Since decades, the pharmaceutical products have been considered as the highly regulated products meant for human use for accomplishing desired therapeutic benefits for treatment of diverse ailments. Despite continuous innovations by the pharma industry, there has been a repeated set back owing their poor quality and manufacturing standards.

The adoption of systematic approaches has been originated from a thought provoking article that appeared in The Wall Street Journal more than a decade back (i.e., September 2002) was an eye opener for the federal agencies. It stated that “although the pharmaceutical industry has a little secret even as it invents futuristic new drugs, yet its manufacturing standards lag far behind the potato chips and laundry soap makers”.

Figure 1 portrays multiple sources of variability during drug product development owing to variability in drug substance(s), excipient(s), process(es), packaging material(s), etc.

With a brief background, the concept of Quality by Design (QbD) was not new to the world. Originated by J.M. Juran, an American Engineer, in early 1970’s, the concept to quality was implemented by several technology driven industries like, telecommunications, automobiles and aeronautics to develop the quality products and services. Subsequently, the concept was adopted by healthcare industries in 1990’s to produce the medical devices.

Introduction of QbD to pharma industry, however, was quite late. After the growing concern and criticisms on the quality of pharmaceutical products, the ICH instituted a series of quality guidances like Q8, Q9, Q10 and Q11, and PAT, all emphasizing the adoption of systematic principles of QbD as its 21st century quality initiatives. The principal endeavor of ICH has been to accentuate sound science and risk-based understanding of the pharma manufacturing by adopting rational and systematic approaches. Endorsement of such rational paradigms by key global regulatory agencies like US FDA, EMEA, MHRA,

TGA, MCC, SFDA, Health Canada, and many others is unequivocal testimony to their immense significance for all the potential stake holders, viz. patients, industrial scientists and regulators.

With the growing pressure from these federal statutes, the pharmaceutical industry has been thoroughly coerced to gird-up their loins by reorienting its strategies. Regulatory agencies, today, neither require “Quality by Chance” nor “Quality by Testing (QbT)”, rather only require the “Quality by Design (QbD)”.

Based upon the Juran’s quality philosophy, pharmaceutical QbD embarks upon systematic development of product(s) and process(es) with desired quality. As a patient-centric approach, the QbD philosophy primarily focuses on the safety of patients by developing drug products with improved quality and reduced manufacturing cost by planning quality at first place to avoid quality crisis.

Beginning with pre-defined objectives, QbD reveals the pharmaceutical scientists with enhanced knowledge and understanding on the products and processes based on the sound science and quality risk management. Adoption of QbD principles, in particular, tends to unearth scientific minutiae during systematic product development and manufacturing process(es).

Besides QbD, PAT tools have also garnered wide attention in the corner stone of FDA’s quality initiatives for designing, controlling and analyzing the quality of a manufacturing process. Implementation of PAT tools in both lab and plant for efficient monitoring and controlling of the process(es).

One of the integral tools in the QbD armamentarium while developing optimized products and processes has been “Design of Experiments (DoE)” employing apt usage of experimental designs. Amidst a multitude of plausible interactions of the drug substance with a plethora of functional and non-functional excipients and processes, adoption of systematic approaches lead to evolution of the breakthrough systems with minimal expenditure of time, developmental effort and cost.

With the objective of developing an impeccable products or processes, earlier this task has been attempted through trial and error, supplemented with the previous knowledge, wisdom and experience of the formulator, termed as the short-gun approach or one factor at a time (OFAT) approach.

 Using this methodology, the solution of a specific problematic product or process characteristic cannot be achieved, and attainment of the true optimal solution was never guaranteed. However, the QbD-based approach usually provides systematic drug product development yielding “the best” solutions. Such approaches are far more advantageous, because they require fewer experiments to achieve an optimum formulation, reveal interaction among the drug-excipient-process, simulates the product performance and subsequent scale-up. Figure 2 illustrates the QbD-oriented development of drug product embarking upon the comprehensive understanding of the quality traits associated with a product(s) and process(es).

With the percolation of such systematized approaches, the domain of pharmaceutical product development has endowed a newer look towards drug formulation development and subsequent patient therapy. Owing to the immense benefits, the applications of QbD are galore such as in drug substance manufacturing, formulation development, analytical development, stability testing, bioequivalence trials, etc. Figure 3 illustrates the flow chart portraying the QbD-based development of drug products in an industrial set-up.

Steps in QbD-based drug delivery optimization
The holistic QbD-based philosophy of product development revolves around five fundamental elements viz. defining the quality target product profile (QTPP), identification of critical quality attributes (CQAs), critical formulation attributes (CFAs) and critical process parameters (CPPs), selection of apt experimental designs for DoE-guided, precise definition of design and control spaces to embark upon the optimum formulation, postulation of control strategy for continuous improvement. Figure 4 illustrates the five step methodology for drug product development employing QbD-based approach.

Defining quality target product profile (QTPP)
The quality target product profile (QTPP) is a prospective summary of quality characteristics of the drug delivery product ideally achieved to ensure the desired quality, taking into account the safety and efficacy of the drug product. During drug product development, QTPP is embarked upon through brain storming among the team members cutting across multiple disciplines in the industry. Critical quality attributes (CQAs) are the physical, chemical, biological or microbiological characteristic of the product that should be within an appropriate limit, range or distribution to ensure the desired product quality.

 There are various types of CQAs associated with the drug products such as drug substance CQAs, excipients CQAs, packaging material CQAs, etc. The identification of prime CQAs from the QTPP is based on the severity of harm a patient may get plausibly owing to the product failure. Thus after defining the QTPP, the CQAs which pragmatically epitomize the objective(s), are earmarked for the purpose.

Prioritization of input variables through risk assessment
Material attributes (MAs) and process parameters (PPs) are considered as the independent input variables associated with a product and/or process, which directly influence the CQAs of the drug product. PPs, further, can be non-critical or critical process parameters (CPPs). Ishikawa-fish bone diagrams are used for establishment of cause-effect relationship among the input variables affecting the quality traits of the drug product. Figure 5 illustrates a typical cause-effect diagram highlighting the plausible sources variability and their impact on drug product CQAs.

Prioritization exercise is carried out employing initial risk assessment and quality risk management (QRM) techniques for identifying the “prominent few” input variables, termed as critical material attributes (CMAs) and critical process parameters (CPPs) from the “plausible so many”. This process is popularly termed as factor screening. Comparison matrix (CM), risk estimation matrix (REM) and failure mode effect analysis (FMEA) constitute the commonly employed risk assessment techniques.

Using these techniques, various MAs and PPs are assigned with different risk levels viz. low, medium and high risk based on their severity and likelihood of occurrence. The moderate to high risk factors are chosen from patient perspectives through brainstorming among the team members for judicious selection of CMAs. QRM is rational approach which not only provides holistic understanding of the risks associated with each stages of product development, but also facilitates mitigation of risks too. Figure 6 portrays the flow layout of overall risk assessment plan employing risk assessment and risk management for identifying the potential CMAs employing a prototype REM model.

Factor screening studies
Along with risk assessment studies, factor screening studies employing experimental designs are used for identifying the critical variables influencing the product and process performance. The low-resolution experimental designs (e.g., Fractional Factorial, Plackett-Burman and Taguchi designs) are highly helpful for screening and factor influence studies. Before venturing into product or process optimization, prioritization of CMAs/CPPs using such QRM and/or screening is considered to be obligatory.

Design-guided experimentation & analysis
Response surface methodology is considered as a pivotal part of the entire QbD exercise for optimization of product and/or process variables discerned from the risk assessment and screening studies. The experimental designs help in mapping the responses on the basis of the studied objective(s), CQAs being explored, at high, medium or low levels of CMAs. Figure 7 provides bird’s eye view of key experimental designs employed during QbD-based product development. Factorial, Fractional-Factorial, Box-Behnken, Central composite, optimal and mixture designs are the commonly used high resolution second-order designs employed for drug product optimization. Design matrix is a layout of experimental runs in matrix form generated by the chosen experimental design, to guide the drug delivery scientists. The drug formulations are experimentally prepared according to the design matrix and the chosen response variables are evaluated meticulously.

Modelization & validation of QbD methodology
Modelization is carried out by selection of apt mathematical models like linear, quadratic and cubic models to generate the 2D and 3D-response surface to relate the response variables or CQAs with the input variables or CMAs/CPPs for identifying underlying interaction(s) among them. Multiple linear regression analysis (MLRA), partial least squares (PLS) analysis and principal component analysis (PCA) are the key multivariate chemometric techniques employed for modelization to discern the factor-response relationship. Besides, the model diagnostic plots like perturbation charts, outlier plot, leverage plot, Cook’s distance plot and Box-Cox plot are also helpful in unearthing the pertinent scientific minutiae and interactions among the CMAs too.

 The search for optimum solution is accomplished through numerical and graphical optimization techniques like desirability function, canonical analysis, artificial neural network, brute-force methodology and overlay plot. Subsequent to the optimum search, the optimized formulation is located in the design and control spaces. Design space is a multidimensional combination of input variables (i.e., CMAs/CPPs) and out variable (i.e., CQAs) to discern the optimal solution with assurance of quality. Figure 8 illustrates the interrelationship among various spaces like, explorable, knowledge, design and control spaces.

Usually in industrial milieu, a narrower domain of control space is construed from the design space for further implicit and explicit studies. Figure 9, in this regard, portrays a typical design space overlay plot for wet granulation process employed for manufacturing tablet dosage forms.

QbD validation, scale-up and production
Validation of the QbD methodology is a crucial step that forecasts about the prognostic ability of the polynomial models studied. Various product and process parameters are selected from the experimental domain and evaluated as per the standard operating conditions laid down for the desired product and process related conditions carried out earlier, commonly termed as checkpoints or confirmatory runs. The results obtained from these checkpoints are then compared with the predicted ones through linear correlation plots and the residual plots to check any typical pattern like ascending or descending lines, cycles, etc. To corroborate QbD performance, the product or process is scaled-up through pilot-plant, exhibit and production scale, in an industrial milieu to ensure the reproducibility and robustness.

QbD-oriented measurement of process capability
Process capability measures the inherent variability of a stable process that is in a state of statistical control in relation to the established acceptance criteria. Defined metrics are used for assessing the capability of pharmaceutical processes such as process capability indices (i.e. Cp and Cpk), and process performance indices (i.e. Pp and Ppk). These indices tend to identify the reproducibility characteristics of any process to delivered desired quality in the product. Dynamic tools like check sheets, histograms, quality control charts, scatter diagrams, defect concentration charts, etc. are useful in this regard for observing the minute scientific information on any specific pattern of deviation in the performance of an unit operation.

Postulation of QbD control strategy and continual improvement
A holistic and versatile “control strategy” is meticulously postulated for “continuous improvement” in accomplishing better quality of the finished product. Principally, control strategy embarks upon automated engineering control, pharmaceutical process control and end-product testing. The pertinent information and scientific minutiae gained from each stages of product development are appropriately culminated to form of a control strategy, which can be futuristically employed as a source of information in the product life span.

Software usage during QbD
The merits of QbD techniques are galore and their acceptability upbeat. Putting such rational approaches into practice, however, usually involves a great deal of mathematical and statistical intricacies. Today, with the availability of powerful and economical hardware and that of the comprehensive QbD software, the erstwhile computational hiccups have been greatly simplified and streamlined. Figure 10 enlist the select computer softwares available commercially for carrying out QbD studies in industrial milieu. Pertinent computer softwares available for DoE optimization include Design-Expert®, MODDE®, Unscrambler®, JMP®, Statistica®, Minitab®, etc are at the rescue, which usually provide interface guide at every step during the entire product development cycle. Softwares providing support for chemometric analysis through multivariate techniques like MNLRA, PCA, PLS, etc. encompass, MODDE®, Unscrambler®, SIMCA®, CODDESA®. For QRM execution using fish-bone diagrams, REM and FMEA matrices during risk assessment studies, etc., softwares like, Minitab®, Risk®, Statgraphics, FMEA-Pro, iGrafx, etc., can be made use of.

Epilogue
Today, the federal agencies look for assurance of patient-centric quality “built-in” into the system, rather than through end-product testing. Notwithstanding the enormous utility of QbD-based philosophy in developing optimal drug products, it leads research mindsets to evolve “out-of-box” strategies too. As variability tends to exist at each and every stages of product development life cycle, QbD application needs to be omnipresent.

Apt implementation of QbD paradigms, accordingly, would be pivotal in achieving a “win-win situation” for patients, drug industry and regulators. The practice of systematic QbD implementation for products has undoubtedly spiced up over the past a few decades, yet it is far from being adopted as a standard practice. Federal regulations for generic drug products are already in place. Several initiatives still need to be undertaken to inculcate mundane use of diverse QbD paradigms in the holistic domain.

Apart from these, the synergistic use of in-process PAT and RTRT tools in tandem with process engineering approaches like extensometry and chemometry, can also be helpful in ameliorating product and process understanding and enhancing the process capability for efficient manufacturing. With the growing acceptance of QbD paradigms, in a nutshell, it is rationally prophesized that soon these QbD philosophies will be required to be implemented to innovators, biosimilars, analytical development, API development and even beyond.