During the requirements analysis phase of analytical data systems used in regulated environments, one question is frequently
posed: Does Part 11 require "compliant instrument control" functions in an analytical data system?
Actually, it is not the right way of asking the question in the first place. FDA has a broad charter to protect consumer safety
and public health. Let us therefore take a look at the general principles of 21 CFR Part 111 and discuss what they mean for instrument control.
- FDA is applying risk-based principles to decide where to focus its effort. In the context of drug development and production,
"risk" means potential problems for the safety and effectiveness of drugs.
- FDA expects regulated firms to use a risk-based approach when implementing their business practices, especially regulated
- When scrutinizing a company's business processes and practices, FDA will inspect and enforce the requirements from predicate
rules such as cGMP, GLP, GCP, and other relevant regulations, as illustrated by recent cGMP warning letters issued by FDA.
- Regarding Part 11 requirements, the agency will look at the relevant requirements from predicate rules and a firm's business
practices to decide whether Part 11 requirements apply or not. The new guidance2 makes it clear that FDA may take your business practices into account. Therefore, it is important to determine and document
in advance whether the electronic record or the paper record will be used to perform regulated activities.
- The spirit and intent of 21 CFR Part 11 is to ensure trustworthy and reliable electronic records. It defines the ground rules
for using electronic records instead of traditional paper records. Electronic records can only be trustworthy and reliable
if they are managed in an environment and with technical controls that minimize risks to data integrity and maximize the probability
of detecting data integrity problems.
What does this mean for instrument control functions in analytical data systems? Let us respond with another question: How
can an analytical laboratory prove that a given result was generated according to the defined procedure without proper documentation
of the instrument control parameters used during the analysis? How can the analytical laboratory ascertain the validity of
the result without evidence that the instrument was really doing what the analyst thought it would do and that the instrument
was within the specifications required for the analysis?
Without generating, storing, and managing the metadata, the trustworthiness and reliability of electronic records in an analytical
data system are questionable. Instrument control, associated diagnostics, instrument events, alerts, logbooks, and other "sanity
checks" are part of that metadata. In a way, this is similar to the voice recordings and other data stored in the famous "black
box" of an aircraft, although it is collected for a different purpose. Without the black box, there would be hardly any raw
data or metadata available to investigate an aircraft accident and learn how to prevent a similar disaster in the future.
Levels of Instrument Control
Control of devices and instruments can be implemented at varying levels of sophistication and complexity (see Table 2). When
we first published our discussion of instrument control in the context of Part 11 compliance in BioPharm magazine in 2000,3 most analytical laboratories operated a diverse installed base of instruments, often from a variety of manufacturers. In
many cases, the data systems used at the time for the evaluation of the analytical measurements did not have any direct instrument
control capabilities and data was recorded through an analog-digital converter. This we defined as level-1 instrument control.
This approach is frequently used to integrate an instrument into a system from a different manufacturer. With level-1 control,
analysts are forced to set parameters using the instrument's own panel or keyboard, as the control software cannot manage
these setpoints. In such cases, it is typically impossible to obtain a printout of the instrument settings that are used during
an analysis, so analysts are forced to document instrument parameters manually. Level-1 control also lacks mechanisms for
documenting and detecting mismatches of injections with sample identification and vial numbers, which requires supporting
binary-coded decimal (BCD) or direct barcode input (from an autosampler).
Table 1: Levels of Instrument Control
Since the original publication, the landscape has changed somewhat. We have observed a trend to replace level-1 systems with
systems that implement full instrument control. Awareness of the value of electronic raw data has significantly increased
since then. Throughout industry, strip-chart recorders and integrators generally have been replaced with computerized systems.
Many systems implement at least a rudimentary level of instrument control. Level-2 control supports the basic parameters of
an instrument, such as solvent composition, flow, oven temperature, or detector wavelength. Level-2 implementations are typically
not developed by the manufacturer of the instrument, but by third-party software developers. Oftentimes, the software developers
do not have the official documentation of the instrument's communication protocol and instead use reverse-engineering, user
experience, protocol analyzers, and snippets of unofficial memos. Obviously, instrument manufacturers cannot guarantee these
so-called "solutions" developed by third parties without the official control codes. Therefore, firms using such systems should
anticipate additional effort will be necessary to qualify and validate such systems. Since the manufacturer of the original
instrument may be neither aware nor responsible for the implementation of the communication protocols, instrument firmware
updates may result in non-functional communication between instruments and the data system. Level-2 instrument control can
work. However, the implementation is not based on the documented properties of the instrument, so if it works, you don't really
know why. Furthermore, the implementation of error handling and logging is typically weak for systems in this category. When
selecting a system that is supposed to control instruments from other manufacturers, verify that the control codes were officially
obtained from the instrument manufacturer to avoid significant additional qualification effort.
Table 2: Characteristics of networked instruments
In most cases, manufacturers implement full instrument control (level 3) for their own systems. Provided the supplier has
sufficient compliance experience, you can expect these systems to maintain complete sets of raw data and metadata along with
the proper documentation. In this category, one also can expect quite sophisticated error reporting and handling, which makes
it easier to verify that an analysis was indeed completed without technical failures or to diagnose an error if it occurred.
Despite the lack of standard communication protocols to date, some systems offer "generic instrument control" capability.
This allows integration of instruments using XML-based vendor specific instrument adapters (VSIA) through a de-facto standard
serial interface (RS232) and data acquisition through an appropriate interface. This approach can be used for full control
of any instruments with an RS232 control interface and analog data output.4
Some manufacturers have implemented an additional level of instrument capabilities that can be controlled from within the
data system. These functions are the basis for the execution of detailed and sophisticated instrument diagnostics as well
as other service functions. This includes provisions for preventive maintenance, remote diagnostics, and early maintenance
feedback (EMF), a technique initially used in the aeronautics industry (to alert technical personnel to perform maintenance
jobs proactively before failure). Systems implemented at this level provide sophisticated support for tracking device serial
numbers, firmware revisions, EMF alerts, tagging of device components, service history, and diagnostics. This information
is not only handy and important for inventory tracking of validated equipment, but it also conforms to the device check controls
required by Part 11.
Level-4 instrument control implementations are based on bidirectional, handshake-based communication between the controller
and the instrument. This means that the recipient of a data record actively acknowledges receipt of the record by notifying
the sender, preventing loss of communication and "misunderstandings" between devices.