The clinical study database concept - How can I picture that?
What is a clinical study database? A clinical study database contains clinical study data and metadata about the clinical data. Study data is initially saved on hardware like a server or a hard disc. The data is stored in a database language which is usually the server’s operating system language like SQL or UNIX. The data in the clinical study database is structured by the Clinical Data Manager who utilizes Clinical Data Management software to do this. The readable format of the clinical study database, ready for (statistical) analysis, is stored in electronic files. So-called datasets, in SAS-, SPSS or MsExcel format for example.
How can I picture myself a clinical study database?
Think of the sight of a stately oak in summer… A clinical study database compared to a stately oak in summertime…
The acorns represent specific data field choices like MALE/FEMALE for the gender field, YES/NO for a yes/no question and MILD/MODERATE/SEVERE for the adverse event severity variable.
The leaves visualize the CRF fields. E.g. the field date of birth (BRTHDTC), subject initials (SUBJINIT) and adverse event description (AEDESC).
The branches show the grouping of collected data to the required datasets. Like all demography data, e.g. date of birth, subject init and gender, in a demography dataset (DM.sas). All adverse events with start- and stop dates, severity, outcome and so on in the adverse event dataset (AE.sas). And all study medication dosing data in a medication dosing dataset (MDDOS.sas).
The trunk of the tree represents the clinical study database. All datasets that together contain all study data collected.
Out of sight, the tree’s roots represent your choice of a structured format for your data.
You are free to structure your data anyway you like. As it suits the person going to analyze them to be able to draw interpretations and conclusions for the Clinical Study Report. The pharmaceutical industry took the initiative for a worldwide standardization of naming conventions and data structure. Which is carried out by the Clinical Data Interchange Standards Consortium (www.cdisc.org). The FDA’s tools for analysis are able to analyse datasets adhering to CDISC standards. Datasets structured according to CDISC shorten the market application time needed by the FDA.
A stately oak in summer... From the tiny acorns variable details to the big trunk database. A clear and lively picture of what a clinical study database actually stands for.
© 2010 ProCDM, Maritza Witteveen
Interested? Subscribe to ProCDM's ezine with clinical research tips and receive the e-book "Five simple strategies to get reliable quality clinical data".
Zeist