How to set up a database?-a five-step process
- PMID: 30510791
- PMCID: PMC6230828
- DOI: 10.21037/jtd.2018.09.138
How to set up a database?-a five-step process
Abstract
Database set-up directly impacts the quality and viability of research data, and therefore is a crucial part of the quality of clinical research. Setting up a quality database implies following a strict data-management process. Too much collected information threatens the quality of the information required to achieve the objectives of the study. Therefore, the data that will be collected and managed have to be cautiously discussed and selected. Case report forms (CRF) are the tools the most frequently used to collect the data specified by the protocol. An informative and well-structured document simplifies database design and data validation. Key elements are about choice of sequential or thematic structuring, information and type of information that should be entered and the importance of data standards and coding guide. Final database must be structured with unique ID patient, with one record per subject or per measure. Specific information must be provided for each variable according to the database specifications. The quality of the results is directly related to the quality of the collected data. The CRF should then be completed as fully and accurately as possible. Data validation relies on three key points: the CRF completion guidelines, the Edit Checks process and the Data clarification process. Various open source or business software applications provide all functionalities to set up a clinical data base and CRF. The General Data Protection Regulation (GDPR) standardizes and strengthens the protection of personal data across the EU and for other country's data being "processed" within the EU. The General principles include lawfulness, fairness and transparency, restricted use of data, data minimization, accuracy, limited storage, confidentiality and probity, and accountability.
Keywords: Database; General Data Protection Regulation (GDPR); software.
Conflict of interest statement
Conflicts of Interest: The authors have no conflicts of interest to declare.
Figures
References
-
- Everyone needs a data-management plan. Nature 2018 15;555:286. - PubMed
-
- Househ MS, Aldosari B, Alanazi A, et al. Big Data, Big Problems: A Healthcare Perspective. Stud Health Technol Inform 2017;238:36-9. - PubMed
-
- Kelsey JL, Whittemore AS, Evans AS, et al. Methods in Observational Epidemiology. Second Edition. Oxford, New York: Oxford University Press; 1996:448 p. (Monographs in Epidemiology and Biostatistics).
Publication types
LinkOut - more resources
Full Text Sources