November 2024 SOCRA Source Journal - Journal - Page 33
this was a trend and not a
standalone deviation. The FDA
investigator further questioned
how consistently missing this
test would impact the study’s
overall data quality.
Follow-up actions by the clinical
research site started while the
FDA investigator was still on
site, and the clinical investigator
explained why he thought that
this was not clinically signi昀椀cant.
Clinical research professionals
reached out to the sponsor to
ask about the impact of missing
the biopsies on the study’s
overall data quality. The sponsor
con昀椀rmed that the biopsy
was not related to the study’s
primary objective. Therefore,
the missing data did not impact
data quality. The sponsor
agreed to document this in an
email and sent it to the site.
The sponsor also noted that
this trend had occurred across
all clinical research sites. The
population for the study was
very ill and they were not
expected to come back for
an invasive procedure such
as a biopsy after their disease
had progressed. When this
information was provided to
the FDA investigator, the FDA
investigator agreed that the
deviation made sense.
The lesson learned from this
audit was that it is best to have
a formal chain of documentation
beginning with the sponsor.
In this case, clinical research
professionals should have
requested a protocol revision
or a clinical research sitelevel waiver for an unfeasible
procedure. Writing a NTF each
time a planned deviation occurs
was inef昀椀cient and led to further
questioning.
In another example, the auditors
identi昀椀ed an issue where clinical
research professionals who
were not trained were directly
involved in the informed consent
process. To address this issue,
they wrote an NTF to say that
this issue had been identi昀椀ed
and recti昀椀ed. Unfortunately, the
FDA investigator did not 昀椀nd this
suf昀椀cient. The FDA investigator
also found that the delegation
of authority log did not list the
individuals as being part of the
informed consent process. This
issue had also been identi昀椀ed at
other clinical research sites.
Lessons learned included that
a corrective action would have
been more ef昀椀cient in allaying
the FDA’s concerns about the
study. Since this re-occurring
issue prompted the FDA
investigator to dig further into
the informed consent process,
a corrective and preventative
action plan could address the
root cause of the issue and
be applied across all clinical
research sites or just one site.
A corrective and preventative
action plan would also have
demonstrated improvement to
the process as a whole, as well
as the site’s commitment to
overall quality issues.
Regardless of the outcome of
an FDA inspection, a clinical
research site can always
implement a corrective and
preventative action plan to
rectify issues identi昀椀ed during
an audit. Clinical research
professionals must be careful.
NTFs are not a corrective
action and should not be
treated as a corrective action.
Unfortunately, it is easy to cross
the line between NTFs being an
effective way to document an
issue and casual misuse of this
documentation tool.
NTFs do not always include a
corrective action, as they are
usually generated after the
corrective action has occurred.
They are not source documents.
Occasionally, FDA has
reprimanded clinical research
sites for having too many BandAid solutions and not actually
昀椀xing the issues. Incorrect use
of NTFs can become a map for
an auditor or FDA inspector to
follow. Rather than generating
an NTF, it is more important to
train all parties involved and
assess and communicate the
impact of the deviation to the
clinical team.
The following Warning Letter
examples are from the FDA’s
website. The sponsor received
a Warning Letter from the FDA
due to overuse of NTFs. The
key information is highlighted
in bold. The FDA uses the term
“memos to 昀椀le” instead of NTF.
“1. Failure to secure clinical
investigator compliance with
the investigational plan and
applicable FDA regulations
[21 CFR 312.56(b)].
Our investigation found
that the sponsor failed to
take any action to secure
compliance while the
study was ongoing except
to generate numerous
memos to the 昀椀le after all
subjects had completed
the study. According to an
FDA interview with a sponsor
manager involved with study
3014, these memos to 昀椀le
served as a mechanism to
train the clinical investigator.
However, this same manager
conceded that because the
majority of these memos to
昀椀le were generated after all
subjects had completed the
study, there wasn’t much
value in training the clinical
investigator. We note that
generation of numerous
memos to 昀椀le after all
subjects have completed a
study does not adequately
secure compliance of a
clinical investigator.
SOCRA SOURCE © November 2024
33
