FeFe women's research panel logoStart recruiting free

For researchers

How to find participants for a study when the condition is rare

Sample sizes fall apart when the condition is uncommon or stigmatised. Here are the recruitment routes that work, what each one costs you, and where bias creeps in.

8 September 2026 · 3 min read

A woman scientist carrying research papers in a laboratory
Photo by Pavel Danilyuk on Pexels

Most recruitment advice assumes a general population sample. The moment your inclusion criteria name a condition, that advice stops working. A general panel of one hundred thousand people might hold a few hundred women with endometriosis who are willing to talk about it, and only a fraction of those will pass a properly written screener.

This post covers the routes available to you, in the order most teams should try them.

Key points

  • Estimate the real eligible population before setting a deadline.
  • Choose channels based on condition rarity and clinical access.
  • Track where recruitment bias enters the sample.

Start by working out how rare your sample really is

Before you choose a route, estimate how many people in a panel of a given size would qualify. Take the prevalence of the condition, cut it to the age band and country you need, then cut it again for anyone who has to be currently in treatment, currently symptomatic, or diagnosed within a certain window. Each extra criterion multiplies the loss.

If the number that comes out is under one in a thousand, no general panel will fill your study at a sensible price. Plan for a specialist route from the start rather than discovering it three weeks in. Our guide to recruiting for women's health research walks through the feasibility maths in more detail.

Route one: a screened health panel

A panel that already holds condition, treatment and cycle information can tell you the size of your eligible pool before you spend anything. You pay per completed response and you get a known incentive, a known consent position and a documented lawful basis.

The limit is panel size. A specialist panel is smaller than a general one by definition, so ask for the eligible count rather than the headline membership number.

Route two: clinics and NHS sites

Clinical recruitment can support diagnosis verification, but NHS research may require HRA approval, local capacity and capability arrangements, and other reviews depending on the study. [1] Women who gave up on the system years ago will not appear. For many women's health topics, those are exactly the people you needed.

Route three: patient groups and charities

Charities reach people no clinic sees. They will usually want to know what the research is for and what participants get out of it, and some will ask to see your materials first. Budget time for that conversation. Expect a sample that skews towards people who are engaged, articulate and often more severely affected, which is fine for qualitative work and risky for prevalence claims.

Route four: social media and community forums

Fast and cheap to start, and the hardest to defend. You cannot verify who answered, professional survey takers follow study links, and the same post shared into the wrong group can flood your quota in an hour. If you use this route, screen hard, cap by source, and record where each response came from so you can drop a bad source later.

Where bias enters, whichever route you pick

  • Self-selection. People with a strong story sign up first.
  • Digital exclusion is associated with age, income and other social factors, so online-only recruitment can produce systematic gaps. [2]
  • Diagnosis as a gate. If you require a formal diagnosis, you exclude everyone still waiting for one, which in conditions like endometriosis and PCOS is a large and systematically different group.
  • Language. English-only screeners narrow your sample in ways your report will not mention unless you make it.

Name the ones that apply in your limitations section. A reviewer will find them anyway.

Practical steps that raise fill rates

  1. Loosen one criterion. Ask which single criterion is doing the most damage to your eligible count, then ask whether the study really needs it.
  2. Accept self-reported diagnosis with a follow-up question about who diagnosed it and when.
  3. Pay properly. Underpaying is the most common cause of a stalled study. See how much to pay research participants in the UK.
  4. Run in waves. Field a small first wave, look at the profile of who came through, then adjust before you spend the rest of the budget.
  5. Ask for re-contact consent at the point of the first study, so a follow-up does not mean recruiting from scratch.

References

  1. Health Research Authority. What approvals and decisions do I need?
  2. Office for National Statistics. Exploring the UK’s digital divide. 2019.

Common questions

How many participants can I expect from a general panel for a rare condition?

Estimate it before you field. Take the prevalence, apply your age band and country, then apply every additional criterion such as current treatment or recent diagnosis. If the result is under one in a thousand panel members, plan for a specialist panel or clinical recruitment instead.

Should I accept a self-reported diagnosis?

For most non-regulated research, yes, with a follow-up question about who made the diagnosis and when. Requiring documented diagnosis excludes everyone still waiting for one, which in conditions such as endometriosis is a large group that differs systematically from the diagnosed group.

Is recruiting through social media acceptable?

It is acceptable if you screen hard, record the source of every response and cap how many completes any single source can supply. It is not acceptable as your only route if you plan to make claims about prevalence or representativeness.

Recruiting for a women's health study?

Filter a screened panel by condition, treatment and cycle stage. Free to set up, and you only pay for completed responses.

Start recruiting free

Keep reading