Study design
Record who the study is from, its ethics and data rules, and the method, and confirm each part Review & launch will hold.
Who can open this: Principal investigator, Steering committee (full access), Biostatistics, Coordinator
Who this is for
Study design is the first Setup group. Its sidebar label is "Study design", and its line is "Define who the study is from, its ethics approval and data governance, and the method that governs its rounds." The tabs are the pages: "Study identity", "Ethics and data governance", and "Method".
Only the Principal investigator can edit Setup. Steering committee (full access), Coordinator, and Biostatistics can open it and see "You can read Setup but not change it. The principal investigator edits these steps; the fields below are shown as they are saved." A save from anyone else is refused with "You do not have permission to change this." Coder, Adjudicator, and Steering confirmer do not get the Setup section. A workspace administrator who is not on the study team sees "You do not have access", "You are not on the team of “{study}”", and "Only people on a study's team can open its pages. Ask its principal investigator to add you."
Waiving required ethics documents is the Principal investigator only. The page says "Only a principal investigator can waive your organization's requirement." The server says "Only a principal investigator can waive the organization's required documents."
Steps
Work the three tabs, then confirm the pages that ask for a confirmation. Review & launch counts Identity (Study identity and Ethics and data governance together) and Method among the ten steps. An error on either step is "Blocks launch". A warning that still leaves the step able to launch is "Needs attention", and that step still counts toward the ten. See Review and launch.
Table: which Setup screen counts which steps. Review & launch sits outside the four groups.
| Screen | Counted steps |
|---|---|
| Study design | Identity (Study identity and Ethics and data governance together), and Method |
| Panel design | Eligibility, Size & mix, and Profile & integrity |
| Round design | Item bank, Rating & consensus, and Rounds & feedback |
| Governance | Consent & comms, and Analysis plan |
| Review & launch | None. It is the destination, not one of the ten. |
Study identity (Setup, Study design). The frame is "Study identity": "These details appear in emails to panelists." The fields are "Research team name", "Affiliation", "Organization name", "Study name", "Principal investigator", "Reply-to address", and "Protocol label". Confirm with "Confirm the study identity". "Withdraw confirmation" clears it. Launch waits on "Confirm the study identity before launch.", on a blank header or footer ("1 email header or footer field is empty.", or "2 email header or footer fields are empty."), and on "Enter a reply-to address panelists can write to, like team@example.org."
Two notes are advice and do not hold Ready or launch: "Emails will carry the Delphi Studio logo until you upload one or use your organization's." and "No protocol label: emails print no protocol line. Add one if the study has a versioned protocol." A blank protocol label still holds the first real invitation; that check is "Protocol label" under Ready to send on Review & launch.
"Pseudonym code" has the field "Code" and the button "Save code". Staff screens, team emails, and exports show each panelist as this code and their number. Panelists never see it. "It can change until the study launches." After launch the card says "The study has launched, so the code is fixed."
After launch, the identity confirmation is what locks. The server refuses a change to it with "The identity confirmation is locked: the study has launched. Changes need a controlled amendment." The other identity fields stay editable, and the edit is logged.
Ethics and data governance (Setup, Study design). The frame is "Ethics and data governance": "Record approval, attach the documents, and confirm how panelists' data is kept and removed." The sections are "Ethics review" ("Which ethics review the study falls under appears in every email footer."), "Approvals and documents", "Withdrawal and removal terms" ("Three questions. The answers are the terms a panelist agrees to."), and "Data retention" ("How long identifiable data is kept, and whether a panelist can ask for removal.").
"Approvals and documents" says "Attach each document the study needs. A file is kept by name, size and date until file storage arrives." A row reads "Attached", "Needed", or "Not attached".
Launch is blocked until the ethics review is chosen and its facts are present: "Choose the study's ethics review: approved, exempt or not required. It appears in every email footer." An approved study also needs "Add the IRB number. It appears in every email footer." and "Add the IRB contact email. Panelists send ethics questions there." An exempt study needs who determined it and when. A study that needs no review needs why. A missing required document blocks launch ("1 required document is missing: {name}.", or "2 required documents are missing: {names}."), unless the Principal investigator uses "Launch without required documents" and confirms "Launch without required documents?". The waiver turns the missing documents into a warning ("The requirement is waived, so launch is allowed without it."). It never excuses an expired approval ("The IRB approval expired 1 day ago.", or "The IRB approval expired 2 days ago."). Retention above the organization's limit blocks launch. "Confirm the data retention policy." and "Confirm the withdrawal terms." need attention and still allow launch. The consent form cannot be published until those terms are complete, confirmed, and still match the retention policy.
After launch, the server locks "The retention policy", "The withdrawal terms", "The protocol document", and "The document waiver", each with "is locked: the study has launched. Changes need a controlled amendment."
Method (Setup, Study design). The frame is "Method": "Define what the study is for, which method governs it, and how structural questions are judged." It is a walk of six steps, labeled "Objective", "Method", "Classification", "Governance", "Decision rules", and "Review", shown as "Step N of 6". On Review, "Confirm the method" records the confirmation. Launch waits on "State what the study should achieve.", "Choose a method framework.", "Add at least one rating round in Rounds & feedback.", and "Confirm the method before launch." A later edit asks "The method changed after it was confirmed. Confirm it again."
A single-round plan, gaps in a framework's dimensions, and a group's target that cannot meet a check are warnings: the step can say "Needs attention" and still count as ready. "More than 4 rounds is unusual and risks panel fatigue. This is advice only: it does not hold up launch." The round list itself allows at most 10 rounds.
After launch the page says "The method is locked: the study has launched. Changes need a controlled amendment." Later changes that do exist are the recorded amendments named on Review & launch under "What launching locks", each with a reason. There is no generic amendment for the scale, the zones, the protocol, or RAND/UCLA.
A save that lost a race is refused and never merged: "This section was changed somewhere else since you opened it. Reload to see the latest version." A confirmation is refused with "A confirmation can only be made by the person saving it.", "A confirmation must carry today's date.", or "The confirmation does not match the content being saved."
What the app will not do
A default logo, and a protocol label left blank, do not stop launch. The first real invitation still waits until "Protocol label" is set.
"Launch without required documents" does not excuse an expired approval, and it does not waive retention above the organization's limit.
The documents card keeps a file's name, size, and date until file storage arrives. It does not store the file.
Launch locks the identity confirmation, the retention policy, the withdrawal terms, the protocol document, the document waiver, and the method. It leaves the other Study identity fields editable.
Related
Checked against the app on 2026-10-31.