Epidemiologist — Patient Population Sizing

What you would do

• Build and review evaluation frameworks for assessing commercial therapeutics and pipeline development, specifying methodological rigor standards.
• Size patient populations for specific drug indications, converting prevalence and incidence into diagnosed, treated, and addressable patient counts.
• Evaluate whether population-size estimates are methodologically sound, well-sourced, and appropriate for commercial market assessment.
• Review real-world data sources including insurance claims, EHR systems, and disease registries, flagging limitations and data quality concerns.
• Apply population-sizing expertise across the project's set of drugs and indications to ensure consistent standards and identify patterns in estimation errors.

Who they want

• US-based epidemiologist with postgraduate training in epidemiology, biostatistics, or public health and 5+ years actively practicing
• Hands-on experience converting real-world data sources (insurance claims, EHR, registries) into patient population estimates and market-sizing models.
• Familiarity with biotech and pharmaceutical development programs, including understanding of clinical trials, indications, and commercialization pathways.
• Expertise in prevalence and incidence calculations, and ability to explain the funnel from total disease population to treated patients.
• Comfort with evaluating trade-offs in data sources: registry completeness, claims data gaps, EHR coding accuracy, and each source's bias.

Main skills

What the interview asks about

1. 1.Patient population calculation

   The core of this work: converting epidemiological data into actionable patient counts. Interviewer assesses whether you distinguish prevalence from incidence and understand the diagnosed-to-treated funnel.

   For example: “Hypertension affects 42% of adults over 30. EHR data shows 85% are treated. Insurance claims identify 30 million on medications. What's your addressable population?”

2. 2.Real-world data strengths and limits

   Every data source has blind spots. Interviewer checks whether you recognize the limitations of claims data, EHR systems, and registries-not just use them naively.

   For example: “You're using Medicare claims data to size a population. The indication needs genetic testing for diagnosis, but claims capture test codes inconsistently. How does this affect your estimate?”

3. 3.Spotting methodological errors

   Catching estimation errors that look plausible but are wrong is the value you bring. This tests your ability to identify missing confounds, bias, and unsupported leaps in logic.

   For example: “A team estimates 50,000 addressable patients using meta-analysis of five RCTs from academic medical centers. The studies didn't include community practice. What would you flag?”

4. 4.Pharma development context

   Understanding drug development helps you assess whether population estimates make sense for the clinical and commercial context-early-stage vs. approved, first-line vs. adjunctive.

   For example: “You're sizing a Phase 2 diabetes drug for a genetic subtype. Registry data: 500,000 people. Current SOC reaches 15% of eligible patients. How does development stage inform addressable population?”

5. 5.Rubric design and calibration

   You'll evaluate AI-generated population estimates. Building and refining rubrics requires translating epidemiological rigor into teachable criteria.

   For example: “You're writing a rubric criterion: 'source appropriateness.' A response uses published prevalence from 2005 and current insurance claims. How would you score this?”

6. 6.Documentation and reasoning clarity

   Population sizing must be transparent so commercial and clinical teams can challenge it. Interviewer checks whether you can distinguish between stated assumptions and hidden ones.

   For example: “An estimate of 200,000 patients uses one EHR study showing 8% prevalence. The estimate doesn't note that EHR misses undiagnosed patients. How would you flag this?”

A task you may get

Given prevalence and treatment rate data from multiple sources, size the patient population for an indication, document assumptions and limitations, and critique another estimate.

How to prepare

• Review 2-3 published prevalence studies you know well. Note case definitions, data sources used, and populations that may have been missed.
• Find pharma market-sizing reports or clinical development rationale from company presentations. Identify estimates and think through what data would justify them.
• Practice converting published prevalence into addressable population accounting for diagnosis and treatment rates and eligible subgroups.
• Think of a population estimate you've made that turned out wrong. What data source or methodological assumption was flawed?

The facts

Pay

$150–175/hr

Hours

Part time, 15 hours a week

Where

Remote · Remote — US based or deep US market experience

Field

Life, Physical, and Social Science

Posted

8/5/2026

Places left

5

We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.