A study set up from its protocol

Upload the protocol: the study, its eCRF draft, profile and QA environment are set up, and the log says what to check.

Setting a study up by hand from a 120-page protocol takes days: the visits and windows from the schedule of activities, a form for each procedure, the study profile, the QA environment. YouEDC Capture can do the first pass itself: upload the protocol, and the study is set up from it. What it could not do, or did on an assumption, it says in the auto-setup log, row by row, with the protocol’s own words and page. Nothing reaches a site before a person checks the log and releases version 1 of the eCRF.

The protocol is read once into a study model: facts with their quotes and pages. Each part of the set-up is built from it, and the log collects what to check before the release.

Where to start

Open New study and follow Set up from the protocol. Choose the protocol file (DOCX, or a PDF with text, not a scan), give the study a name and a code if you want others than the protocol’s title and number, and press Set the study up. Anyone who may create studies on the portal can do it; the study counts towards their projects like one made in the wizard.

The page then shows the progress of the set-up: the upload of the file, then the steps one by one (reading the protocol page by page, finding the schedule of activities, creating the study and its eCRF draft, the QA environment, the log), and opens the log when the study is ready. A protocol of a hundred pages takes a minute or two. You can leave the page: the study appears in your studies, and the job is under My background jobs. If the file cannot be read, the page says why and nothing is created. One set-up runs at a time per person: pressing the button again opens the progress of the one already running.

A draft reviewed in Protocol to study build (YouEDC Studio) gives the same result: its Create a study button sets the study up and writes the same log.

What is set up

• The study in set-up, its code from the protocol number and its name from the title, the sponsor, the type (trial, observational, registry).
• The eCRF draft: visits with their days and windows from the schedule of activities, the forms of each visit from the form library, the logs (adverse events, medications) in one visit. Version 1 is not released.
• The study profile: phase, design, blinding, control, number of groups and the ratio, planned patients, ages, treatment and follow-up, dictionaries.
• The QA environment with the draft published to it, when the portal has QA environments: you can try the eCRF on test patients at once.
• The protocol file in the inbox of the Trial Master File, when the TMF is on.

The log

The log is in the eCRF builder (the line “Set up from the protocol” at the top) and on its own page. Each row has a code (A-03: the area and a number), a status, what was done, the quote of the protocol and where it is (section and page). Statuses:

• Done: read unambiguously and set up.
• To check: set up on an assumption (a visit without a day, a visit counted from an event, a new form no library form fits, a patient questionnaire).
• Contradiction: the protocol says different things in different places, for example the age range in the synopsis and in the inclusion criteria. The set-up used the first one; ask the sponsor.
• Not set up: to do by hand, and the row says what and where (eligibility criteria, randomisation, the SAE deadline, laboratory panels, the team and sites).
• Licence needed: MedDRA, WHODrug and licensed questionnaires. The portal never brings them: your organisation loads them under its own licence; a licensed questionnaire is set up as an empty form.

The log of the training study YE-330-302, rows waiting for a decision: the protocol gives the age range as 18 to 75 in the synopsis and 18 to 70 in the inclusion criteria, and the treatment period as 26 and 24 weeks.

Decisions and the release

Rows “to check” and “contradiction” hold the release of the eCRF. The project administrator decides each one: Accepted as is, Corrected or Rejected, with a comment (a correction or a rejection needs one). Rows “not set up” and “licence needed” are marked Done by hand or Not needed when you get to them; they do not hold the release. Every decision and every undo goes to the audit trail. The log also downloads as Excel for the study file.

While a row waits for a decision, the eCRF versions page says so and does not release the version. In a hurry, tick Release without deciding them and say why in the reason: the release then leaves a line in the audit trail with the number of undecided rows.

The versions page holds the release of version 1 while 7 rows wait for a decision. “Release without deciding them” lets it through with the reason in the audit trail.

Try it on the training protocol

The page has the button Set up the training study: YE-330-302, a made-up phase III study in type 2 diabetes written for YouEDC (the drug and the sponsor do not exist). It has 10 visits, a run-in period, a phone visit, laboratory panels, two questionnaires and two contradictions left in on purpose: the age range and the treatment period. The study is your demo with its QA environment: it does not count towards your projects, and it is deleted in one click with everything in it. One training study per person; delete it to set it up again.

How the protocol is read

No AI. Rules read the document: the headings and their numbers (ICH M11 or the usual 1, 2, 5.1…), the sections they open (synopsis, schedule of activities, inclusion and exclusion criteria, randomisation, safety), the rows of the synopsis table and the schedule of activities. Each fact keeps its quote, section and page, and the facts of different sections are compared. The same file always gives the same set-up: the log keeps the SHA-256 of the file and the version of the rules, so an inspector can trace every row back to a page of the protocol.

A protocol is read best when its headings are numbered and the schedule of activities is a table. A scanned PDF has no text to read: upload the DOCX.

A protocol amendment in a running study

When the protocol is amended after the eCRF has been released, open eCRF versions → Protocol amendment (the project administrator and the data manager). Upload the amended protocol, or pick it from the study’s Trial Master File, and give it a label such as “Amendment 1”. It is read with the same study model in the background and compared with the protocol the study was set up from (or with the previous amendment). A study built by hand is compared with the eCRF version in force: then a visit or form named differently in the eCRF shows as a change, and you reject such rows. The same file uploaded twice opens the same amendment.

Each difference is a row in the style of the log, with the old and the new words of the protocol and their pages:

• Into the draft: a new visit with its day, window and forms; a visit no longer in the schedule; a visit moved or its window changed; a form added to a visit or taken off it; a new form from the library for a new procedure. The row says how many patients it touches (for example, those already past the window of a new visit).
• Study profile: planned patients and sites, countries, ages, treatment and follow-up.
• By hand: what the portal does not change itself, and the row says why: the eligibility criteria and their checks, randomisation, blinding and groups, the SAE deadline, how a visit is done (phone or in person), a visit counted from an event, and a form patients already have at a visit it leaves.

The data manager accepts or rejects each row; a rejection needs a comment. Then the project administrator presses Write into the eCRF draft: the accepted rows go into the draft of the next version (in the QA environment when the study has one) and the study profile. The version in force does not change, and nothing is deleted: a visit the amendment removes is deactivated and its data stays, a form with data at a visit stays in it. After that, run the draft over the working data and release it on the eCRF versions page: the reason of the release is filled in with the amendment’s number and date, and the release marks the amendment released. When the study keeps informed consent versions and the amendment changes what patients do, the page reminds you that a new consent version may be needed; the consent is never changed automatically. The upload, every decision, the writing and the release are in the audit trail; the rows download as Excel.

Amendment 2 of the demo study, written into the eCRF draft: a new visit with the urinalysis, the window of the last visit widened, the larger sample size rejected with the sponsor’s reason.

To try it, the training protocol has its amendment: version 2.0 of YE-330-302 adds visit V8.1 and the urinalysis, widens the window of V10 and plans 270 patients. Set up the training study, release version 1, and on the amendments page press Read the amendment of the training protocol.

What it does not do yet

• Eligibility criteria are counted, not yet written into the eligibility form one by one.
• Randomisation, YouEDC Safety, laboratory tables and the ePRO schedule are named in the log with their quotes and set up by hand.

Accept all. The button “Accept all rows to check” above the log accepts as is every row waiting for a check. It leaves the contradictions alone: you decide them one by one. Each decision is in the audit trail.