The late aerospace engineer Edward Murphy is often quoted as saying, “If there’s more than one way to do a job, and one of those ways will result in disaster, then somebody will do it that way.” So much so that the infamous “law” was attributed to him: Anything that can go wrong, will—at the worst possible moment.
Engineers are trained to take a “worst case” approach to problems, then to design safeguards or preventive measures. “What if the maximum weight put on the bridge is exceeded by a factor of five?” “What if the reactor core isn’t sufficiently cooled?” Or, more recently, “What if a plane were flown at top speed into this structure?” Unfortunately, pharmaceutical scientists haven’t often received this training. Neither have those who run QA and QC labs.
For many of them, it’s all about perfection: optimizing theories, models and techniques. Failure is not an option. However, the industry faces some very different threats today. For instance, sophisticated drug counterfeiters who add adulterants to “fool” analytical instruments so that “products” pass quality tests. Given the economic and safety risks involved in each pharmaceutical industry decision today, more non-engineers in the industry are also learning to take a “worst case” view, and use engineering tools to manage risks, on the plant floor and in the lab.
One of the most powerful tools to help in this daunting task is Failure Modes and Effects Analysis (FMEA), whose roots are in the aerospace industry. The technique is designed to help determine all the possible ways a process or product can fail, rank these risks and ensure that development and design minimize the likelihood of failure. FDA included sections on FMEA in its 2006 guidance on Quality Systems.
We hear that pharmaceutical companies have been using FMEA in various applications, and usually with great results. We’ve even been able to publish some of these cases. But is the industry missing opportunities to apply the concept more broadly, or “out of the box,” and to harness risk data proactively in its everyday operations? What if part of the FMEA could be automated and if systems could pull data automatically from asset management systems, from batch recordkeeping software, LIMS and data historians?
What if records of past problems weren’t filed away but kept as living documents, and used throughout manufacturing and development operations. FMEA requires cross-functional teams working together to rank the risks, improving communication between them. As the technique were more broadly applied, silos would disappear. Such data would routinely become part of any new drug application (NDA). Consider the heparin recall.
An international team of scientists applied risk management techniques to identify the contaminant that led to the recall. Given the status quo, they had to create development data for a drug that was commercialized decades ago, and concoct “worst case” scenarios that weren’t even considered in the past. What if this information had been gathered during development?
FMEA has even been applied to assess outsourcing risks, an approach used by one electronics supplier to determine whether or not to source parts from Asia. (See “Using FMEA to Assess Outsourcing Risk,” in the August 2007 issue of ASQ’s Quality Progress for a technique that might be adapted to pharmaceutical situations.) IT tools are becoming available, and more drug companies are already trying to use FMEA proactively. Write in and let us know how you’re using the technique.
*For any trivia buffs, Bloch was not a quality systems thinker or pal of Deming’s; he wrote the novel that inspired Hitchcock’s film, Psycho.
