Good Clinical Practice is a set of internationally recognised ethical and scientific quality requirements which must be observed for designing, conducting, recording and reporting clinical trials that involve the participation of human subjects. Compliance with this good practice provides assurance that the rights, safety and well-being of trial subjects are protected, and that the results of the clinical trials are credible (EU Directive 2001/20/EC in Article 1 (Scope)).
When inspectors identify non-compliance at a particular investigational site with potential impact on the benefit-risk balance, the question arises: Was this finding confined to just this one investigator, or could it be an indication of a more systemic fault in the trial?
There is no easy answer to this question, and each case will have to be looked at individually. In general, 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. As GCP inspections are normally confined to 2-3 investigational sites, which is normally a very limited number compared to the number of sites included in the study, other means may be needed to demonstrate that non-inspected sites are GCP compliant or to quantify the impact if a substantial number of non-inspected sites were to have the same problems as the inspected site(s).
Sensitivity analyses excluding just one or two inspected sites are often not meaningful since these do not address the uncertainty regarding the non-inspected sites. If it is suspected that the GCP inspection finding is likely to be a more systemic problem, but still limited to for example a certain country or region, a specific assessment tool or instrument not used at all sites, or patients experiencing a specific event, sensitivity analyses excluding the affected patients may be helpful in the benefit-risk evaluation.
Sensitivity analyses aimed at investigating the maximum impact of a finding observed at one site on the study results assuming that the finding would have observed at all sites if inspected may also be useful.
If the problem causing an inspection finding leaves an identifiable “footprint” in the clinical database, targeted analyses of the database may be another way to shed more light on the issue of how to generalise the inspection findings. For example, investigators systematically applying identical scores of a psychometric test to patients throughout a study (despite such a scoring pattern being very unlikely in the course of the disease) could be identified by carefully designed searches in the clinical database.
