How do I change my project's DUSTER configuration after I created it?
The ability to edit a REDCap project's DUSTER configuration is not currently available yet. For now, if you want your dataset to include more clinical variables automated by DUSTER, then you will need to create a new REDCap project via DUSTER.
Also, keep in mind that you could edit your project's fields, instruments, etc. as you could with any REDCap project. However, doing so may prohibit DUSTER from functioning normally in your project.
Where can I find out which clinical variables are available?
A wiki is currently available here. Note that this wiki is only a temporary solution and a permanent resource is planned.
How do I request a new clinical variable or feature?
Send an email to 'duster-support@stanford.edu' with details of your request.
How large of a cohort and how much data can be supported in a single DUSTER project?
There is not a direct and straightforward answer to this question.
Currently, the general recommendation is to limit a cohort in a DUSTER project to 600 records or less (i.e., patients). This recommendation is subject to change over time as DUSTER develops and improves.
As the size of a DUSTER project grows in both cohort size and the number of clinical variables, a project will eventually reach a point where attempts to fetch and save data may not work as intended.
Larger cohorts numbering in the thousands may have no trouble fetching data if data collection windows are nonrepeatable and have few clinical variables. On the other hand, smaller cohorts numbering in the hundreds may have trouble fetching data if data collection windows are repeatable and contain many clinical variables.
Additionally, an important factor that can be difficult to measure is the amount of data each individual patient may actually have. For example, consider a DUSTER project which contains a repeatable collection window that begins on the calendar day of hospital admission and ends on the calendar day of hospital discharge with each instance representing an inpatient calendar day. Then consider fetching data on this project for a cohort of 500 patients with a median hospital admission of 4 days. Now consider fetching data on this project for a cohort of 500 patients with a median hospital admission of 30 days. The amount of clinical data fetched and saved for the latter cohort will be much larger in comparison to the former cohort.