To assess the integrity of a regulatory submission, US Food and Drug Administration (FDA) reviewers typically sample and then scrutinize 10% of the submitted data for data integrity issues. Clinical trial informed consent forms and financial disclosures, manufacturing batch or lot quality control review checklists, investigative drug inventory reconciliations, eCRFs, approved nonclinical study protocols versus results versus later interpretations—all of these and more are scrutinized.
FDA reviewers look at data within the submission and ask:
- Are the records accurate and complete, attributable and original, legible and consistent?
- Is the source data available?
- Have any outliers been omitted?
- Do dates and times match with clinician work hours?
This article lays out several ways in which to build data integrity checks and controls early into the development process. Both regulatory affairs and quality management play crucial roles. Proactive compliance necessitates cross-functional coordination and a holistic view. Leaving data integrity decisions to any one group alone is a sure path to unpleasant surprises. Thus, the earlier an effective data integrity control program can be put in place, the less risk to the submission, to the patient, and to your organization’s bottom line.
Effective submission data quality begins with assessing the risks presented by the clinical site or investigator, just as a firm assess its suppliers for risk to the product and patient safety. This, however, is where regulatory affairs, clinical management, and quality teams typically miss an opportunity to demonstrate strategic thinking.
When it comes to the quality of data in a submission, the financial risk to the company also needs to be taken into account. Any new medicinal product is expected to generate revenues. Thus, the greater the revenues expected, the greater the risk to the sponsor’s financial bottom line from including poor quality data in a submission.
FDA reviewers look at the data quality of a submission first, before they consider any actual proof of safety or efficacy claims. Companies that do not have quality data in their submissions call into question whether FDA safety and efficacy experts will even see the submission. Lack of data integrity equals lack of regulatory agency approval.
With my clients, I frequently review their existing strategy, or lack thereof, with their compliance teams (regulatory affairs, quality, and clinical management). By looking at it from a product development lifecycle perspective, we can avoid the burden of data quality falling on any one group.
There are three key messages to convey:
- Compliance teams are all responsible for the integrity of records supporting new medicinal product development;
- Think prevention, not reaction; and
- Use planning tools such as a clinical regulatory integrated strategic plan (CRISP) to help identify points to verify and/or control for data quality.
Principle of Integrity
Integrity of information, whether that information is generated for or by the company, requires controls that guarantee record authenticity and reliability (see the US Federal Rules of Evidence). For the FDA, this means that data in a submission, or that support a submission, must be attributable, legible, contemporaneous, original, and accurate (FDA uses the acronym ALCOA in training its reviewers).
Two officials in the FDA’s Division of Bioresearch Monitoring (BIMO), Michael Marcarelli and Jonathan Helfgott, suggest several proactive data quality controls to consider:
- Better sponsor and clinical investigator understanding of data quality expectations within the Good Laboratory and Good Clinical Practice regulations;
- Mock FDA audits or compliance gap assessments of clinical sites; and
- Early, and often, sponsor monitoring and intervention to ensure good data integrity practices are effective at the clinical site.
Clinical Regulations Training
In a 2010 BIMO survey of four years’ worth of data quality-related FDA actions—pre-approval inspections, warning letters, and so on—40% were directly attributable to non-compliance by the clinical investigator and/or the sponsor’s oversight team.
First, sponsors need to better train their personnel on the current expectations around clinical and laboratory controls, not just basic regulations. This includes International Harmonization Conference (ICH) activities and FDA guidance documents.
Second, sponsors need to make it easier for clinical investigators to comply with sponsor expectations by providing checklists of good data quality characteristics, listing out requirements as part of the investigator’s brochure, and incorporating FDA good recordkeeping expectations into contracts.
Mock FDA Audits/Gap Assessments
BIMO data also shows that a FDA compliance gap assessment of a clinical site by the sponsor that is designed to show the site what to expect in an FDA inspection has significant merit. All the warning letters issued to clinical sites during the four year period covered by the BIMO survey were only issued to sites who had never undergone either a real FDA inspection or an FDA compliance gap assessment.
The key, of course, is to ensure a balanced critique of the site that includes gaps and recommendations. Finding fault alone is not going to help a clinical site provide a sponsor better quality records for a submission. The sponsor, or its third-party, must be able to provide recommendations, priorities, and suggestions relevant to the site and its business constraints (e.g., suggesting a small clinical site invest in a multi-million dollar electronic data capture system is probably not a realistic recommendation).
Early FDA compliance gap assessments or other sponsor reviews can help sensitize the clinical site to the importance of data quality while allowing the site to make necessary corrections. Examples of data to look at early on include case report forms versus source data, safety reports, investigational product inventory records, and informed consent documents.
An effective, proactive data integrity strategy enhances the quality and credibility of the data supporting your new product’s safety and efficacy. This, in turn, improves the chances for your new medicinal product to achieve FDA approval.