During the past decade, pharmaceutical manufacturing and quality functions were thrust into an unaccustomed spotlight. After an extended spate of consent decrees, noncompliance and drug quality problems, FDA began to look for root causes and to consider lessons in manufacturing science that regulators and manufacturers might learn from other industries.
At FDA’s Center for Drug Evaluation and Research, the Office of Pharmaceutical Sciences held a series of meetings to analyze these issues. In quick succession the Process Analytical Technology (PAT) team was formed in 2002, the 21st Century cGMP’s were published, outlining a risk-based approach to regulation and compliance, and the PAT Guidance appeared in 2004. During this period, a challenge was issued to drug companies by then-Commissioner Mark McClellan, who was quoted by The Wall Street Journal as saying that the science of drug manufacturing was “behind that of potato chip and soap making.”
Today, drug companies still face quality and compliance problems ranging from supply shortages to consent decrees and warning letters. We asked a number of expert industry observers, including some who had led the call for change last decade, to share their thoughts on how pharmaceutical manufacturing has progressed, and whether anything had actually changed. This is a brief summary of some of what they had to say.
Examining the big picture, experts agree that the “sea change” predicted for drug manufacturing still hasn’t occurred. “McClellan’s quote still resonates today,” says Bikash Chatterjee, President and CTO of Pharmatech Associates, Inc. “Our processes have not demanded the level of sophistication required in other market segments where margins are much tighter, so it’s natural to expect them to evolve at a different rate.”
Although some companies have implemented processes that run at 6 Sigma, the industry overall still operates at the same Sigma level of 2.5 to 3 that it did last decade, says G.K. Raju, Executive Director of the Pharmaceutical Manufacturing Initiative at MIT, and President of Light Pharma, Inc. At FDA’s Science Advisory Board meetings last decade, Mr. Raju documented the high cost of traditional QC and its contributions to pharma’s unusually high cycle times.
“Progress seems to be slow when one looks at reports on shortages, recalls and other quality issues,” says Ajaz Hussain, CSO at Philip Morris International, and former Deputy Director of OPS and head of the PAT Team at FDA. FDA had launched PAT and the 21st Century initiatives, in part, to prevent just these types of problems from recurring. “Efforts towards developing guidelines have been laudable,” he says, “but additional emphasis is needed on certain fundamental aspects of quality such as effective QMS, training, root-cause investigations and setting of specifications based on an analysis of variance.”
However, there are bright spots. More people seem to realize that changing operations is not simply doing some Lean Sigma training, but changing mindsets, says Thomas Friedli, professor at the University of St. Gallen, whose research group has been closely tracking pharmaceutical operational excellence progress for a number of years. “The complexity that drug plants have to master has increased, so a plant that maintains operational performance over the years has in fact improved,” he says. Traditional pharma leads in Lean and Six Sigma efforts, but biopharma is now devoting more resources to operational excellence in manufacturing, and to consolidating isolated improvement activities, Friedli says. Currently, the University of California Berkeley and NSF have launched the Initiative for Research in Biopharmaceutical Operations to study this issue, while, funded by the Sloan Foundation, MIT and Georgetown are exploring related issues and the impact of globalization on regulation and innovation.
Increased Use of Modern Manufacturing Tools
Another positive sign, Raju says, is the fact that more drug companies are applying established industrial engineering tools such as process capability analysis and statistical process control. These methods are specifically called out in FDA’s Process Validation Guidance, says biopharmaceutical consultant James Blackwell, but more companies are realizing their value in catching trend deviations, providing new process insights and driving continuous improvement. “Looking for trends once a year is not cGMP anymore,” he says.
“Today, even generic manufacturers routinely have control charts on the shop floor, a big change from a decade ago,” Chatterjee notes, “while most companies use some kind of scorecard for product performance.” In addition, he says, inline detection, vision systems and barcode readers are common at most facilities. Ten years ago, they were still a novelty. Chatterjee notes the need for greater flexibility, particularly where shift changeovers are concerned. Mergers and acquisitions are driving manufacturing to greater agility, since many companies now must accommodate technologies rapidly, even with limited or no development understanding. He points to flu vaccine manufacturing, and the push to cell-based vaccines, as a success story. A few years ago, there was only one cell-based vaccine pilot. Today there are eight. “The market needed something, and we applied the innovation engine to figure out how to do it,” he says.
FDA enforcement and guidance trends are also having a positive, if painful, impact. For one thing, the Agency’s emphasis on science- and risk-based reviews and inspections will likely change the short-sighted emphasis that some companies give on “delivering quality to the regulator rather than the consumer,” says Gawayne Mahboubian-Jones, manager of Quality be Design at Philip Morris, and a frequent speaker and trainer at FDA staff training workshops.
Benefits of New Process Validation Guidance
FDA’s revised Process Validation Guidance will stimulate improvements in manufacturing, some say. “It has already increased the industry’s awareness of the need to take a lifecycle view of product quality and the importance of good science and process characterization and increased efforts,” says Blackwell.
Chatterjee notes that the guidance will allow manufacturers to take overall equipment effectiveness (OEE), SPC and Lean, and drive manufacturing equipment performance levels up far more rapidly. In addition, he says, it will facilitate the implementation of platform-based technologies. “This may be the most significant byproduct of the guidance, which is lost on those who remain fixated on the three stages,” he says. “If we can, during those stages, demonstrate that we not only understand and are monitoring those variables that affect our process, and setting limits around them, we can improve processes without having to revalidate everything; now we are able to make improvements to our manufacturing approach.”
Of special significance is the use of matrix validation, he says. “When you do this, you allow people to use more nimble or continuous manufacturing solutions, and qualify them for new or existing applications.”
Clueless on the Cost of Poor Quality?
One obstacle to more scientific manufacturing has been a persistent lack of understanding of the cost of quality. “Until the industry grasps the fact that this is an essential driver, and that the Taguchi Loss Function is not a digital (1 or 0) function, they cannot start down the road toward significant improvement,” says Mahboubian-Jones.
Blackwell agrees. “More efficient development and manufacturing come from a better understanding of the cost of quality throughout the product lifecycle, and the more efficient use of resources and technology to develop robust processes and to approach scaledown, platform technologies, and DOE,” he says. “Most pharma managers still don’t have a handle on the cost and risk of noncompliance.”
Chatterjee notes past progress in this area, in the late 1980s for instance with activity-based costing, which “came and went,” and with Lean accounting, which tied product costs not to work centers but to areas between them and allowed companies like Procter & Gamble to achieve quick and significant results. Today, he says, at most pharma companies, the financial buckets are not measured to the same magnitude. “If I have to make 100 lots of a drug to meet a market forecast this month and I only make 80 because the rest failed, those 20 are now accounted for in a different bucket. Nobody actually calculates the true loss, the true cost of those lots; they just track the scrap costs,” Chatterjee says. “The reality is that you lost that market share, that revenue opportunity,” he adds.
For OTC drugs, it can take six months to a year before you realize the impact. “If you’re trying to meet a minimum order for Costco, Walmart or Sam’s Club, if you don’t ship that number on that day in that quantity, you’ve lost that shelf space, possibly forever. How do you measure that, and its impact on your business revenue?”
That’s where cost of poor quality becomes a very important number,” he says. “Then, when you factor in the cost of deviation and CAPA investigation, sales and forecast inaccuracies, material on hold, inability to ship on time, the numbers are tremendous and it doesn’t take much to realize much of that gain back, but we’re not connecting these pieces together and considering what the true cost is of our inability to supply.”
Right First Time
The key to improving manufacturing lies in improving drug development. “We develop drugs using methods from the 1950s and 1960s,” says U.K.-based consultant Hedley Rees. “You need to scotch the idea that you discover drugs by accident by screening thousands of compounds and then getting into clinical as soon as possible. It’s only when we have hard clinical evidence in Phase III that people begin to think about how we actually make these products, and by then it’s too late.” The key, he says, is Quality by Design, but not in doing it as a scientific and technical activity but by reengineering the whole process, by involving all stakeholders at an early stage, not just R&D but CMC and procurement.
The current approach also limits pharma’s ability to connect manufacturing issues to impact on the patient, says PMI’s Hussain. “This is especially true,” he says, “if we only focus on the ‘mean values’ of attributes of a product used in clinical trials. Prior-knowledge, pilot studies, Phase I–Phase II studies are a means to reduce uncertainty on which product attributes are critical or not,” he says, “while structured experiments and databases can provide relevant information.”
“If you were to see side-by-side case studies of companies that had adopted QbD vs. those that didn’t, you’d see clear benefits in terms of the cost of poor quality, scrap rate, inability to meet market demand,” says Pharmatech’s Chatterjee. “The benefits are clear. But that kind of comparison is never done.”
The new FDA guidance on applying QbD for generic drugs is very much in line with what brand-name pharma should do and what generic companies don’t do, Chatterjee says. “If a generic company were to adopt three quarters of the methodology, not only would their bioequivalency lots match their scaleup lots the first time out of the box, but the variability they would see as they go to commercial manufacturing would be significantly less.”
Hussain sees the connection between IT for data, information and knowledge management and how it is integrated with Quality Management Systems to be critical to quality assurance and efficiency improvements in drug manufacturing. Pharmaceutical manufacturers have not been investing in the types of technology and IT required to better handle knowledge management and quality, says Raju, instead focusing on transactional systems. In the end, he says, this only encourages the documentation of 2.5-to-3-Sigma processes, rather than helping in transformation.
In addition, he says, industry may be reluctant to invest in technology that may reveal problems. Are regulators to blame for this? Since the last decade, FDA has invited companies to discuss such issues openly with them, and those who have report mixed results. Critics and surveys have suggested lack of coordination between reviewers and inspectors as a major source of problems. Recently, Raju’s company, Light Pharma, is working under contract with FDA to help improve those connections.
Strengthening those ties will be critical, says Hussain. “The PAT Guidance and PAT team, which united CMC review, cGMP compliance and inspection, recognized the need to reduce technical and regulatory uncertainty. At the moment, I can’t see such an integrated approach,” he says.
Since 2000, FDA has swung from “carrot” to “stick” mode. However, the end result will be positive, observers say, if it drives manufacturers to acquire better knowledge of their processes. Industry didn’t take the carrot [for instance, with PAT], so regulators are now using the stick, but this stick is based on product and process understanding, a much more effective weapon than the blunt “quality and compliance” approach they took in the past, says Mahboubian-Jones.
One ongoing challenge is the fact that manufacturing’s role is viewed as less of a priority than R&D and marketing. Although pharma is hiring more experts from other industries, says Sam Venugopal, consultant with PricewaterhouseCoopers in Cleveland, it is not likely that a CEO will come from the operations track any time soon.
“Companies need to understand that manufacturing skills and understanding are an essential component of the mix in a boardroom. Until that happens, decisions will continue to be made which look sensible on the balance sheet, but are disastrous on the factory floor,” says Mahboubian-Jones. The future of drug manufacturing belongs, he says, to small, flexible, forward-looking companies which create a base by filling niches deemed “uncompetitive” by Big Pharma. “Once this base is established, they will progressively out-compete the current generation of companies. There are clear signs that this process has started, and it seems to be most vibrant away from the conventional manufacturing centers in the Western world,” he says.
Overall, there is reason for optimism, he says. “Today’s positive examples suggest that pharma is capable of transforming its manufacturing and the quality of its products. The future has great potential. The question is: Will that potential be realized by the current generation of companies, or will it be a new generation of more flexible, more responsive, less autocratic companies which take up the gauntlet and bring about the transformation which is required?”