Illustrative case study. Company names and identifying details are fictional. Technical details reflect real industry practice and current regulation.

Incubation testing — the cannery’s verification, holding sample cans at elevated temperature and checking for swelling — is the industry’s last line of defense: the test that catches the underprocessing, the leakers, the contamination that every other control missed. At Blue Ridge Canning, the incubation testing was catching nothing — because the incubation temperature was wrong: the incubator, set for 35°C (the thermophilic screen), was actually running at 28°C (a thermostat fault, unnoticed for a year), and the mesophilic spoilage — the organisms that grow at 28°C anyway — was being caught, while the thermophilic indicators (the underprocessing alarms) were incubating below their growth temperature. The verification was running, the samples were held, and the test was blind to the failures it existed to find.

Background: a cannery running routine incubation

Blue Ridge Canning (fictional) canned about 60,000 cases of vegetables a month — 65 employees — the incubation testing: sample cans from every lot, incubated at 35°C for 10 days (the thermophilic screen — the underprocessing indicator), checked for swelling. The program was routine, the records complete, the results consistently negative — “no incubation failures in over a year.” The incubator — the equipment that made the testing meaningful — hadn’t been calibrated in two years.

Challenge: the failures that weren’t

A customer complaint — swollen cans, thermophilic spoilage (Geobacillus, the underprocessing indicator) — in a lot that had passed incubation testing: the sample cans, incubated and checked, showed no swelling — but the production cans, in the customer’s warmer warehouse, swelled. The investigation checked the incubator: 28°C, not 35°C — the thermostat had drifted (the fault, gradual, unnoticed), and the incubation — the verification — was running 7°C cool. The thermophiles (Geobacillus, growing optimally at 55°C, minimally at 35°C) — didn’t grow at 28°C: the test couldn’t detect them, because the temperature was below their growth range.

Investigation: the cool incubator

The incubator’s thermostat — examined during the investigation — had drifted gradually over a year: the setpoint said 35°C, the actual was 28°C, and the drift was invisible — because the incubator wasn’t monitored (no independent thermometer, no calibration, no verification). The incubation records — “35°C, 10 days, no swelling” — documented the setpoint, not the actual: the verification was verifying the paperwork, not the temperature. The year of “no failures” — the program’s proud record — was the year the test couldn’t fail, because the temperature was too low for the indicators to grow.

The thermophile biology made the 7°C critical: Geobacillus stearothermophilus — the classic underprocessing indicator, the organism whose survival means the retort was inadequate — grows at 35–75°C, optimally at 55°C. At 28°C, it’s dormant: the spores don’t germinate, the gas isn’t produced, the cans don’t swell. The incubation at 28°C was testing for mesophiles (which grow at 28°C — and which the retort kills easily, so they’re rarely the problem) while missing the thermophiles (the actual underprocessing indicators). The test was screening for the wrong organisms.

The production lot — the one that swelled at the customer — had a retort deviation (a temperature fluctuation, minor, unrecorded as a deviation): the underprocessing was real, the thermophiles survived, and the cannery’s incubation (at 28°C) missed them — while the customer’s warehouse (at 30°C+, the summer heat) grew them. The verification failed; the customer’s warehouse succeeded — the worst possible division of labor.

Root cause: the uncalibrated incubator

1. Incubator running 7°C cool. The thermostat drift — 35°C setpoint, 28°C actual, unnoticed for a year — put the incubation below the thermophiles’ growth range. The test couldn’t detect the underprocessing indicators.

2. No independent temperature verification. The incubator — uncalibrated for two years, unmonitored — documented the setpoint, not the actual. The verification’s foundation (the temperature) was unverified.

3. “No failures” misread as success. The year of negative results — the program’s pride — was the blindness, not the excellence. The absence of failures (at the wrong temperature) was interpreted as process control, rather than test inadequacy.

Corrective actions: heating the verification

Immediate: the incubator was recalibrated — the thermostat replaced, the actual temperature verified at 35°C (and the thermophilic screen was supplemented: a second incubator at 55°C, for the optimal thermophile detection). The affected lot — the customer’s swollen cans, plus the cannery’s retained samples (re-incubated at the correct temperature, where they swelled — confirming the underprocessing) — was recalled and destroyed.

Within 30 days, the incubation program was rebuilt: the incubators are now calibrated quarterly (the temperature verified, not just set), monitored continuously (independent data loggers — the actual temperature, recorded), and alarmed (any deviation from the 35°C/55°C targets notifies QA immediately). The incubation records now document the actual temperature (from the loggers), not the setpoint — the verification verifies reality. The sampling was also strengthened: the thermophilic screen (55°C, the optimal detection) is now the primary incubation, with the 35°C as the supplement — the program targets the indicators, not the conveniences.

The retort deviation — the minor fluctuation that the incubation should have caught — was investigated: the deviation procedure was tightened (the “minor” fluctuations now trigger the full deviation protocol), and the retort monitoring was enhanced (the temperature control, the chart review — the deviation that caused this is now impossible to miss). Within 90 days, the company’s other canneries were audited for incubator calibration, finding similar uncalibrated equipment at one more facility.

Results: the verification that verifies

Twelve months later: the incubators at the correct temperatures (the loggers prove it), three incubation failures detected (all investigated — two were true underprocessing deviations, caught before shipment — the program working as designed), and the “no failures” record replaced by the honest one: the failures that occur are found, because the test can find them.

Lessons learned: what you’d do Monday morning

Calibrate your incubators — today, with an independent thermometer. If the actual isn’t the setpoint, your incubation testing is fiction — and the “no failures” record is the blindness, not the excellence. Monitor continuously, alarm the deviations, and document the actual.

Then check your incubation temperatures against your indicators: are you screening for thermophiles (55°C optimal) or just incubating at a convenient 35°C? The test has to match the hazard — the underprocessing indicators need their growth temperature, or the verification is theater. And investigate your “no failures”: the prolonged absence of findings (in any verification) should trigger the question — is the process perfect, or is the test blind?

Yeah, but actually — the year of “no failures” is the warning of this case: the program’s proudest metric — zero findings — was the symptom of its failure. Your verifications should find things occasionally; the perfect record is suspicious, not reassuring. Calibrate the equipment, match the temperature to the indicator, and welcome the failures — because the incubation that catches the underprocessing is the last line of defense. The thermostat drifted 7°C; the program drifted into blindness. The thermometer — the independent one, the calibrated one — is the difference between verification and ritual.