The deployment of Lean Labs and Six Sigma methodologies by biotech and pharmaceutical companies have delivered much needed improvements to an array of manufacturing and quality processes. From the receiving dock for raw materials to the final fill and finish of drug product, Biotech is rife with processes that could stand for optimization. However, during an entire manufacturing run, Quality Control groups must consistently keep pace; and, at times, set the pace for batch scheduling and production output while battling inherent limitations. From my experience working as a QC analyst, I think routine water bioburden testing, often overlooked for process change, is in desperate need of an OE improvement.
The standard practice of sampling a water system, filtering onto a membrane, incubating for up to five days and reading results creates a long and risky incubation time. Whether sampled at the NPCW source or WFI port in-suite, water is constantly being used at-risk, as results for today’s water won’t be known until well after today’s batch has moved far downstream, carrying with it any contamination.. This pattern is widely accepted as a norm, simply, “the nature of the beast”. I see a great opportunity to put the Six Sigma DMAIC (Define, Measure, Analyze, Improve, Control) methodology to use. We can assess and evaluate how automated rapid detection technology can improve this process.
Define. I have already touched on the general basics of the problem we are assessing. A standard water bioburden test, involving an inefficient 3 day or a 5 day incubation time, leads to an ever present at-risk status for water usage. The results read each day by a QC analyst only tell them that last week’s water was clean. Every day since then, including that actual day, is still questionable. Yet, manufacturing must still continue on. We would define an efficient process to be a shorter incubation time, yielding quicker results.
Measure. The specific metrics behind measuring this problem can be identified from many departments across a manufacturing site. For example, Manufacturing and Operations metrics include holding costs for final product, while Engineering groups focus on wait time of manufacturing as water systems are released after shutdown or maintenance operations. The QC department measures materials and manpower costs needed to plate and incubate samples, read results and perform any needed emergency follow-up assays pending a hit.
Analyze. Analysis of these metrics would point to a root cause being the 3 to 5 day incubation time. But why must we incubate this long? . First, we can blame Mother Nature. Water borne microorganisms found in pharma water, such as WFI, are highly stressed. Subjected to trauma like line sterilizations and an extremely low nutrient water environment, they may have adapted in various ways by producing shock proteins, sporulation, or even just a entering a basic vegetative state. Unfortunately, when filtered and plated onto a nutrient rich medium agar, there is an unavoidable lag phase, where these microorganisms must wake up and start up normal cell processes again to begin forming colonies. This is the underlying Catch-22 to QC Microbiology. The lag time for waterborne contaminants to form colonies inherently dictates the second need for a long incubation time, the ability of the analyst to count colonies. Now, this is no stab at an analyst’s ability to count, I used to be one (although my skills in calculus always left something to be desired) but let’s face it, stressed bugs form stressed, smaller, slower colonies. What may be there after the first day may not even be perceptible by the human eye until the third or fourth day. We have to extend incubation times to allow the colonies to merely become visible. Also, the limitations of the human eye may lead to analyst error in counting numerous colonies and speed at which they can read plates.
Improve. So how do we improve the process? Mores so, how do we solve our defined problem and shorten incubation time? Do we develop and then shamelessly market a super nutrient-rich growth medium that microbes just can’t resist? One that they will make them grow faster, jump higher, and hit more free-throws (It’s What Bugs Crave!) Or should all analysts be required to wear eye glasses? Do we have an on-site Lasik Surgery Day even? Of course not. Frankly, if the problem of long incubation times is a result of an uncontrollable characteristic of microbes growing at their own pace, we must be able to detect them sooner. This is where the utilization of rapid automated microbial detection and enumeration is an answer, if not the only answer. Rapid detection technology can see far beyond the limitations of the human eye and can detect even the smallest of slow growers well before even the best eyes money can buy. High speed imagers can also count colonies faster than an analyst would, in some cases counting many plates in the time it may take an analyst to count just one. In conjunction with counting faster, imagers can count far more; further extending how we determine what exactly is “TNTC” or Too Numerous To Count.
Control. Routine bioburden testing is an industry wide staple assay; however, as most QC assays do, is always subject to variation through analyst error. Changing over to a rapid detection technology, moreover, an automated rapid detection technology like the Growth Direct, stands before a wall of red tape. Validation and turn-over of technology must be fast and concise. The system must be capable of delivering repeatable time to detection to return the investment made to alleviate the long incubation time. Furthermore, it should deliver consistent accuracy while maintaining throughput. Big shoes to fill, but the Growth Direct does fit these requirements. I think further process capability studies within the industry will continue to affirm this.
Lean Labs and Six Sigma processes are helping us push the envelope in an already dynamic and progressive industry by utilizing new and rapid technologies. Although this was a very brief and general usage of Six Sigma methodology on this topic, I believe the principles still apply. Now, if I only I could think of a name for that super tasty agar growth medium… I’ve got it, Microbade!
Research Assistant, R&D – Microbiology
Rapid Micro Biosystems.