The Hardest Resource to Find Is People | AI Learning Guide

Think about why eligibility criteria exist. They reduce noise, so a real effect is easier to see. But that same filtering shrinks the pool of people who can join. Then a second filter applies: can those people actually get to a site, and are they willing to accept the chance of a placebo? Many never hear about the trial at all. Now add the retention problem. Even a trial that enrolls well can lose participants to side effects, travel burden, or simple life changes. When enough people leave, the trial becomes underpowered. That does not mean the drug failed; it means the trial could not tell whether it worked. So recruitment and retention are not administrative details. They set the pace of the entire program, and they depend on human circumstances that no design tool can optimize.

0:00 / 0:00

Auto-play

Eligibility criteria narrow the pool for scientific reasons, but the pool is then narrowed again by reachability, willingness, and the burden of participation. Retention adds a second filter: participants who leave reduce the trial's ability to detect a real effect.

What limits enrollment and retention

Narrow eligibility criteria: specific disease stage, prior treatments, age range, and excluded conditions
Limited sites: participants may need repeated travel to a research center
Awareness: many eligible patients never learn a trial exists
Willingness: some decline the chance of placebo or an unproven treatment
Dropout: side effects, inconvenience, relocation, or loss of interest reduce the final analyzable group

A trial that loses too many participants can become underpowered: it may fail to show a real benefit simply because too few people remained to measure it. The result is not 'the drug does not work' but 'this trial could not tell.'