You get the call on a Tuesday. Environmental monitoring results are back, and a swab from the slicing line came back positive for Listeria. Not on product. Nobody is sick. But the clock is running, and every decision you make in the next hour will either solve the problem or push it down the road.
What happens next separates programs that genuinely protect plants from programs that produce paperwork.
The remediation trap (and why smart teams fall into it)
Here's the honest version of what happens in most plants: when a positive comes in, the first call goes to sanitation. Intensify cleaning. Get three negatives. Resume production.
The pull toward sanitation first is understandable. The line is down, the plant manager needs a timeline, and the customer has a delivery window; so intensifying cleaning feels like the responsible move in the moment. Investigation-first thinking sounds right in a training room but feels impractical at 6 AM with a scheduler standing at your door.
The problem is that remediation-first responses don't change conditions. They confirm that your sanitation can suppress contamination temporarily. A few weeks later, the same site lights up again. As Steve Tsuyuki, who spent decades building environmental monitoring programs across more than twenty facilities at Maple Leaf Foods (including through the 2008 Listeriosis outbreak), put it during a recent FSMA Friday session: it's whack-a-mole.
The counterintuitive truth is that investigation-first is actually faster dealing with the second or third occurrence of the same positive site. When you find and fix the actual source the first time, you're not standing at the same drain six weeks later explaining to your operations VP why the line is down again.
The Listeria control equation: it's not just a sanitation problem
Before any productive seek-and-destroy response is possible, you need a shared framework for what you're actually controlling. A framework that emerged from the AMI Listeria Control workshop presentation team in the late 1990’s, and that Tsuyuki still uses today, identifies five controllable factors:
- Controlled traffic patterns: segregation of raw and RTE zones, and the movement of people, equipment, and product between the zones.
- Good manufacturing practices: personal hygiene, glove use, handwashing, employee behaviors and housekeeping during production
- Infrastructure integrity: floors, walls, ceilings, drains, standing water, and weeping or dripping water
- Sanitary equipment design: ability to clean and access all surfaces; managing wear and tear from repeated cycles of production and sanitation
- Effective sanitation: routine cleaning AND non-daily periodic tasks
Look at that list carefully. Only the last item is owned exclusively by sanitation. Infrastructure belongs to maintenance. Equipment design involves engineering and procurement. The top three are largely operational responsibilities that line supervision and plant leadership control day to day.
This is why FSQA can't run a seek-and-destroy investigation alone. When Tsuyuki's teams investigated positives across Maple Leaf's network, the effective responses always involved operations, maintenance, sanitation, and FSQA together. And always led by the plant manager..
Under 21 CFR Part 117, environmental monitoring is a verification activity, and not a control measure. It tells you whether your existing controls are working. When a positive turns up, one or more controls have already failed. The swab just revealed it. The real work is figuring out which one failed and why. This is the same logic your food safety plan and HACCP system already run on: monitoring verifies that controls are working, and a deviation tells you one of them didn't.
What good "seek" actually looks like on the floor
When a positive comes in, most practitioners want to go straight to remediation planning. Tsuyuki recommends something different for the first step: go to the floor and watch.
Before anyone picks up a swab, gather your cross-functional team and observe the affected line during production. You're looking for anything that has changed, anything unusual, anything operators experience every day but managers rarely see.
Ask these questions during that initial observation:
- Where is water accumulating, and where is it coming from?
- What do GMPs look like in practice versus on paper? (Check your GMP standards against what you actually see.)
- What recent changes have occurred: equipment moves, new products, line reconfigurations, staffing changes?
- What challenges is the sanitation crew dealing with that don't show up in records?
- What do frontline operators know that nobody has asked them about?
That last question deserves real weight. If a conveyor gets moved in and out depending on the product run, or a drain backs up intermittently, or a floor section deflects water toward a specific zone, the people working that line every day know. They just haven't been asked in the right context.
Once you've done the observation and gathered that input, your team can identify two or three plausible causes, build a hypothesis-driven sampling plan, and start swabbing with purpose rather than pattern.
Purposeful sampling: investigative vs. routine
Investigative swabbing for seek-and-destroy looks nothing like your routine environmental monitoring program. Routine monitoring is scheduled, systematic, designed to verify your program over time. Investigative sampling is targeted and hypothesis-driven, designed to catch contamination in motion.
Two approaches are worth knowing:
Time-series sampling involves swabbing the same area repeatedly at defined intervals during a production run or sanitation cycle. Instead of a static snapshot, you're tracking movement: where does the organism first appear, and how does it spread? In a case Tsuyuki described, time-series sampling on a slicing hall completely reframed the investigation. The team initially hypothesized harborage in the equipment. The data showed the positive appearing consistently at the point where meat logs entered the hall, on the casings, and first on the hands of workers peeling those casings. The source was incoming raw material. Without time-series sampling, they'd have kept cleaning the wrong things.
Post-rinse or rinsate sampling uses gravity and the sanitation cycle to reveal internal harborage. Following a regular sanitation process and immediately after the final rinse step, you wait. Internal components (bearings, enclosures, equipment cavities) continue to weep water long after visible surfaces are clean and rinse water has gone to the drains. Swabbing equipment undersides or equipment legs where the leg meets the floor, and areas where dripping water collects can reveal contamination that routine surface swabbing never reaches. In one documented case, this technique found meat debris and contamination trapped on an electrical enclosure seal, a location completely inaccessible to routine sanitation and swabbing.
Neither of these runs as a solo FSQA exercise. Both require production, maintenance, sanitation, and FSQA to coordinate around them. They require a collaborative and genuine investigation culture, not just filling out a corrective action form. This is where your root cause analysis process either earns its value or exposes its gaps.
The drain problem: where you look vs. where it starts
Drains are one of the most revealing indicators of whether a program is genuinely investigation-oriented or reflexively remediation-focused.
When a drain positive turns up, the easy read is that the drain is the source. Clean it, confirm negatives, move on. But drains are usually the destination, not the origin. Water comes from somewhere. Listeria travels to the drain because water is carrying it there.
One technique Tsuyuki uses: temporarily cap the drain, then swab the surrounding area. If contamination is still present, it's coming from somewhere else. In every case where he's applied this, the drain itself was not the source.
In the whack-a-mole case he described, the actual source was a small jump conveyor moved in and out depending on the product run. When investigators finally swabbed inside the electrical cabinet of that conveyor, they found Listeria. Not the drain. Not the floor. A mobile piece of equipment harboring contamination. No amount of drain cleaning was going to solve that.
Design out vs. manage through procedures
When seek-and-destroy works correctly and you've found the actual source, the corrective action takes one of two forms: design out the problem, or manage it through procedure.
Procedures require training, ongoing compliance monitoring, and verification. Human behavior varies. Even well-trained employees have bad shifts. Procedures are necessary, but they carry inherent drift risk.
Design-based solutions remove the human variable entirely. Tsuyuki's analogy: a sign saying "no skateboarding" requires enforcement. Metal pegs spaced to prevent skateboarding require nothing. In food manufacturing, the equivalents include replacing hollow conveyor rollers with solid ones, installing sloped floors that eliminate standing water, upgrading to monolithic or crack-free flooring surfaces that eliminate harborage sites, or redesigning equipment to eliminate internal cavities that trap soil.
When a design-based correction is available and feasible, it's almost always more durable. And once you've made a design improvement, the discipline is to never let it be reversed. Never allow a designed-out risk to become a procedurally managed one. For a look at how this kind of continuous improvement thinking applies in food manufacturing, the principles carry directly.
If your EMP data lives in binders, you're investigating with one hand behind your back
Some teams are building this investigation infrastructure digitally. Here's what that looks like in practice.
The investigation process Tsuyuki describes runs on data: time-series results that need to be mapped across shifts, corrective actions that need to trace back to specific swab results, trending that needs to surface recurring sites before the third or fourth positive forces the conversation. When that data lives in spreadsheets and paper logs, the investigation is slower, harder to coordinate, and nearly impossible to defend at audit.
Digital environmental monitoring programs change that. When swab locations, results, and photos are captured at the time of collection, that data is available without pulling spreadsheets or waiting for someone to compile it manually. When a positive triggers a CAPA workflow, that record captures not just the result and the response, but the root cause documentation that answers an auditor's third question: not "what did you do?" but "why did it happen and what changed?" When results are centralized and trended across facilities, patterns that might take weeks to surface in a spreadsheet can become visible across facilities when data is centralized and trended.
When an environmental positive triggers a configured workflow, your cross-functional team gets notified and an investigation checklist generates automatically. That's not a feature description; that's what replaces the phone tree at 2 AM.
No software replaces the investigative judgment, the cross-functional culture, or the interpretive experience that practitioners like Tsuyuki bring to an environmental positive. But if your EMP data isn't centralized and traceable, you're making that judgment call with incomplete information.
Three questions to assess where your program actually stands
Before you leave this page, run through these:
- What happens in the first hour after you receive an environmental positive? Is the first response an observation walk with a cross-functional team, or a call to schedule additional sanitation? The answer tells you your program's default orientation.
- How do you decide what to swab during an investigation? Is investigative sampling driven by hypotheses about contamination sources, or by your routine sampling map with more sites added? Those serve different purposes and require different designs.
- When a corrective action closes, what evidence exists that the root cause was addressed? If the CAPA says "intensified sanitation, three negatives confirmed," the next question is: what was the source, what was the transfer vector, and what changed? If those answers aren't in the record, the seek part of seek-and-destroy wasn't completed. A well-structured CAPA process captures that documentation, not just the closure date.
Environmental monitoring is valuable because it creates the opportunity to catch contamination in the environment before it reaches finished product. That value is only realized when positives trigger genuine investigation.
Watch the full session with Steve Tsuyuki
This piece covers the framework. The FSMA Friday session goes deeper: specific investigation protocols, Tsuyuki's case examples from Maple Leaf, the full Listeria control equation in practice, and how to build an investigation culture in a plant where production pressure is real.
