The evidence base for new medicines still describes a narrower population than the one being treated.

Enrolment diversity in clinical research has received sustained regulatory attention on both sides of the Atlantic, and the direction of travel is genuinely positive. The distribution of that improvement is not even.

Where progress is real

Sex balance in most therapeutic areas is now broadly reasonable, and reporting of enrolment demographics is far more consistent than it was a decade ago. Regulators asking a question repeatedly does change behaviour.

Where it is not

Older adults with multiple conditions remain systematically under-enrolled, largely because the exclusion criteria that keep a trial clean also exclude the typical patient. Pregnancy and lactation remain evidence deserts by design, which leaves clinicians extrapolating.

Rural participation lags because trial sites cluster at academic centres, and participation cost, including time off work and travel, falls hardest on the people least able to absorb it.

The design changes that help

Decentralised visits, reimbursement that covers real costs rather than nominal ones, community site partnerships, and pragmatic trial designs that accept messier data in exchange for a population that resembles the clinic.

What to ask of any new approval

Read the enrolled population, not the indication. If the indication is broader than the population studied, the difference is the clinician's judgement, and it should be labelled as such rather than implied.