In 2013, The US-FDA reported
that laboratory processes and deficiencies associated with laboratory controls
were ranked in the top three most frequent causes of observations following
US-FDA inspections. And same report also cited that an increase of 50% in
warning letters related data integrity.
During Annual meeting of
International Society of Pharmaceutical engineering (ISPE) held in 2014 at Las
Vegas, it was reported that the FDA has identified that a dozen Indian
pharmaceutical manufacturers who had problem with data integrity practices at
their facilities. That number is significant, since India and china account for
80% of API production.
Other regulatory bodies,
including the European Medicine Agency have made similar observations. It is
expected that this trend will continue to grow.
Data integrity is currently
one of the highest cited area in regulatory observations yet data integrity is
not a new requirement.
For years the basic
principles have been described in international GMP guidelines. Here I will
highlight the meaning and principles of Data Integrity.
Definitions:
DATA:
Data is the information
derived or obtained from ‘Raw data’.
RAW
DATA:
Original records and
documentation retained in the format in which they were originally generated
(Paper or electronic) or as a ‘True copy’.
META
DATA:
Meta data is the data that
describes the attributes of other data and provide context and meaning.
Examples:
1. For
example, Analyst-A reported impurity-A results as 0.05% from HPLC chromatogram.
Raw data:
HPLC Chromatogram
Data: 0.05%
Meta Data:
Analyst-A, Impurity-A
2. For
example, Operator-A recorded in BPCR as Reactor-A temperature raised to 20°C
and maintained for 1 hr.
Raw data: BPCR
Data: 20°C, 1 hr
Meta Data: Operator-A,
Recorder-A
DATA INTEGRITY:
- Data integrity is a policy of the firm which assurance that all data are accurate, complete, intact and maintained within their original context including their relationship to other data records throughout the data life cycle.
- In short, data integrity aims to prevent unintentional changes to data or information. i.e ensuring data integrity means protecting original data from accidental or intentional modification, alteration, malicious intent (fraud) or even data deletion (data loss).
ALCOA+ Principle:
- As
per Regulatory, data should meet certain fundamental elements of quality as
follows whether they are recorded on paper or electronically.
- ALCOA
is commonly used acronym short for “Accurate,
Legible, Contemporaneous, Original
and Attributable.
- Later
on Complete, Consistent, Enduring and available also added to ALCOA principle
which then termed as ALCOA+.
- As
per ALCOA+ Principle data should be Accurate, Legible,
Contemporaneous, Original, Attributable, Complete, Consistent, Enduring and
available.
ACCURATE:
The term Accurate means data are
correct, truthful, valid and reliable. This means an honest, accurate and
thorough representation of facts describing conduct of study.
Example-1:
A manufacturing instruction state as follows
1. Take
25 gram of RM1 and add to 100 L water
2. Mix
for 20 min. Check complete dissolution
3. Heat
to 70°C.
Now
while the solution was being heated for whatever reason the temperature rose to
72°C.
Is
it deviation? Obviously Yes! So what does one do? Report? Ideally Report? What
happens then Investigation, Risk assessment, CAPA, Massive documentation and
probability of Auditors comments?
So,
is there an easier remedy? Simply write 70°C in BPCR instead 72°C?
Example-2:
The result of Impurity from a HPLC
chromatogram getting out of specification results as 0.11% against limit 0.10%.
Ideally OOS initiation, Investigation, Impact assessment, CAPA and training.
So,
is there an easier remedy? Simply write adjust integration parameters and
adjust impurity result to 0.09% instead of 0.11%.
Other
examples for inaccuracy:
· Not
or inadequately qualified/ calibrated / maintained equipment or instruments
used.
· Not
or inadequately validated method / process used.
· Investigation
of OOS results & Deviation not done or doubtful.
So, never compromise accuracy at any situation
record actual accurate details.
There
will be times when source documents are in complete, inconsistent, or wrong. If
changes need to be made modifying a document always need to done in complaint
manner. When the source is electronic, Audit trails can provide transparency to
prevent data from being altered in a way that it is difficult to detect.
Finally, Data must correctly reflect the
action / observation made.
LEGIBLE:
Data
should be readable and understandable and must be possible to interpret data
after it is recorded.
Example-1:
The
typo error in date was identified in the document as 20/09/2016 instead of
19/09/2016. During the correction good documentation practices were not
followed due to that old entries are not readable or not understandable. i.e.
generated inlegibel document
Example-2:
During
the issuance BPCR, it was noticed that the Xerox (True copy) of the master BPCR
not legible due to printer problem but same was issued. i.e. generated
inlegible document.
CONTEMPORANEOUS:
Data
must be recorded at the time it was generated and observed. The documentation
should serve as an accurate attestation of what was done and what was decided
and why i.e. what influenced the decision at that time.
Example-1:
A manufacturing instruction state as
follows
1. Take
25 gram of RM1 and add to 100 L water
2. Mix
for 20 min. Check complete dissolution
3. Heat
to 70°C and maintain for 15 min.
But
during recording of BPCR, Ideally record 70°C without noticing online for
actual temperature.
Example-2:
During the HPLC analysis, the
online entries (Updation of balance usage, pH meter usage logbooks) were not
made and all entries were made after completion of analysis.
ORIGINAL:
Original Record can describe
the first source capture of data or information. If corrections or revisions
need to be made to original record, changes should not obscure prior entries.
Example:
In case of HPLC, The first
source data is electronic copy of chromatograms and in case of balance the first
source data is paper weight print which comes under original data.
ATTRIBUTABLE:
Attributable means
information relating to originator of the data. i.e. when documenting data on a
paper every written element is need to be tracked back to the authorized
individual who is responsible for recording it. It requires the signature and
the date.
Audit trail in the
electronic system make it very obvious who created record, when it was created,
who made a change, when the change was made and reason or the change. A
complaint system will automatically track this information and enable
electronic signature. Data is attributable to a unique user with secure password
and role based permissions.
COMPLETE:
Complete data can be describe
all relevant data is present and available. i.e. Complete data is data with all
required data.
Example:
In case BPCR, BPCR is only a
data not a complete data, a complete data includes Raw material issuance slips,
on demand slips, in-process analysis reports, Labels..etc.
CONSISTENT:
All elements of record, such
as sequence of events follow on and are dated or time stamped in expected sequence.
i.e consistent practices to be followed like Good documentation practices…etc.
For example correction of
wrong entries to be done in same manner for all documents.
AVAILABLE:
Data/Documents should readily
available for review and auditor or inspection over the lifetime of document.
Records must be available
for review at any time during the required retention period, accessible in readable
format to all applicable persons who are responsible for their review whether
for routine release decisions, investigations, trending, annual reports, audits
or inspections.
