

Independent technical due diligence on a biosecurity, surveillance, or outbreak-forecasting product — modeling validity, claim verification, operational fit.

  • Structured diligence memo
  • 60-minute decision-maker debrief
  • Open evaluation framework, confidential application
View engagement details →
Why this practice exists

Most causal inference experts don't do scenario modeling. Most infectious disease modelers don't do rigorous causal inference.

Summit Epi Analytics sits at that intersection. The work is grounded in two decades at the interface of causal inference, network methods, and infectious disease — from a $2M NIH R01 on network-informed HIV prevention trial design, to operational global health collaborations with the Rwanda Biomedical Centre, the Botswana–Harvard Partnership, and the WHO Immunization programme.

The combination is rare on purpose. Trials that ignore interference misunderstand the risk of the populations they are meant to protect. That is the gap this practice was built to close.

More about the practice →

Recent & representative work

Infectious disease surveillance program development and monitoring

  • Co-author on the 2025 IJID Regions paper led by Rwanda Biomedical Centre staff, evaluating airport-based pathogen surveillance.
  • Recommending adaptation to surveillance programs in response to emerging outbreaks of concern.

Causal inference under interference

  • Foundational publications in Epidemiologic Methods (2016) and Statistics in Medicine (2023, with Goyal et al.)
  • CRAN package treatSens for sensitivity to unmeasured confounding.
Operating principles

Independence is the product. Accompaniment is the method. Rigor is respect.

Clients pay for the willingness to tell them what they don't want to hear. We protect that structurally — through what we publish, what we co-author, and what we refuse to take on. The principles below run every inquiry before it reaches contract.

Publish openlyEvaluation methodology is public; only the client-specific application is confidential.
Reserve the right to publishFindings can be published with reasonable embargo. No permanent suppression.
Validation engagements are treated as developmental projects: our evaluation is independent of presupposed conclusions.
Loyalty to end usersWhen a tool's claims would mislead the populations it touches, we say so in the headline, not the appendix.

Read all six principles →

Engage

A short scoping call is the right first step.

Tell us what you're trying to decide, who else is at the table, and the deadline you're working against. We respond to every well-formed inquiry within five business days.

Request a scoping call