Data Integrity – Approach in Clinical Research
Bodo Lutz,
Data Integrity Officer,
Novartis
Investigations conducted by Health Authorities in the last years uncovered several serious data integrity issues and many companies have received 483s and warning letters related to data integrity. Examples include analytical tests which were repeatedly performed until achieving desired results, computer systems that allow deletion and change of raw data, and back-dating of documentation.
Data integrity is a critical component of any clinical study as interpretation of study results is only possible when all data is trustworthy, reliable, consistent and accurate. There is also an ethical obligation that the involvement of human study participants requires that study data meets the highest standards. Mishandling of data can jeopardize the efficacy of medicinal products and safety of patients.
In the pharmaceutical and medical device industry the principals of data integrity and quality are important in ensuring compliance with international standards of Good Clinical Practice. ICH E6 (GCP guideline for medicinal products) and ISO 14155:2012-01 (GCP guideline for medical devices) refer in several sections to quality throughout the data lifecycle, from getting them collected, monitored, transmitted, transformed, processed, analyzed, reported, backed up, and finally archived either as carbon copies or in electronic version.
Regulators and Notified Bodies expect that study data handling follows ALCOA, the acronym for attributable, legible, contemporaneous, original, and accurate, as study data are used as basis for regulatory decisions. In addition study data should always be complete, available, consistent and enduring.
This presentation will be focused on examples on how data quality can be improved by strengthening culture, processes and systems related to data integrity within the Sponsors, CROs and Investigational Sites, and by establishing (Quality by Design) and maintaining adequate manual quality control, automated control steps, verification and validation
|
