Data integrity and data quality are critical success factors for artificial intelligence (AI) and machine learning (ML) solutions in life sciences. Simply performing computer system validation or managing computer systems under CGMP conditions is not enough to ensure data integrity and data quality for data sets where AI/ML is intended to be applied.
Despite the best efforts of those responsible for data integrity, the potential for human error is directly and indirectly impacted by the corporate, the national/regional , and quality culture of an organization.
For successful establishment and sustainability of a quality culture, “the mindset and behavior... must start at the top and be emulated by individuals at all levels and in all functions within the company.”
Data integrity is of paramount importance to ensure patient health and safety and to improve shareholder value, particularly for virtual companies. Startups finding themselves in the throes of managing complex drug development programs realize they may face great risk if they do not begin with the end in mind and integrate data integrity practices early on.
During a recent meeting of data integrity professionals, a fundamental question was posed by a member of the group: “How can one prevent or detect malicious intent as it relates to changes to information and the impact to data integrity?”
Continuous improvement in data integrity can advance a firm on the journey toward a mature culture of quality, particularly through the implementation of QA on the shop floor. Batch record review (BRR) and product disposition are often complicated by data integrity issues and poor data quality.
When creating and managing electronic documents, document metadata deserves as much attention as document content. Firms that do so can improve compliance and even gain potential competitive advantage by realizing electronic documents (with appropriate metadata) as real business assets.
At least 20 percent of all warning letters issued by the CDER Office of Manufacturing Quality in 2017 included explicit observations by inspectors of blatant data integrity violations in laboratory operations. It is incumbent upon us to enforce and reinforce basic CGMP and GDocP practices in our own laboratories and those of our suppliers
There remains a great deal of confusion in the life sciences industry as to how to implement practices to best ensure data integrity. This confusion frequently stems from the assumption that data integrity is something to “turn on” at a particular stage of the product development life cycle.