Proactive Compliance: Putting the "P" in CAPA

Pharmaceutical companies invest plenty of time and resources dealing with problems that occur in the manufacturing process once they happen. But how much effort are they making to actually prevent quality issues from occurring in the first place?

With much-publicized quality lapses surfacing at some big drug companies, it’s no wonder that pharmaceutical manufacturers are concentrating more than ever on the corrective and preventive action (CAPA) process and the systems they use to manage these events. At the same time, there is a growing emphasis on continuous improvement to help integrate quality controls into manufacturing processes, thus preventing problems before they occur. 

Nearly all drug firms have a corrective and preventive action process and system that ultimately strives to improve manufacturing processes, reduce downtime and eliminate regulatory complaints. Many, however, view their CAPA responsibility as a pure regulatory concern — in other words, identify the root cause of the problem, fix it, find a way to prevent it from happening again, and get on with business. It’s only been in recent years that drug companies have focused on making broader, more sweeping gains aimed at improving quality — what consultants refer to as “continuous improvement.”

A key reason many pharmaceutical firms have been reluctant to make wholesale changes on the plant floor is that the amended processes must then undergo recertification, often a time-consuming ordeal in itself. “There are some difficulties for pharmaceutical companies to really change their processes, because they have to go through the recertification process,” says Catherine Converset, a partner at Productivity Group Europe in Paris, where she has helped pharmaceutical industry clients with continuous improvement and lean manufacturing initiatives for more than a dozen years. 

“Part of the challenge here is the regulatory concern,” agrees K. R. Karu, industry solutions director for the pharmaceutical industry at Sparta Systems, a quality management and CAPA software vendor.

“Pharmaceutical companies tend to put handcuffs on themselves. They are resistant to change because they are afraid the change is not going to be compliant.”

Those concerns are not unfounded. Every time pharmaceutical companies make substantive changes in processes, procedures or systems — including just about every action workers perform on the plant floor — they must have those changes approved by regulators. The smallest change in the way a product is produced triggers yet another round of FDA review and authorization. It’s no surprise, then, that some manufacturers are wary of making major changes in their manufacturing processes unless forced to do so.

Converset says drug companies have been trying to adopt continuous improvement methods for at least a decade, but only recently have they focused on improving their processes and process controls. “They are doing a lot of work today on quality and on how to better monitor their CAPA process to be more efficient when handling a CAPA event,” she adds. “The pharmaceutical industry understands that it has to make a lot of changes, and it is now really challenged to make continuous improvement efforts.”

An impending challenge for the industry, Converset says, is the need for drug manufacturers to switch from batch process production to continuous flow manufacturing. “Batch production is a structural limitation for the industry,” she says. “The move to continuous flow requires a lot of changes, and it requires that companies have the ability to closely monitor their processes.”

She recommends that companies embrace the concept of quality by design. This requires that companies carefully define and monitor the parameters of the manufacturing process so that drug quality is guaranteed. “Quality is built into the design of the process, but you need to constantly monitor the process,” she says. 

“We see some continuous improvement that is directly related to CAPA events,” Converset observes. “But a lot of continuous improvement is aimed at reducing cost and lead time, reducing inventory and reducing errors.”

Admittedly, the pharma industry has embraced the concepts of continuous improvement and lean manufacturing principles later than some other industrial sectors, most notably the automotive industry. “Pharmaceutical companies have come to the lean party a little later than other manufacturers, but they are getting there,” Karu says. 

Of course, many pharmaceutical manufacturers have extensive controls in place to enable them to closely monitor processes and receive alerts when something in a process starts to go out of specification.

“Pharmaceutical manufacturers leverage systems that offer thresholds and alert mechanisms to warn when the process is moving outside the control limits, so that the appropriate actions can be put into place,” says Deborah Kacera, regulatory and industry strategist at Pilgrim Software, a vendor of risk, compliance, quality management and CAPA systems. 

Kacera says the annual product review undertaken by drug manufacturers typically will point to potential problem areas where preventive actions should be undertaken on a proactive basis. “The annual product review is an example of looking at all the data from various sources — from investigations of root cause analysis, batch record reviews, change control reviews and other nonconformance reviews to see all the data systematically,” she says. “This review consists of the ongoing vigilance, trending and aggregating of nonconformities to determine if there is a pattern emerging that can be intervened. This type of information and its analysis can lead to preventive actions that will improve both product and process control.”

Pharmaceutical firms also can take preventive action proactively by improving training for employees on the plant floor. “Training is an essential element of a compliant workforce and a regulatory requirement,” Kacera asserts. “It is critical for workers to be certified in the areas of best practices, the standard operating procedures they are responsible for using, product information, problem solving, and whatever else is important based on their role and responsibilities,” she adds. “Ongoing current good manufacturing practices training with effectiveness testing is an essential part of the prevention equation.”

Complicating manufacturers’ efforts to take preventive action is the fact that CAPA itself is not clearly understood within the industry. As a result, companies tend to interpret it in their own way. “The first thing I do when I meet with pharmaceutical company executives is to ask them to define CAPA,” Sparta’s Karu says. “The reason is that there are so many different ideas of what it is.” 

This relative lack of a clear definition, in turn, makes the definition of what constitutes a preventive action equally unclear. According to the U.S. Food and Drug Administration, which regulates and enforces CAPA procedures for the industry, corrective actions are those performed in response to an observed problem, while preventive actions are taken to prevent such problems. 

But even the FDA recognizes that confusion exists within the industry over the difference between preventive and corrective actions. “While industry typically has procedures that specify corrective and preventive actions taken in response to a failure of some other unexpected event, these ‘preventive’ actions are really only corrective actions in response to the event,” says Francis Godwin, Acting Branch Chief of the Generic Drug Manufacturing Assessment Branch at the FDA. “A truly preventive action is one taken to improve a system or process before any disruptive event occurs, and thus is not a reactionary action. 

“A preventive action is an action taken by a firm to prevent an occurrence of a nonconformity or undesirable situation,” Godwin adds. “A corrective action is an action to eliminate the cause of an undesirable situation that has been observed, and to ensure it does not recur.” Godwin says the specific definitions appear in the glossary of the ICH “Guidance for Industry, Q10 Pharmaceutical Quality System,” which can be found at http://www.fda.gov/downloads/Drugs/Guidances/ucm073517.pdf .

To sum up, a true preventive action focuses solely on reducing the possibility of things going wrong, without any connection to a prior manufacturing inconsistency or other problem. Godwin says preventive actions can be triggered by a variety of things, including continuous improvement initiatives, periodic product reviews, updated risk assessments, trend analyses and other gains in knowledge.  

“Preventive action is when you don’t know what could go wrong,” concurs Pilgrim Software’s Kacera. Nonetheless, she adds, once such a possibility has been identified, the company can apply the CAPA process to reduce or eliminate the likelihood of its occurrence. “After the annual product review, if there could be a change made to a product specification, then creating a preventive action CAPA is a way to help understand what has changed and what can be done to mitigate the problem,” Kacera says. 

Once the root cause of a problem has been identified and corrected, says Sparta Systems’ Karu, there is the opportunity for preventive action to follow. “The preventive part is twofold: first, to take steps to keep this from reoccurring at this site, and second, to be proactive and make the same improvements at any other locations where you use that same piece of equipment or process,” he says. 

“In my view, the biggest way companies can prevent manufacturing problems and do continuous improvement is to apply the lessons they learn across the enterprise,” Karu adds. “This is about tribal knowledge versus institutional knowledge — in other words, companies need to ask themselves where else could they possibly have the same circumstances arise?”

Karu offered the experience of a large pharmaceutical firm with 50 manufacturing plants worldwide. “The customer put a global CAPA system in place, and within a three-year period, the company reduced its manufacturing deviations by 70%.”

It’s exactly those kinds of longer-term process gains that CAPA is designed to foster, the FDA’s Godwin says. “Both the corrective and the preventive portion utilize increasing product and process knowledge to continually improve the process: corrective to address events as they arise; preventive to take action before they arise,” he explains. “It is in the manufacturer’s own interest to continually improve their processes, even if the process has yet to yield an irregularity or undesirable event.”

One drug firm that is using a CAPA system to enable better management of manufacturing issues is Pharmaceutical Associates Inc. (PAI). The company manufactures and distributes a variety of liquid pharmaceuticals, including contract-manufactured liquid products, private-label formulations, and PAI’s own line of generic pharmaceutical liquids. In addition to the company’s 300,000-square-foot manufacturing plant in Greenville, S.C., PAI is building a new state-of-the-art facility on the same site.  

PAI is using Pilgrim Software’s CAPA and complaints management system to track and trend exceptions, says Mark Wita, corporate vice president for Quality Assurance and Regulatory Affairs. “Before we had numerous standard operating procedures for investigations or inspections — nothing was uniform,” he says. “Everyone had their own spreadsheets for handling these issues. The problem was how to handle complaints in a uniform manner. Pilgrim unifies all the different areas involved with potential problems in your company.” 

In this case, the CAPA software is used to monitor only those issues that have been noticed and entered into the system as a result of some complaint or quality issue that was brought to the company’s attention. At PAI, for example, those events are entered into the system by plant supervisors, quality assurance staff and other managers.  

PAI’s approach to handling a CAPA event is typical. “First, we define the problem and list the most likely probable cause,” Wita says. “Next, we check to see what we did the last time if we had a similar issue. If not, we may have to do some experimentation to see what might have caused the problem. Then we assign a root cause. Typically we find that the cause of errors or anomalies can be traced to one of four sources — man, machine, material or method. 

“Next we take corrective and preventive action. This could mean telling the vendor to replace the material. Or we may need to change the inspection process, or make changes to the SOP or batch record. Then we train our people on what needs to be done to bring about those changes. And we continue to monitor the process to make sure it doesn’t happen again.”

Wita says he has heard workers claim that a certain manufacturing anomaly was just an “isolated” event. “But when it happens three times in a year, it can’t be an isolated event,” he adds. 

By closely examining the aggregated data the CAPA system puts together on nonconformances, Wita says he is able to spot potential trouble areas where the company needs to take action early, instead of waiting for something to go wrong. “You look at the process as a whole and look for trends,” he says. “Then you make a judgment on whether something needs to be corrected. Finally, you monitor the process and review the data.”

Despite the apparent confusion that exists within the industry over the difference between a corrective and a preventive action, Wita says PAI has been able to harness its CAPA system to help it prevent problems that have already occurred as well as prevent some that could possibly arise down the road. “For us, it’s both,” he explains. “You can predict some problems and prevent them, but there are some unforeseen problems that are going to crop up.” 

As an example, while conducting PAI’s annual review process, Wita says the company has identified areas where processes can be improved on a proactive, rather than reactive, basis. “With our annual review, we identified some improvements we could make proactively to our processes.” Of course, he adds, “If you make a change, you might have to revalidate that process.” 

PAI also is able to take preventive action by leveraging in-process monitoring to detect equipment that may be starting to go out of specification. “We do in-process monitoring to make sure our processes are in control,” he says. “Process controls should be well within the design controls for quality, so that you aren’t drifting outside your regulatory requirements. If you see drift, it’s telling you there’s something wrong.”

Yet another way for pharmaceutical firms to apply continuous improvement to prevent manufacturing errors is through improved training of employees, especially when it comes to overcoming employee attitudes that can impact quality. “Good training can overcome any kind of cultural issues you may have,” he says. In other words, he says, employees should be trained to believe that, as he puts it, “If you see something, say something. You want to train people not to be afraid to bring forward any quality problems.”

Human error due to poor training, shoddy adherence to work standards, or carelessness on the job often are found to be the root causes of manufacturing deviations or other quality problems. “If you need to do three CAPAs related to weighing problems, then you may have a training problem,” Wita adds. 

PAI tracks its quarterly error or exception rate based on type of cause. “We then compare it to the same rate for those causes in previous years, and finally to the quantity of product produced,” Wita says. “We look for a number of repeat occurrences that is above the norm for that cause. 

“We also track what shift the errors occurred on,” he continues. “So if you see that the second shift had more errors, then you want to ask what is different — for instance, is there new management on that shift?”

Problems also can be caused or related to time of year, he says, because some types of problems are more sensitive to cold or hot weather. 

A useful feature of the CAPA system is its ability to track the level of risk according to the type of incident. “When you are entering an exception into the Pilgrim system, you assign it a risk of high, medium or low. This enables us to subdivide CAPA events by risk level and by cause.”

PAI’s new plant will, by design, push the company’s quality levels still higher, Wita believes. “The new plant will do in two shifts what we currently do in two shifts and a weekend shift, and in four manufacturing tanks it will produce what we currently require 30 tanks to produce,” he says. “The new equipment will be computer-controlled to continuously monitor various aspects of each batch, ensuring a more safe, uniform product,” Wita adds.