The see vs. solve challenge

Manufacturers are continuously working to improve processes for better yields, faster time to market and fewer regulatory investigations. Aegis Analytical’s Randy Tatlock, who has more than 27 years of experience working for process manufacturers, provides insight into challenges and technology solutions.

EMI approach for Proactive Problem Solving

“In manufacturing, we work on what’s screaming the loudest right now. We may have different people doing the 10 investigations, so it’s sometimes hard to recognize the amount of time wasted. Once you add up that time, it becomes clear that it’s worth investing some extra time with the help of an investigational tool to eliminate a problem versus treat its symptoms”

Why is problem solving difficult today for life sciences manufacturers?
The biggest challenge when manufacturers face a process deviation is that the data needed to solve problems is not easily accessible, and it is not available in a form that is easily used. A second challenge is that in manufacturing, we are not in the habit of proactive problem solving. We tend to be distracted by other priorities and focus more on after-the-fact fixes so we can move on to the next pressing task. Clearly the better approach is to examine a problem’s root cause and eliminate that setback, so that it doesn’t reoccur in future batches.

How can technology solutions help?
The reality for manufacturers is a time-crunched day-to-day world in which stopping to examine the root cause of a problem is a luxury. Software solutions that access and aggregate manufacturing data in a way that enables analysis are extremely helpful, because they save time that would typically be spent by users on the manufacturing side or IT.

Tools like Aegis Analytical’s Discoverant help you actually solve a problem versus simply seeing that a problem exists. This software provides on-demand access to data from disparate sources throughout the manufacturing enterprise – such as paper records, MES, LIMS, historians, DCS, Oracle databases and data warehouses in one validatable environment. Other solutions exist, but data access is isolated to specific systems. A valuable software solution not only provides users with access, but it fuses data in a way that allows configuration and analysis that is most useful for problem solving versus simply seeing information on a computer dashboard.

What kinds of issues do manufacturers typically see but not solve?
I think every manufacturer gets caught up in a world of treating the symptom and not the root cause. For example, consider a process that has a deviation in temperature, pressure, or flow rate and the reason is unknown. You investigate and find out that batch of material is okay to use because, although it violated a particular parameter specification, it meets acceptance criteria. Because of pressing priorities of the day, you see the problem, you move on, and you never solve the larger problem. So two months later you’ll likely have the same situation. Taking some time and using analytics with a tool allows you to identify the root cause and solve the problem, so that it doesn’t happen again.

What risks are associated with not solving such problems?
By letting problems continue, you really face product, supply chain and regulatory compliance risks. Recurring deviations are a systemic problem because you’re continuing to use the same resources and keeping batches on hold when the products are actually viable. You’re holding up the supply chain and spending additional man hours each time to investigate and release batches that could be more predictable. It’s easy to see how this unnecessarily keeps an organization on hold, wasting time and money. It also puts the manufacturer at risk, because the next time, parameters could very likely be over the acceptable edge and waste the batch because of the same root cause. It could also expose the organization to additional regulatory scrutiny based on the recurring deviations.

In reality, if you could stop and spend additional time using the right analytical tools to find the root cause and permanently solve the problem, you would save the additional hours wasted each time the deviation reoccurs. And even better – you can set up a system like Discoverant to proactively identify problems for you by providing the Enterprise Manufacturing Intelligence that helps overall quality and predictability.

What is Enterprise Manufacturing Intelligence?
EMI is a trend of viewing an organization’s information holistically. We can think of a pharmaceutical enterprise as all of its sites and information from top to bottom – everything that goes into its processes from raw materials to the final outcome. I like to call it a “supply chain of data or information.” Data flows from rudimentary details (such as vendors of materials and tank temperatures) all the way up to the output of the final product and its quality. Like a supply chain, you have to manage everything from beginning to end so that it’s effective all along the way. You have to be able to track and observe information from beginning to end of all processes.

When we define intelligence, it can be either a dashboard type view of information or a more analytical view of data that leads us to greater insight. It is the intelligence factor in EMI that lets us solve versus see problems. Intelligence is more actionable than pure information, because we have considered multiple data, thought about what was going on and come up with a definitive conclusion that can be utilized for improvement.

What qualities are critical considerations for EMI tools?
Configurability is the key to giving users data they can use for intelligence. There are three requirements for an EMI solution that will truly lead to problem solving versus simply seeing issues.

The first is enabling you to access data from all sources – going back to the holistic view of the enterprise, including all of its processes, sites and batches. Many solutions on the market today are integrated with a specific source product, such as a data historian. It has a process control function to store continuous and process control data, but it is not built to efficiently get other sources of data and provide analysis on it. A valuable EMI solution must be “source system agnostic,” so it is all inclusive and can objectively represent all relevant data in the enterprise.

A second requirement is providing self-service to data from source systems. Many solutions are closely tied to source data, such as lab information, and depend on the user knowing what he or she is looking for to make an investigative query. A better EMI solution will have a virtual data warehouse – providing data in combinations that are not hard coded. Self service is especially important because it doesn’t require IT intervention. Users can get what they need on their own without burdening IT for each data request. Quick turn-around problem solving is paramount to achieving maximum organizational benefit.

A third requirement tied to configurability is the existence of signals that can be defined by users to flag an issue. The right EMI tool will let a user set desired critical process parameters and proactively look at results outside of the set parameters to signal they are potentially going to have a problem – or reactively to examine a problem. An effective EMI solution lets users drill down into the details of the flagged problem to investigate the root cause to solve it. Using multiple systems tied to disparate data sources would require much more time and would require the systems to be in synch with one another to allow investigation. It is much simpler and requires less maintenance to use EMI software that fuses all sources of data into a single portal.

What else can manufacturers do to create a problem solving environment?
It is helpful for manufacturers to assess how much time they are actually spending – and often wasting – on investigations. Do they have good handle on what deviations are really recurring? Could they go back and see that they’ve worked on the same problem 10 times in one year – translating to supply chain disruptions and wasted time? In manufacturing, we work on what’s screaming the loudest right now. We may have different people doing the 10 investigations, so it’s sometimes hard to recognize the amount of time wasted. Once you add up that time, it becomes clear that it’s worth investing some extra time with the help of an investigational tool to eliminate a problem versus treat its symptoms.

You can even use an EMI tool to create a Pareto chart showing the highest impact problems to allow you to better prioritize. You could, for example, configure a system to connect to a Corrective and Preventative Action (CAPA) system. You could then analyze at a dashboard level all deviations and find that 100 of those were caused by processing variations. If you can prove with data that 75 of those deviations were caused by a common part of the process, then you can solve that particular problem by further using the EMI tool to identify the root cause and eliminate it. Ultimately, you would have solved a systemic problem that a CAPA system alone wouldn’t have showed you. The EMI solution provides the enterprise, holistic view that leads to actionable solutions.