A site coordinator writes a patient's vitals on a paper source form during the visit. Later that day, or the next, someone re-keys those same numbers into the EDC system. Nothing about that process looks careless. But it means every data point in that trial passes through two separate moments of human transcription the original recording and the re-entry and every one of those moments is a chance for a transposed digit, a misread decimal, or a value copied into the wrong field to enter the record undetected until a monitor happens to catch it during source data verification.
This isn't a hypothetical. It's the default state of a large share of clinical data collection, and it's exactly the pattern that regulatory data integrity standards were written to catch.
ALCOA+ is the framework regulators use to judge whether clinical trial data can be trusted: data must be Attributable, Legible, Contemporaneous, Original, and Accurate — the original five criteria the FDA introduced in the 1990s plus Complete, Consistent, Enduring, and Available, the four principles added as data increasingly moved to electronic systems. Each of these is a specific, checkable property of a data point, not a vague quality goal:
A paper source document that gets manually re-keyed into EDC struggles against several of these at once. The re-keyed entry isn't the original record the paper form is. The moment of transcription introduces exactly the kind of copying error that "accurate" is meant to guard against. And if the re-entry happens hours or days after the visit, contemporaneity is already stretched thin before anyone reviews the data at all.
The regulatory risk is real, but the more immediate cost shows up in how much time a site spends managing it. Every transcribed data point is a data point a monitor eventually has to source-data-verify tracing the EDC entry back to the paper original to confirm they match. That verification work scales with every study, every visit, every field, and it's time a CRA or site coordinator spends checking for errors instead of doing anything else. And when a discrepancy does turn up, resolving it means a query, a site response, and a correction cycle that delays the very data lock the trial is working toward.
| Paper Source + Manual Re-Entry | Basic Digital Forms (no direct capture) | Cloudbyz eSource | |
|---|---|---|---|
| Number of transcription steps per data point | Two recording, then re-entry | Often still one re-entry step | Zero captured once, directly |
| Contemporaneous capture | Depends on when re-entry happens | Improved, but not guaranteed | Built in data is source and record at the same moment |
| Audit trail / attributability | Weak paper rarely shows who entered what, when | Partial | Full, timestamped, e-signature-backed |
| 21 CFR Part 11 alignment | Requires separate controls layered on | Varies by tool | Native |
| Site data-entry effort per patient | Highest full paper documentation plus EDC entry | Reduced | Cloudbyz reports roughly 8 fewer hours of site effort per patient |
| Source data verification time per patient | Highest every field checked against paper | Reduced | Cloudbyz reports roughly 2 fewer hours of SDV time per patient |
If more than one of these gives you pause, the risk isn't abstract it's already built into how your data gets collected.
Cloudbyz eSource captures clinical data directly at the point of care the entry made during the visit is the source record and the EDC record at once, with no separate re-keying step in between. It supports FDA 21 CFR Part 11-compliant e-signatures and eliminates the paper documentation that would otherwise need to be transcribed later. Because the record is created once, with a timestamp and an attributable signature at the moment of entry, contemporaneity and attributability aren't things a site has to work to demonstrate after the fact they're built into how the data was captured.
Cloudbyz reports that customers using eSource have eliminated 90% of paper-based processes and seen an 80% improvement in data quality, with roughly 37% less time spent on the overall data process. At the patient level, that translates to around 8 fewer hours of site effort and roughly 2 fewer hours of source data verification time per patient — time that goes back to the site and the monitor instead of to checking one system against another.
If your team is still tracing EDC entries back to paper source documents to verify them, it's worth seeing what direct electronic capture actually removes from that process. Book a demo with Cloudbyz team.