Enrollment Reports Sponsors Don't Have to Double-Check

Sponsors and CROs read dozens of site reports a month. Most get skimmed. A few get quoted in enrollment calls, used to justify extending your contract, or cited when a sponsor decides where to send the next protocol. The difference usually isn't performance — it's whether the numbers hold up under questioning.

Report the funnel, not the total

A report that says "4 randomized this month" tells a sponsor nothing actionable. A report that shows how 180 inquiries became 4 randomizations tells them where the protocol is leaking.

At minimum, break out:

• Inquiries received, by source
• Pre-screens completed
• Pre-screen failures, with top three reasons
• Screening visits scheduled, completed, and no-showed
• Screen failures, with the specific criterion that failed
• Randomized

The screen-failure criterion breakdown is the single most valuable column you can produce. When a sponsor sees that eleven of your fourteen screen failures tripped the same lab threshold, that's a protocol conversation, not a site performance conversation. Sites that report this consistently get taken seriously when they ask for an amendment.

Define your denominators once and never move them

The fastest way to lose credibility is to have last month's "pre-screened" number mean something different from this month's. Decide now what counts:

• Does a voicemail left count as an inquiry contacted?
• Is a patient who reschedules twice counted once or twice in screening visits?
• When does a pending consent move out of the pipeline?

Write the definitions down, put them in a footnote on the report, and keep them stable across the whole study. If you must change a definition mid-study, restate prior months on the new basis and say so explicitly.

Separate what happened from what you expect

Sponsors tolerate bad news. They don't tolerate surprises. Keep actuals and projections visibly distinct — actuals in the main table, projections in a separate section with your assumptions stated.

A defensible projection looks like: "Nine patients in active pre-screen, six expected to reach screening based on our running pre-screen-to-screen rate, four expected to randomize. Projecting three to five randomizations next month." A range with reasoning beats a confident single number every time, because when you miss a range you still look like you understand your own pipeline.

Explain variance before you're asked

Every report should carry two or three sentences of narrative. Not apology — explanation with a decision attached.

"Screening dropped from nine to four. Two weeks of our PI's availability were lost to coverage at the hospital. We've added Thursday evening screening slots starting the 14th and expect to recover by month end."

That's cause, impact, and remedy in three lines. It reads as a site that is managing the problem rather than reporting it.

Make the operational details boring

A few habits that pay off disproportionately:

1. Send on a fixed cadence. Same day each month, even when the numbers are bad. Erratic reporting reads as hiding.
2. Keep the format identical. Sponsors build mental models of your report. Reordering columns forces them to re-read.
3. Reconcile against EDC before sending. If your report and the sponsor's system disagree, you will spend the next call defending arithmetic instead of discussing strategy.
4. Keep a dated archive. When someone asks in month nine what you projected in month three, you want the actual file, not a memory.

The goal isn't to look good. It's to be the site whose numbers nobody has to verify.