Registries: dictionaries, consent, sites, import and API, exports for the statistician and 50 000 patients

YouEDC Capture

• Answer lists of a study: a list is written once and used by any number of fields, and changing it changes them all; ready-made lists (sex, yes / no / unknown, severity, ECOG, smoking, federal district); dependent lists (district → region); a “From a dictionary” field with the Ministry of Health МКБ-10 uploaded from НСИ: the investigator finds a diagnosis by code or words, the code is saved Answer lists and dictionaries
• Outcomes of participation: the “Lost to follow-up” and “Died” statuses, each status with the date it took effect (in the card, the certified casebook and the export); a log of attempts to reach the patient; moving a patient to another site, keeping the code and the data, with the history of moves Patients et matrice des visites
• Informed consent: versions of the consent form with the approval date and what they allow (linking with medical records, recontact, secondary use), each patient's consent, a “re-consent needed” list, a basis without consent for retrospective studies, data entry only after consent, two kinds of withdrawal — withdrawing the use of the data takes the patient out of every export Consentement éclairé
• Data cuts: the whole study's data frozen at one moment for an interim analysis or a registry report, with a SHA-256 fingerprint and an integrity check; a cut downloads as it is or de-identified (pseudonyms S0001, dates shifted within a patient, age instead of the birth year, no free text and no queries); a new Statistician role sees de-identified cuts only; a cut holds the Consents sheet; a patient who withdraws the use of their data after a cut leaves every later download of it, with a note in the file and the audit trail, while the cut and its SHA-256 stay as they were; a patient's pseudonym is the same in every cut Data cuts and the de-identified export · Centres et équipe · Consentement éclairé
• A form library in the builder: nine ready-made forms (demographics, diagnosis, vital signs, therapy, AE, questionnaire, outcome, source, end of participation), a copy of a form of this or another study, the whole eCRF of another study with visits, conditions and answer lists (in place of the current one while there are no patients); eCRF from Excel (a fields sheet and a visits sheet) and back to Excel Form library, copying and eCRF from Excel · Constructeur d’eCRF
• AND / OR conditions and checks across forms: a field shows, a check fires and a visit is expected only when conditions on fields of this and other forms and the patient's card hold; a check compares with a field of another form (“the AE start date is before screening”); values of other forms to try in the preview; conditions in Excel, the DVP and the investigator guide AND / OR conditions and checks across forms · Contrôles de saisie · Constructeur d’eCRF
• An export builder for the statistician: the forms and variables you need from the live data or a cut, the long layout (a row per visit) or the wide one (a row per patient, SBP_V1, SBP_V2), SPSS, Stata, SAS (XPT with a formats program), R (CSV with a script), Excel and CSV; variable and answer labels, numbers and dates already in the file, several answers make a 0 / 1 variable per answer Export builder and formats for statistics · Export de données · Data cuts and the de-identified export
• A data quality report: completeness of each variable by site, taking show conditions and “no data” marks into account, outliers by the modified z-score with a link to the page, the entry delay from the visit to the first save, loss to follow-up and patients at risk of it; the same in Excel Data quality report
• Registries of tens of thousands of patients: the pages are checked on a copy of the demo registry with 50 000 patients and 300 sites; the lists of patients, review and signatures show 100 rows a page with filters over the whole study; in a study of more than 3 000 patients the summaries of the overview, the site risk and the quality report are kept for 10 minutes with “Counted at” and “Count again”, and the calendar shows one site at a time; the Excel and CSV export is several times faster, and the audit trail CSV starts downloading at once Vue d’ensemble et paramètres de l’étude · Patients et matrice des visites · Événements, frise du patient et calendrier des visites · Signature électronique de l'investigateur · Export de données
• A Sites page in each study: the site's status (being activated, enrolling, enrolment closed), PI, enrolment target and team; inviting people straight to a site; Team is a menu item of its own Centres et équipe · Site directory and organisation
• Sites of a study from Excel: a row per site with the PI and their e-mail, the site is found in the directory or gets a card; the PI gets an invitation and invites the investigators and coordinators of their site (the My site item), and the manager can make any investigator a site PI Centres et équipe
• Whole patients from one Excel workbook: the patients, their consents and every form of every visit load in the background, through the same checks as typing; the layout is the export's, so an exported file loads back; the file is shown before loading; the audit trail names the person who loaded it; the one-form import also refuses the data of a patient without a consent that stands Import de données depuis Excel et CSV
• A write API for hospital and laboratory systems: the study manager gives each system a key (sites, the right to enrol, a term), the system reads the study design, enrols patients and sends visit forms in batches of up to 500; every value goes through the same checks, consent gate and audit trail, which names the key; a key can be revoked, and stops when its manager leaves API for hospital, laboratory and sponsor systems
• Possible duplicate patients: enrolment can take the date of birth and the initials, the portal keeps only their fingerprint with the sex and finds a patient enrolled twice, at one site or at two; Cyrillic and Latin initials match; the data manager decides each pair, and the decision goes to the audit trail; the check can also be made later on the patient card, in the whole-patient workbook and through the API Patients et matrice des visites
• CDISC variable names in the eCRF builder: under the field code a CDASH-style name appears, from the question, the field type and the form (“Start date” on an AE form → AESTDAT, “Systolic BP” → SYSBP), with where it goes in SDTM; one click takes it, and an empty code gets it by itself. Warnings when a name belongs to another domain, does not fit the field type or is filled in SDTM automatically. These are hints: any code can be saved Constructeur d’eCRF