A/B Testing for Health Programs: Can We Borrow from Tech Without Harming Patients?

A plain-language guide to experimentation in public health and development programs

A/B Testing
Impact Evaluation
Health Programs
Author

Nichodemus Amollo

Published

November 5, 2025

Why Tech Loves A/B Tests (And What Health Can Learn)

In tech:

  • A/B testing is used to:
    • Optimize click-through rates
    • Test new features
    • Improve revenue

In health and development:

  • Stakes are different:
    • We’re dealing with well-being, safety, and sometimes life-or-death

But we can borrow experimentation ideas carefully.


What A/B Testing Looks Like in Health

Examples:

  • Testing two reminder messages for clinic appointments
  • Comparing two formats of health education materials
  • Evaluating two outreach strategies for vaccination

Key safeguards:

  • Ethical review and approvals
  • Informed consent where appropriate
  • Clear principle: no group is offered worse than current standard of care

How This Relates to RCTs

A/B tests are, in essence:

  • Simple randomized controlled trials with:
    • A small number of arms
    • Large samples
    • Clear short-term outcomes

Differences:

  • In tech: outcomes are often clicks and purchases
  • In health: outcomes are attendance, adherence, health status, equity

How a Junior Analyst Can Contribute

  • Help design:
    • Outcome measures
    • Randomization procedures
  • Help implement:
    • Data collection tools
    • Analysis pipelines
  • Help interpret:
    • Not just “did it work?” but “for whom?” and “why?”

Experimentation in health requires both rigor and humility. If you can bring both, you’ll be a huge asset.