Data export

Excel or CSV for analysis, with the dictionary and queries; export history.

Export gives you all study data for the analysis: Excel or CSV together with the dictionary of variables and the queries, and CDISC ODM for other systems. The file is laid out as in YouEDC 0.2, so existing analysis templates and scripts keep working.

How to export

Open “Export” in the study menu, choose the format and press “Download”. “Answer options as text instead of codes” replaces the answer codes (1, 2) with their text (“Mild”, “Moderate”). That is easier to read by eye; for statistics people usually take the codes.

What is inside the file

The file is written as the data is read, so even a registry of 50 000 patients downloads in seconds rather than minutes; text in Excel always stays text, a value starting with “=” never becomes a formula.

• Excel. A “Patients” sheet and one sheet per form, named with the form code. Then come the “Dictionary” of variables with the answer codes, and “Queries”.
• CSV (zip). It holds subjects.csv, one file per form (VS.csv, LB.csv…), codebook.csv, queries.csv and README.txt. The separator is “;” and the encoding UTF-8. It suits SPSS, R and Python.

One row of a form sheet describes one form of one visit. Its columns are SUBJECT (patient), SITE, VISIT, INSTANCE (event number), STATUS (draft, complete, verified), SOURCE (web, epro, import), UPDATED_AT, then the values by variable code.

• “No data” marks are exported as ND, UNK or NA.
• When a “no data” mark has a reason, the export gets a No data marks sheet (missing.csv): patient, site, form, visit, instance, variable, the mark and REASON. The Events sheet of a study whose unscheduled visit asks why the patient came has VISIT_REASON and VISIT_REASON_TEXT; the dictionary lists both codes. A de-identified cut keeps the codes and drops VISIT_REASON_TEXT, which is free text.
• Values are exported as they are stored: numbers with a dot (7.4), dates as 2026-09-28, times in UTC.
• In Excel, number fields are written as numbers, so sums, filters and sorting work straight away. Answer codes, dates and ND marks stay text. CSV writes everything as text, as YouEDC 0.2 did.
• A value that starts with “=” is written as text, so Excel does not take it for a formula.

Tables in the export

Each table in a form gets its own sheet (its own file in the CSV), named by the form's and the table's codes: VS_LB_CHEM. This is the long layout, like the CDISC LB domain: one line per row of the table in each form. First come the same columns as on the form's sheet (patient, site, visit, status), then ROW (the row's code: GLUC, or the number 1, 2, 3 of a growing table) and ROW_LABEL (the row's name). When the rows have their own units and normal ranges, UNIT, NORM_LO and NORM_HI follow, and then the values of the columns.

• A grid's values are also on the form's sheet, a column per cell: LBORRES_GLUC, LBORRES_CREAT. So the whole visit stays on one line.
• A result “<5” is exported as it is, as text.
• A deleted row of a growing table is not exported. Its values and the reason it was deleted stay in the audit trail and in the casebook.
• In the export builder a table's sheet comes in the long layout. For the wide layout choose the form: a grid's cells are already columns there.
• In ODM a table is a repeating group (ItemGroup Repeating="Yes"); the repeat key is the row's code or number.

CDISC ODM 1.3.2

ODM is the CDISC standard format for exchanging clinical data between systems. One XML file holds both the study design and the patients’ data, so another EDC, CDISC tools and YouEDC Studio can read it. You do not need to check the file before you send it: it is built to pass the official ODM 1.3.2 schema.

• A visit becomes an event (StudyEventDef). An unscheduled visit repeats, and the event number is written as StudyEventRepeatKey.
• A form becomes a FormDef with one group of fields. A field becomes an ItemDef: the name is the variable code, the question is the field label.
• A number is written as float, a date as date, text as text. Single choice and yes / no fields get a list of codes with labels (CodeList). Multiple choice is written as codes separated by commas, as in Excel.
• A whole number and the scales are written as integer, a time as time, a date and time as datetime, a partial date as partialDate (2024-03). A drop-down list and a checkbox get a CodeList. The formula of a calculated field and the format of a text go into the variable's description; a text for the investigator is not exported.
• Hard and soft ranges become RangeCheck marked Hard or Soft, and units become MeasurementUnit.
• ODM 1.3.2 has no “no data” marks. So a field that allows them gets a companion field next to it, for example WEIGHT_ND with the value ND, UNK or NA.
• The first event, “Patient card”, holds the patient’s sex, year of birth, enrolment date and status. Sites become Locations, and every patient carries their site.
• ODM codes are made from the form and field codes. Visit names in other alphabets are written in Latin letters: «Визит 6 мес» becomes SE.Vizit_6_mes.

Form statuses, SDV, queries and the “if – then” rules are not part of ODM: they stay in the Excel export and in YouEDC itself. The “answers as text” box does not apply to ODM, because ODM always holds the codes together with their labels.

Here is a part of the file of the demo study REG-DIAB-01: the systolic pressure field with its unit and ranges, and below it the first adverse event of patient REGDIA-01-0002 with severity 3.

<ItemDef OID="IT.VS.SBP" Name="SBP" DataType="float" SASFieldName="SBP">
  <Question><TranslatedText>Systolic BP</TranslatedText></Question>
  <MeasurementUnitRef MeasurementUnitOID="MU.mmHg"/>
  <RangeCheck Comparator="GE" SoftHard="Hard"><CheckValue>50</CheckValue></RangeCheck>
  <RangeCheck Comparator="LE" SoftHard="Hard"><CheckValue>300</CheckValue></RangeCheck>
  <RangeCheck Comparator="GE" SoftHard="Soft"><CheckValue>90</CheckValue></RangeCheck>
  <RangeCheck Comparator="LE" SoftHard="Soft"><CheckValue>180</CheckValue></RangeCheck>
</ItemDef>
…
<SubjectData SubjectKey="REGDIA-01-0002">
  <SiteRef LocationOID="LOC.01"/>
  <StudyEventData StudyEventOID="SE.Adverse_event" StudyEventRepeatKey="1">
    <FormData FormOID="F.AE">
      <ItemGroupData ItemGroupOID="IG.AE">
        <ItemData ItemOID="IT.AE.AE_TERM" Value="Dizziness"/>
        <ItemData ItemOID="IT.AE.AE_SEV" Value="3"/>

Example: registry data for the statistician

The data manager of the demo study REG-DIAB-01 exports Excel with answer codes. The file has 8 sheets: Patients, DX, VS, LB, PRO, AE, Dictionary, Queries (with the Russian interface the sheets are called «Пациенты», «Словарь», «Запросы», as in 0.2).

• On the LB sheet, the row REGDIA-01-0004 · 01 · Month 6 · 1 · complete · import with HbA1c 7.4 came from the laboratory file. The article on import tells how.
• In the Dictionary, the variable HBA1C has the type number, the unit %, the required mark and the allowed range 3–20.
• The Queries sheet holds the automatic query on the HbA1c of 11.8% for patient REGDIA-02-0004.

Statisticians usually prefer CSV. In R a form sheet reads as read.csv("LB.csv", sep = ";", fileEncoding = "UTF-8-BOM"). In Python it reads as pandas.read_csv("LB.csv", sep=";", encoding="utf-8-sig").

Only some forms and variables, a row per patient, or files for SPSS, Stata, SAS and R with labels are made by the export builder.

For an analysis that must be reproducible, or for a statistician outside the team, export from a frozen data cut: it never changes, carries its SHA-256 and downloads de-identified.

Who can export, and what the audit trail keeps

• The study manager and data managers can export. A data manager with access to only some sites gets only the patients of those sites, and the page says so.
• Every export is written to the audit trail: the format, the number of forms and patients, “answers as text”, the sites.
• The bottom of the page shows the export history: when, who, format, how many patients, file name. Exports made in YouEDC 0.2 are marked “YouEDC 0.2”.
• The Portal Owner can limit the number of exports per day.

Database lock

Before the analysis the database is locked. Data entry and changes close for everyone, and export stays open. The checklist on the export page shows what is not finished yet: open queries, forms in draft, forms without SDV. It is a hint, not a ban.

The study manager locks the database in the study settings with “Lock the database”, giving a reason and their password. The database can be unlocked again, also with a reason and the password, for example when enrolment continues after an interim analysis.