Your Easy Guide to Clinical Studies

Raw Data (RD) is considered the original data derived from some primary source (e.g. laboratory reports, participant questionnaires, medical examinations).

 

RD has not undergone any processing, either manually or through automated processing.

 

It is critical to clearly define any RD used in a study. Aspect to include are:

• Description of the type of RD (e.g. weight = continuous data, gender = coded data, questionnaire = score data)
• The source (Source Data) from where RD was retrieved (e.g. medical records, laboratory reports, paper CRF)

 

Sometimes RD must be processed in order to convert RD into a format that can be analysed and visualised. This usually entails some type of cleaning (harmonisation) or transformation. Any such data processing procedures are described and documented in the metadata of the study.

More

As a SP-INV, familiarise yourself with how to specify and document study data:

• List variables needed for study evaluation
• Define RD and its source (e.g. participant weight is located on paper-CRF from medical examinations at screening)
• In the event of derived or processed data, indicate formula used (e.g. computation of pain score based on participant questionnaire)
• Define data format and prioritise standardised or international formats
• Define quality checks needed to locate computation errors
• Document implemented processing procedures in the metadata of the study

 

Examples of data formatting

 

• Date formats: 31st January 1999, 31/01/1999 or 31.1.99 99.01.31, 31011999, or today
• Computation:
  • metrics: 1,70 m or 170 cm
  • time: 1,15 min or 75 sec
• Coding: female=F male=M, or female=1 male=2

 

By comparing entries in the electronic database (eCRF) with the original RD, a monitor can confirm during monitoring that the respective data entered in the database is correct.

More