After logging in on TrialKit Coder website directly or accessing it via the EDC menu option, a screen similar to what's shown below will be displayed. This is where all existing coding projects can be accessed, or additional coding projects can be created. 

Access to the various functions throughout the system are permission dependent and may not be visible/accessible to all users. 


See the annotated image and descriptions below.




1. Table of coding projects - This lists the current coding projects which have been created. Each project is tied to a specific dataset which can be updated over time. 


2. Add New Project - Use this to create a new coding project. This only needs to be done once for each datasource being coded. If updating an existing datasource, the upload option will be used with the same data source name (file name).


3. Upload CSV - Use this to upload data source files to code. After uploading, a new data source, a project must be created to perform the coding on it. Read more at the bottom of this article.

Note, This option only displays for standalone coding. Live TrialKit EDC studies will not have this option, since the data source comes in automatically from EDC.


4. Code This - Opens the corresponding project for coding. Read more here.


5. Dashboard - Reports on metrics regarding the coding project. Read more here.


6. Edit/Delete - Edit the name or workflow levels of the coding project, or Delete the entire dataset. WARNING: Deleting will delete the coding project and all underlying data. This will not be available for coding projects tied to a live TrialKit EDC study. 


7. User preferences (change password and logout) and options to switch to a different study or database.



To Add a new project (item 2 above):



Configuration Name - Name the project. If doing standalone CSV importing (described below), be sure to first upload the file before doing this step. If using the mobile app, be sure the name of the configuration matches the name of the CSV file, without the file extension on the end. 


Medical Coding Type - Select which dictionary will be used for the dataset being coded. The options here depend on which dictionaries have been enabled. Contact your account representative with any questions about this.


Review Levels - Define which user roles need to perform review on coded items. This step is optional and can be left blank if workflow is not needed. Provide both the applicable role and a name for the level of review. Up to two review levels can be defined, but none are necessary.

Data Source - Choose a data source for the coding project. This will be either a form from TrialKit EDC or an imported data file (described below).




Standalone File Importing (not applicable to studies being conducted in TrialKit EDC)


To prepare a CSV file for importing and coding the terms, a few minor requirements exist. Any number of additional columns (data points) can be imported to assist with the coding process, but there are basic columns needed for the auto-coder to work and for the system to be able to differentiate records when adding new ones or updating records over time.


The following columns are needed, at minimum, for event coding (additional columns of data can also be included):


Column Name
Type
Description
PatientIdentity
String
Used to Identify an individual Patient
Date
String
Used to Create a Tx Starting Date and Visit Date
AETERM
String
MedDRA Search Term (MedDRA Only)
RecKey
String
Individual Record Key, used to insert new records or update existing records


The following columns are needed for medication coding:


Column Name

Type
Description
PatientIdentity
String
Used to Identify an individual Patient
Date
String
Used to Create a Tx Starting Date and Visit Date
CMINDC
String
WHO Indication Term (WHO Only)
CMROUTE
String
WHO Route Term (WHO Only)
CMTRT
String
WHO Search Term (WHO Only)
RecKey
String
Individual Record Key, used to insert new records or update existing records



Any other column names and data can be included and will come out in medical coding exports as well. The dataset must not have any duplicate columns.


Important: RecKey is a unique identifier for each record. This must be unique to prevent any overwriting of records. This is also what makes updating coding lists a possibility. By adding new lines to a file, or updating existing items from a previous import, the RecKey column is how the system knows to update an item or create a new one.

Important notes:

  • When updating an existing coding list, the data source file must have the same name as the original.
  • When existing records' verbatim term changes in a source file that is being updated, the system will automatically clear any coding data that may have existed on those records. Records where the verbatim term did not change will not be affected by re-importing source data.




Upload the file containing columns described above in the table, depending on if coding to MedDRA or WHO Drug.

Once The Upload button is selected, a message will display indicating successful upload. The window can then be closed.




Next, create a coding project to correspond to the new data set just uploaded, as described at the top of this article.




The required fields are completed below. The data source should be the name of the data set file uploaded above.




After saving, the table will populate with the new coding project, which can now be coded.



To import a file (Updating a data file):


Once a coding project is created and partially coded, it can be updated by simply updating the file again via the Upload button. The file name must match the original. This is how the system knows which file to update on existing projects.

New records (rows) can be added to the file and existing items can be changed. The system uses the record key column (described in the table above) to determine if it should update an existing record or add a new one.

Existing coding will not be affected unless the verbatim term changes. In that case, the coding details no longer apply and are cleared.




Continue reading how to code a dataset..