From Quality by Inspection to Quality by Design: A Roadmap for Success

The past few years have been difficult for the pharmaceutical industry. Profit warnings have abounded and products worth billions of dollars failed during late development or were withdrawn because of safety problems (1). Several drugs will lose blockbuster status over the next few years, resulting in billion-dollar profit losses (2). What is causing these problems?

For one, products are becoming more complex. Because of advances in molecular sciences and technology, pharmaceutical companies can, in theory, develop drug–device–diagnostic hybrid products, healthcare packages for patients with specific disease subtypes, and targeted treatment solutions (i.e., individualized medicine). But these products, which are complex, delicate, and difficult to manufacture, are beyond the reach of most pharmaceutical companies at the moment.

Another factor is that the US Food and Drug Administration has raised the bar on regulations. Janet Woodcock, deputy commissioner for operations at the US Food and Drug Administration, said in a 2003 Wall Street Journal article: "Production techniques were outmoded, and that just refining procedures and documentation wasn't going to change that" (3). Compliance now requires a quality-systems approach that begins with quality by design during development and ends with scientific process control in manufacturing.

In addition, reported quality performance improvements are somewhat misleading. Although the absolute number of recalls is down (from 471 in 2000 to 254 in 2005 [4]), the ratio of development-to-manufacturing issues has doubled during this period, thus indicating the true source of poor quality performance. Although quality improvements have been made, they have been attained by increased efforts in "testing-to-document quality." This approach is costly, superficial, unsustainable, and simply not feasible for new, complex products coming down the pike.

Joiner and Gaudard sharply summarized the difference between quality by inspection, the industry norm now, and quality by design: "Depending on inspection is like treating a symptom while the disease is killing you. The need for inspection results from excessive variability in the process . . . Ceasing dependence on variability means you must understand your processes so well that you can predict the quality of their output from upstream measurements and activities" (5).

The mandate

Building on Deming's principles, FDA's vision for modern pharmaceutical manufacturing (6) can be summarized in five statements:

• product quality and performance are ensured by effective, efficient manufacturing processes;

• product specifications are based on a detailed understanding of how formulation and process elements affect product performance;

• continuous "real-time" assurance of quality;

• regulatory policies and procedures are tailored to the level of scientific knowledge required for product applications, process validation, and process capability.

• risk-based regulatory scrutiny relates to the extent of scientific understanding of how formulation and manufacturing processes affect product quality and performance and the ability of process control strategies to prevent or mitigate the risk of producing a defective product.

In other words, pharmaceutical development and manufacturing will be synonymous with quality by design, scientific manufacturing, and risk-based regulatory oversight in the future. No longer can the industry afford to be ruled by a "cause no problems" mentality focused on getting enough manufactured product through extensive quality controls to meet market demand.

The solution

If pharmaceutical companies are to meet FDA's challenge to improve the predictability of their drug development and manufacturing processes (see Figure 1), they must create a risk-based framework for identifying, developing, and manufacturing drugs, and acquire the scientific knowledge and technological skill to create and deliver increasingly complex products (6). To achieve these goals, pharmaceutical companies will face a myriad of complex questions:

• Which products should be retained, unlisted, or redeveloped?

• Which technologies should be phased out, acquired, or monitored for impact on quality?

• Which facilities are fit for purpose?

• What is the right sequence of investments, where should investment be focused, and over what time period should investments be deployed?

Figure 1: FDA´s challenge is three-fold: to improve innovation, efficiency, and predictability in drug development and manufacturing.

Based on industry research and the authors' extensive experience with clients in the pharmaceutical industry, we believe companies should take a "value-driven compliance" approach to these questions (see Figure 2). This plan consists of five steps: risk diagnosis, business refactoring, transforming industrialization, compliance-centric business architecture, and compliance-centric IT architecture.

Figure 2: IBM value-driven compliance approach.

Risk diagnosis. The first step is risk diagnosis, which involves identifying root causes of failure, quantifying product risk, and then translating the results to dollars.

By analyzing batch records, deviation reports, out-of-specification reports, nonconformance reports, and customer complaints, the root causes of failure can be identified and indeed attributed to inherent characteristics of products, pharmaceutical technologies, and good manufacturing practices (GMPs) systems.

Using advanced statistical analysis techniques, it is then possible to assign a probability to each root cause and to derive the cumulative "beta" factor for any given source. By adding revenue projections to the equations, it is further possible to derive a probability density distribution that represents the company's revenue at risk over a particular time horizon.

Business refactoring. Next, companies should engage in business refactoring in which they prioritize, schedule, and cost-justify the best set of actions to reduce and manage the risk profile.

Once a company has calculated the beta factors and figured out its potential effect on future revenue, it must decide what to do to reduce the root causes of failure and reduce the potential revenue impact.

One method, developed by IBM's business consultants and academic researchers, uses a mixed-integer linear program to calculate a company's revenue at risk based on the computed beta factors for its products, technologies, and GMP systems, and its projected product revenue streams. The method then determines the best course of action by accounting for the capital cost of every risk mitigation option, the actual risk reduction these options are expected to achieve, and the company's goals and constraints.

The model has been tried in a pharmaceutical environment with 18 months of data on 3500 batches from 13 product families manufactured at three different sites. The root causes of all nonconformance incidents were identified, quantified, and classified in more than 90 categories. The information then was used to calculate the beta factors for all 13 product families and seven technology platforms. The output was used to simulate five future scenarios, including the sequence of actions required to reach the desired end state. One scenario was chosen for further analysis.

Business refactoring not only helps to comply with the FDA's vision for pharmaceutical manufacturing, but it also provides the business case for change, together with the sequence and timing of actions.

Transforming industrialization. Third, companies should build quality into products and processes using science, systems, and technology to reduce risk.

FDA's vision requires significant collaboration across the scientific disciplines (e.g., development, quality, and manufacturing) and functions (e.g., review of chemistry, manufacturing, and controls as well as current GMP [CGMP] investigations) incident to pharmaceutical quality. This collaboration is needed to develop an understanding of the systematic factors affecting quality and performance. The enablers of this collaboration are data, information, and knowledge management.

Product lifecycle management (PLM) is a concept for managing product data throughout a drug's lifecycle from early development to retirement. PLM is the electronic backbone that facilitates the flow of data on products and processes between development and manufacturing, which are presently far too disconnected. It provides an integrated framework for regulatory compliance, strategic sourcing, product specification and data management, manufacturing and corrective actions, and product safety and quality control. The two-way data flow (i.e., passing data back from manufacturing to development to inform the design of new products and processes) is crucial, particularly as products and processes become more complex.

Compliance-centric business architecture. Next, companies should make a cultural revolution to embed quality into everyday operations and reduce compliance costs.

According to a Compliance Program Guidance Manual for FDA staff about drug manufacturing inspections, the business architecture is a framework that:

• harmonizes a company's development and manufacturing operations with its GMP systems;

• manages its activities by a set of standardized quality and compliance macroprocesses;

• controls these activities through business-performance management (BPM) and supporting BPM technologies (7).

The key to moving to a systems-based approach is the definition, ownership, and execution of a quality plan for each GMP system. This strategy helps to ensure each system has clearly defined objectives, quality and regulatory requirements, resources, training needs, performance standards, controls, and targets for continuous improvement, regardless of which department "owns" the GMP system.

Defining and instituting a set of standardized processes to manage quality and compliance across operations improves the consistency and quality of execution, standardizes training needs, and helps instill quality and compliance. Once these processes are harmonized and the information content is standardized, BPM and other technologies can play a role in ensuring process conformance, assessing process performance, and delivering process excellence.

Compliance-centric information technology architecture (CCITA). Finally, companies should provide the visionary architecture to enable value-driven compliance and show feasibility.

Finally, every pharmaceutical company must move toward CCITA. This plan provides a "topology" of systems functionality to support integrated quality, manufacturing, and compliance processes, and a service-oriented technical design that uses existing IT solutions and provides services to the quality and compliance macroprocesses.

The functional architecture identifies the system functionality required to support quality by design and scientific manufacturing. Focusing on functional components (e.g., product-data management or enterprise-risk management) instead of on specific applications helps to ensure that the architecture suits the needs of the business rather than those of the applications vendors.

The final dimension of the CCITA is the technical architecture, which facilitates business integration and supports the functional needs of the organization. It has multiple layers (e.g., infrastructure, data, application, service, integration, and interaction), each of which can be instituted and modified without affecting the others. To work, it must be designed according to the principles of open standards, common specifications, and a service-oriented architecture (i.e., a structure that functions as a collection of services that communicate with each other, either to pass data or to coordinate two or more activities).

Benefits

Companies that embrace the new five-element manufacturing paradigm can cut internal failures from 16% to less than 1%, reduce the cost of goods sold from 20% to 17%, and attain process performance levels of 4.5 sigma. For the top 30 pharmaceuticals companies alone, savings of nearly $10 billion are possible. Greater production effectiveness can speed time-to-peak-sales by as much as two years, generating additional revenue of $600 million per year for a drug with peak sales of as much as $1 billion per year (2, see figure 3).

Figure 3: Process understanding can accelerate time to peak sales by as much as two years.

Companies that quantify their quality and compliance risks and then plan the strategic actions to mitigate those risks while building a foundation for quality by design and scientific manufacturing will likely be rewarded with business certainty, operational efficiencies, reduced regulatory oversight, and, most important of all, the freedom for real process and product innovation. Those companies that do not might not be around for the next revision of the GMPs.

Michael T. Ricci* is the global Value Driven Compliance lead in the life sciences and pharmaceuticals practice at IBM Global Business Services, michael.ricci@us.ibm.com

Heather E. Fraser is the global lead in life sciences and pharmaceuticals at the IBM Institute for Business Value, hfraser@uk.ibm.com

*To whom all correspondence should be addressed.

References

1. C. Bowe, "Companies the Americas: Pulling Vioxx Cuts Merck Profits by 29%," Financial Times, Oct. 22, 2004.

2. IBM Global Business Services analysis (White Plains, NY), 2005.

3. L. Abboud and S. Hensley, "Factory Shift: New Prescription For Drug Makers: Update the Plants" Wall Street Journal (Sept. 3, 2003), A1.

4. "Drug Recall Totals for 2005," The Gold Sheet 40 (2), (Feb. 1, 2006).

5. B. Joiner and M. Gaudard, "Variation, Management and W. Edwards Deming," Qual. Progr. (Dec.) 29–37 (1990).

6. US Food and Drug Administration, Innovation and Continuous Improvement in Pharmaceutical Manufacturing: Pharmaceutical CGMPs for the 21st Century (FDA, Rockville, MD, Sept. 2004) http://www.fda.gov/cder/gmp/gmp2004/manufSciWP.pdf, accessed July 21, 2006.

7. L. Bush, "From CGMPs to the Critical Path: FDA Focuses on Innovation, Quality and Continuous Improvement: Inside and Out," Pharm. Technol.28 (7), 34–44 (2004).

8. FDA, Compliance Program Guidance Manual: Drug Manufacturing Inspections (7356.002), (FDA, Rockville, MD), Feb. 1, 2002, http://www.fda.gov/cder/dmpq/compliance_guide.htm, accessed July 21, 2006.