EMA Offers Inside Look at Clinical Trial Inspections

Deficiencies in a sponsor, CRO or other organisation delegated responsibility by the sponsor but unrelated to the clinical trial site are almost alyways systemic, the EMA says in its points to consider document​ to help inspectors draft their reports.

But deficiencies found at trial sites are more difficult to address. The EMA cautiously notes that identifying good clinical practice (GCP) noncompliance at one site could be an isolated finding, especially if a trial involves multiple sites. But if the same finding is identified “at two different investigational sites (preferably in different countries/regions), it greatly increases the probability that the finding reflects a general problem with the study​.”

Analyses of sponsors’ clinical databases may be a way to “shed more light on the issue of how to generalize inspection findings​,” the EMA notes, particularly if a problem leaves an identifiable footprint on the database. For example, investigators applying identical scores of a psychometric test to patients throughout a study could be identified through specific searches in the clinical database.

The EMA focuses on findings that should be of particular importance for inspectors, noting the top eight GCP inspection findings most likely to impact a risk-benefit evaluation are:

• Deficiencies in the blinding of a medication, particularly if it leads to bias related to patient reported outcomes, psychometric tests and other subjective endpoints;
• Problems with randomization that raise questions about the comparability of the treatment groups and the relationship between treatment and effect;
• Violation of diagnostic inclusion- and exclusion criteria
• Violation of procedures related to the assessment of the primary efficacy endpoint;
• Inadequate reporting of adverse events and other safety endpoints;
• Missing source documentation;
• Faults in data management, statistical programming and analyses, though difficult to detect, especially at the level of sponsor or CRO, have the potential to lead to completely false study conclusions; and
• Fraud and other scientific misconduct.

But even inspection findings outside of those most likely to impact the benefit-risk evaluation can become significant if several of them seem to indicate overall poor data quality, according to the EMA. One such example is a failure to document pre-specification of analyses prior to breaking a study blind so as to avoid analyses tailored to the trial’s findings.

In addition, major ethical flaws, even ones not directly influencing a benefit-risk assessment, “should have an impact on the final conclusions about approvability of an application​,” the EMA notes.