|Year : 2023 | Volume
| Issue : 1 | Page : 1
Design of experiments: A design to improve pharmaceutical research
Bhaswat S Chakraborty
Department of Pharmacology, Shree S. K. Patel College of Pharmaceutical Education and Research, Mehsana, Gujarat, India
|Date of Submission||11-Dec-2022|
|Date of Acceptance||12-Dec-2022|
|Date of Web Publication||20-Jan-2023|
Dr. Bhaswat S Chakraborty
Shree S. K. Patel College of Pharmaceutical Education and Research, Ganpat Vidyanagar, Mehsana, Gujarat
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Chakraborty BS. Design of experiments: A design to improve pharmaceutical research. J Adv Pharm Technol Res 2023;14:1
Many important statistical and mathematical concepts, including that of design of experiments (DOE), were ushered into pharmaceutical sciences a bit later than expected. There are many reasons for this delay which are beyond the scope of this editorial but their usefulness in current pharmaceutical R and D must be emphasized. When we design an experiment – be it causal or observational – we know one thing for sure that examining the effect of one factor at a time is not only time-consuming but also it is mostly futile. Instead, we can look at several factors (independent variables) together at different settings through different runs or experiments and look at their effects on the output or response (dependent) variable. This is a smart way of doing an experiment or even simulating or explaining one.
DOE can be thought of a systematic approach to control and optimize input variables such that the output variable(s) is maximized. In practice, we label independent variables as factors meaning that they have independent effects on the output. In addition, interactions of independent factors are also considered when necessary. Needless to say that random or uncontrolled variables will also have some effect on the output albeit much smaller than those of the controlled variables when an effective DOE is used. As long as the effect of controlled independent variables is significantly larger than the effect of uncontrolled variables, the experiment is successful and worthwhile.
One of the most common designs is a full factorial design (FFD). Now, a full factorial design with as many as 7 factors and say 2 or 3 levels will have thousands of experimental runs. This is rather impossible to practice. Even when factors are fewer and levels are only 2, the total number of runs (say 30-50) can be resource draining as well as pointless in the sense of gathering useful information about which factors are important and at what levels. This is why we resolve to a lower number of runs by statistical means such that no important information is lost at the same time much fever runs are to be made.
Let us take a practical example in pharmaceutics. You are developing an optimal solid oral dosage formulation for a nearly insoluble active pharmaceutical ingredient. You have 6 factors of excipients, bioavailability enhancers, glidants, surfactants, etc., at two levels with 64 possible runs in an FFD. This definitely has to be resolved to a lower number of runs with a fractional factorial or a screening design. There are many examples in literature where a formulation has been optimized by a screening design within DOE such as central composite design or Box-Behnken., Such screening greatly reduces the potential number of experimental runs without losing any important information.
| References|| |
Kamala Kumari PV, Yarraguntla SR, Sharmila M, Gulibindala E. Application of box-behnken design for formulation parameters of eslicarbazepine tablets. Indan J Pharm Sci 2021;83:575-83.
Hassan H, Adam SK, Alias E, Meor Mohd Affandi MMR, Shamsuddin AF, Basir R. Central composite design for formulation and optimization of solid lipid nanoparticles to enhance oral bioavailability of acyclovir. Molecules 2021;26:5432.