Q and
A
Design of Experiment (DOE) Back
Q. What is DOE?
A. Design of Experiment is a structured analytical method for determining
the relationship between factors affecting a process and the output
of that process. It can also be said that DOE is a statistical method
to plan experiments and analyze data so that the maximum amount of
information is obtained with the fewest number of runs. DOE has broad
applications in science and engineering in comparison tests, process
development & optimization, formulation & product design, validation
tests, and process management. For additional information on DOE, refer
to “Design and Analysis of Experiments” by Douglas C. Montgomery,
6th edition, John Wiley & Sons, Inc, 2005.
Q. Why do I need DOE?
A. The classical strategy of experimentation is the one-factor-at-a-time
(OFAT) approach. This method consists of varying one factor at a time
while keeping the others constant. The OFAT approach is time-consuming,
does not focus on the critical factors affecting the process and does
not detect factor interactions. To make matters worse cell culture
development and in particular, formulation development has frequently
used trial and error to find the desired settings leading to inconsistent
results. By detecting factor interactions, concentrating on relevant
factors and eliminating trial and error, statistically designed experiments
reduce variability, improve process performance, and reduce developmental
time and cost.
Q. When do I need DOE?
A. Statistical modeling minimizes the risk of making incorrect decisions,
and is particularly useful to:
• Screen unknown factors and select
the vital factors from the trivial ones.
• Determine how critical factors
interact and affect a process.
• Define the best mixture combination
in a formulation.
• Find the right factor settings
for optimum performance.
• Improve process control and
minimize variability. In other words, designed experiments are useful at any stage of cell
culture development from bench to production. In addition, because cell
culture processes are subject to experimental errors, this statistical
approach to planning experiments is critical to draw meaningful conclusions
from the data obtained.
Q. How do I integrate DOE with OFAT?
A. You cannot use DOE and OFAT at the same time. However, practical information
you have obtained using OFAT can be used for DOE. In fact, the best
DOE results are obtained when you have good knowledge of your system.
Let’s say you know which factors affect your process but want
to reduce process variability. In this case, we use your previous knowledge
of the factors to design an experiment that will bring your process
under control.
Q. Can I use DOE to quickly fix a problem?
A. If you are looking for quick fixes, DOE is not your answer. DOE uses
a sequential approach where the next design is based on the results
obtained from the previous design. This is repeated until the desired
response is achieved. DOE is used to improve the quality, consistency,
and efficiency of a process whether at the research or production level.
It is also used to determine the commercial viability of a product
and/or process, by looking at efficiency versus cost.
Q. How many DOE reiterations does it take to complete a project?
A. The number of designs required to complete a project is highly dependent
on the stage of the cell culture process. Screening assays for example
require more reiterations than validation assays. We always try to
minimize the number and size of the designs to save time and cost,
but never at the expenses of valid information.
Q. What other considerations should be kept in mind when using DOE?
A. First and foremost, the quality of the data obtained from a statistically
designed experiment depends on the quality of the input data, i.e.
garbage in = garbage out. Do not expect reliable results when factor
specifications are incorrect, historical data are used, equipment is
not properly calibrated, and the experimenter has poor technical skills
or limited knowledge of the system. In addition, designed experiments
must follow a structured lay out (random runs, replicates, blocks,
etc) which should not be modified without affecting final data analysis.
This is why interactions with customers are so important during the
planning stages of DOE to ensure proper execution of the experiments.
Q. Do I need a statistical background to use a DOE design?
A. No, it is not required. All you need to do is the laboratory work
according to the layout of the experimental design provided to you.
This design consists of a random combination of factors at specified
concentrations (or settings) in a structured matrix created by our
statistical software, which is very easy for you to follow. Statistical
analysis of the final data is also performed by our software. We do
the design, interpret the results, review them with you and provide
you with easy-to-understand graphs and summary of data evaluation.
Q. How does Cell Culture Solutions design experiments?
A. Experimental design and data analysis are performed using “Design-Expert” statistical
software from Stat-Ease, Inc. By combining our expertise in cell culture
with DOE techniques, we are able to design experiments that provide valid
and objective conclusions.
Q. Is the statistical software properly validated?
A. Stat-Ease, Inc. uses several validation
procedures to test its software.
Because the FDA's 21 CFR Part 11 ruling does not apply to the software,
it can be used for developmental and validation studies in the biopharmaceutical
industry. In fact, the DOE approach is part of a set of tools recommended
by the CBER/FDA for the Process Analytical Technology (PAT) Initiative.
The goal of PAT is “to design and develop processes that can
consistently ensure a predefined quality at the end of the manufacturing
process”. When designing experiments using statistical tools,
Cell Culture Solutions will ensure that compliance issues are thoroughly
reviewed with customers before issuing a DOE protocol. Back to top Back (Return
to “Q and A”)
Cell Culture – Read
about the use of DOE in cell culture.
Customers – Discover
the profile of our customers.
Service
Options – Decide
which service is best for you.
|
