Building a Patient Registry That Actually Fills Studies

Most sites have a registry. Far fewer have a registry that produces enrollments. The difference usually isn't size — it's whether the data stays usable and whether anyone touches the list between studies.

What a registry is actually for

A registry earns its keep in one specific moment: a sponsor sends a protocol with a tight window, and you need to know within a day whether you can find fifteen people who plausibly fit. Everything about how you build it should serve that moment.

That means a registry is not a spreadsheet of everyone who ever called you. It's a searchable set of people who (a) consented to future contact, (b) have enough clinical detail attached to filter on, and (c) have been contacted recently enough that the phone number still works.

Capture consent at the right width

The single most common registry failure is a consent that only covers one study. Someone screen-fails for a diabetes trial, and eighteen months later you can't call them about a related protocol because the authorization was study-specific.

Work with your IRB on a standalone registry consent or HIPAA authorization that covers future contact about research generally. Ask for:

• Permission to retain screening and demographic data
• Permission to contact by phone, email, and text, captured separately
• A stated retention period and an easy withdrawal path

Get this approved once and every downstream conversation gets simpler.

Store the fields you'll actually search on

Sites commonly collect names and phone numbers and nothing else, which makes the registry useless for filtering. Decide on a short list of searchable fields per therapeutic area and enforce them at entry. For most sites that means:

1. Primary condition and rough diagnosis date
2. Key lab or measurement values with the date taken
3. Current medications relevant to your common exclusions
4. Prior trial participation and washout status
5. Distance or travel constraints
6. Reason for prior screen failure, in a coded field rather than free text

That last one matters more than people expect. A registry where screen-fail reasons are coded lets you instantly exclude people who failed for permanent reasons and surface those who failed for something time-limited.

Keep it warm

Contact data decays fast. A list you haven't touched in two years is mostly noise, and coordinators learn to distrust it — which is worse than not having it.

Build a light, standing touchpoint. A quarterly email with a plain-language study update and a one-click "my info changed" link does most of the work. Log bounces and disconnected numbers immediately and mark those records stale rather than deleting them; a stale flag tells the next coordinator not to waste ten minutes.

When you do reach someone for a study, update the record whether or not they enroll. The value compounds only if updating is part of the call, not a separate task nobody has time for.

Segment before you need to

Don't wait for a protocol to start slicing. Maintain a handful of saved segments for the indications you see repeatedly — say, "type 2 diabetes, A1c above 7.5 in last 12 months, no GLP-1" — and refresh them monthly. Feasibility questionnaires get much easier to answer honestly when you can quote a real count.

Measure it

Track how many enrollments in each study came from the registry versus paid outreach or referrals. If registry-sourced enrollments stay near zero across several studies, the problem is the fields or the consent, not the list length. Fix the structure before you spend anything on growing it.