What this page does not do
What this page offers instead: the structural difference between the two models — which is verifiable and stable — and a list of questions to ask your own installation. It is less spectacular and it stays true.
The structural difference
It concerns not features but the distribution model, and it has direct practical consequences.
| REDCap | AnzarSeha | |
|---|---|---|
| Model | Software distributed to a consortium of institutions, installed and administered by each | Hosted application, administered by its publisher |
| Access condition | That your institution is a member and runs an installation | An account |
| Administration | By the institution's IT department | By the publisher |
| Control over hosting | Total — the institution hosts it | None — a real disadvantage, mitigated by export in open formats |
| Maturity and community | Long established, abundant documentation and peer support | Recent, with no established community |
| Starting point of a study | An empty project to build | A dictionary generated from the pathology catalogue |
The two points where AnzarSeha makes a different choice
They are stated as choices, not presented as superiorities.
1. The dictionary is generated, not built
A research project starts in AnzarSeha with an already-typed dictionary drawn from the pathology catalogue: staging, work-up, scores and markers are there, with their units and bounds. The work becomes adjustment rather than construction. The trade-off is plain: outside the covered pathologies this advantage disappears, and the form is built variable by variable, as elsewhere.
2. Extraction from free text is a first-class capture route
The provenance of a value — registry, rule, model quoting its source sentence, human decision — is a column of the data model, not a data-entry convention. This choice targets the retrospective case, where most information sits in reports. Detail: assisted extraction and provenance.
Questions to ask your installation
Rather than a feature table written from a distance, here is what actually separates a usable installation from a nominal one. These questions apply to any tool, ours included.
- Can I open a project this week, or does it require a request? Administrative delay is the leading cause of abandonment.
- Does the log record a reason, or only the change? The distinction is treated in the audit trail.
- Is there a freeze state distinct from the end of data entry? Without one, nothing separates capture time from analysis time.
- Does the export include the dictionary and the missing-value table, or only the data? See the dataset.
- Are reasons for absence distinct from an empty value?
- Who can delete a project, and does deletion leave a trace?
Frequent questions
Is REDCap free?
REDCap is distributed to a consortium of institutions, largely academic and non-profit, under terms set by its developers at Vanderbilt University. What it costs an end user therefore depends entirely on their institution's policy, and should be checked with them rather than inferred from a web page.
Can data be migrated between the two?
In both directions the difficulty is not moving the data but moving the definitions. Columns transfer easily; a typed dictionary, bounds and reasons for absence require rework. Plan a dictionary-building pass either way.
Why does this page not conclude?
Because the conclusion depends on two facts we do not know: whether your institution runs an installation, and which pathology you work on. A comparison page that concludes without those two is concluding about something other than your case.