Data Request
How to ask for data that isn’t already in the catalogue — describe what you need in chat, let the assistant turn it into a structured, checked specification, and submit it for sourcing. Everything you submit is tracked under My Requests.
When to use it
Use Data Request when the cohort you need isn’t available in Data Insight. Instead of querying data that exists, you describe the data you want — indication, modality, sample size, region, biomarkers — and the assistant builds a precise request that PAICON can source against its clinical network.
Open it from Data Request in the sidebar.
What the assistant does for you
- Turns conversation into a specification. Every answer you give is saved into a structured request — indication, modality, specimen and tissue types, stains and IHC results, biomarkers, sample size, region, and more. You can see the specification growing in the panel beside the chat.
- Asks the right questions. Where a detail is missing, the assistant asks — often as clickable options rather than free text, so you pick instead of type. Fields that naturally take several values (specimen types, biomarkers, stains, ICD-10 codes) allow multi-select.
- Resolves the codes. Describe a cancer type or diagnosis in plain words; the assistant matches it to OncoTree and ICD-10 and asks you to confirm. Locations are resolved to country codes.
- Handles more than one cohort. A request can contain several cohorts — for example a HER2-positive and a HER2-negative arm — each with its own criteria and sample size, checked and tracked together under one reference number.
- Checks availability before you submit. A count against connected data tells you how feasible the request is, per cohort, before anything is sent.
How it works
- Describe your need. Start a chat and say what you are looking for in plain language. You can attach up to three PDFs (2 MB each) — a protocol, an ethics summary, a study synopsis — and the assistant reads them for context.
- Answer the questions. Pick from the options offered or type your own. Each answer is saved as you go; the summary card in the panel shows which sections are filled and which are still open. You can change any answer later in the conversation.
-
Check availability. When enough is specified, use Check availability (or ask the assistant to). Each cohort gets one of three results:
Result Meaning What to do Successful A healthy number of matching cases Submit, or tighten further Limited results Fewer matches than the sample size you asked for Relax a constraint, or submit and let PAICON source the remainder Unsuccessful No matches with the current criteria Follow the suggestions the assistant gives — broaden the region, widen the indication, drop a biomarker — and check again - Refine and re-check. Adjust any part of the specification and run the check again. The panel shows the latest count per cohort. Iterate until the estimate looks right; there is no limit on checks.
- Submit. When the check is Successful or Limited, Submit becomes available. You get a reference number and a confirmation page, and the request appears under My Requests. A submitted request is read-only in the chat: the expanded summary card is its final state.
Availability is an estimate. It counts matching cases in connected data as a feasibility signal for scoping the request; it is not a guarantee of final delivery.
What happens after you submit
The confirmation page tells you which of three situations applies:
- Fully available. Your request is submitted for review. You receive a confirmation email with next steps.
- Partially available. PAICON delivers the matched records and contacts its partner network for the remainder; you are told how many of the requested records are covered.
- Not available in connected data. PAICON contacts its partner network to source the data and reaches out to you.
You also receive an email confirmation with the reference number.
Tracking under My Requests
My Requests lists every submitted request with its indication, reference number, cohort count, submission date, requested volume, and current status. The list updates live as the status changes.
| Status | Meaning |
|---|---|
| Submitted | Received and awaiting review |
| Approved | Reviewed and accepted for sourcing |
| Preparing | Data is being assembled for delivery |
| Ready | Delivery is prepared |
| Failed | Delivery could not be completed; the team has been notified |
| Rejected | Declined at review; the reviewer’s note is shown alongside the status |
When a request is Ready, the Actions column offers the download where self-service download is enabled for your account; otherwise it gives a pre-filled contact link to arrange the transfer.
Tips
- Start broad, then narrow. Give the assistant the shape of what you want, then tighten constraints once you see the availability estimate.
- Let it resolve the codes. You don’t need exact OncoTree or ICD-10 codes — describe the indication in normal terms and confirm the match it suggests.
- Use cohorts for comparisons. If you need a case and a control arm, ask for two cohorts in one request rather than two requests; they are checked and delivered together.
- One request per need. Keep unrelated cohorts in separate requests so each has its own specification and reference number.
- Explore first when you can. If the data might already exist in a connected dataset, try Data Insight before requesting; the availability check is a good bridge between the two.