# Troubleshooting Failed QC Results in Microbiology Failed QC in microbiology is almost always a control or reagent problem, and the first thing to check is the control itself, not the patient isolate, the instrument, or the analyst. Expired material, a broken cold chain, one freeze-thaw cycle too many, or a reconstitution error accounts for a large share of out-of-range results, and each of those is fast and cheap to rule out. Quality control (QC) is the planned set of checks that confirms a test system is performing within predetermined limits before patient results are released. A microbiology QC failure investigation is the structured process that follows an out-of-range result: stop release, confirm the failure, find the cause, correct it, and prove the correction worked. Microbiology is harder to control than clinical chemistry. Cultures grow, media vary, organisms change, and many results are qualitative or interpretive rather than a single number. A sterility check, a growth promotion test with a defined organism, a Gram stain control, an antimicrobial susceptibility endpoint, and a molecular negative control all answer different questions. That variety is why no single troubleshooting recipe fits every bench. The logic does fit, and it runs from the most common and most easily fixed causes to the rarest. This guide covers the causes, confirmation steps, fixes, and prevention practices for failed QC across media, reagents, stains, instruments, susceptibility testing, serology, and molecular assays. It is written for the bench technologist, the section supervisor, and the laboratory director who has to sign the corrective action. ## Why Most Microbiology QC Failures Start at the Bench Quality control in microbiology is not one test. It is a set of independent checks that each answer a different question. A sterility check asks whether a medium is free of viable organisms. A growth promotion test asks whether the medium supports a defined organism. A negative control in a molecular assay asks whether the workflow introduced contamination. A kit control in serology asks whether the reagents and the instrument still agree with the manufacturer's expectations. A control material is the sample with a known expected result that is run alongside patient specimens. Acceptance limits are the numeric or qualitative boundaries that define a passing result. Most failures are pre-analytical. The control was stored wrong, thawed too many times, reconstituted with the wrong volume, or used past its expiry. The medium was poured too thin, autoclaved too long, or stored past its shelf life. The inoculum was too heavy. The incubator ran a degree warm overnight. These causes share a feature: they are visible on the bench and cheap to check. The rare causes, such as true reagent degradation or a stubborn instrument fault, take longer to find and should be investigated only after the common ones are excluded. A useful mental model is a chain with five links: the control, the reagent, the operator, the equipment, and the acceptance criteria. A failure can originate in any link, and the investigation should test the links in that order. The chain also explains why two laboratories can see the same out-of-range result and reach different conclusions. One laboratory has a control problem and the other has limits set too tight. Calibration, QC, and proficiency testing are not the same thing. Calibration establishes the relationship between an instrument signal and a known concentration. QC verifies that the calibrated system continues to perform. Proficiency testing, also called external quality assessment or EQC, compares the laboratory's result on an unknown sample with peer laboratories. A failed proficiency test is not a QC failure, although both can share a root cause. Keeping the three concepts separate stops the investigation from drifting into the wrong records. ## The Troubleshooting Table: Match the Symptom to the Cause Use the table as a first pass, then confirm each suspected cause with the check listed in the third column. The fourth column is the fix that resolves the cause, not a workaround that hides it. | Symptom | Likely cause | How to confirm | Fix | |--|--|--|--| | Sterility check shows turbidity | Medium contaminated during preparation or a damaged container | Inspect broth, review autoclave load records | Discard the lot, repeat sterility check, review sterilization cycle | | Growth promotion fails for one organism only | Nonviable organism or a selective agent that is too inhibitory | Subculture the organism on nonselective medium | Replace the organism, repeat with a fresh culture, remake the medium | | pH outside the specified range | Weighing, water quality, or a drifted pH meter | Check buffers and calibration log | Recalibrate the meter, remake the medium, check pH after cooling | | Negative control amplifies in a molecular assay | Amplicon carryover or reagent contamination | Repeat with new reagents and a new set of pipettes | Decontaminate surfaces and pipettes, re-extract, review workflow | | Positive control fails | Degraded control, reagent, or enzyme | Repeat with a fresh aliquot and a new kit lot | Replace the failing component and retest | | One QC point outside 2 SD but inside 3 SD | Random error or an early shift | Plot the Levey-Jennings chart and apply the rules | Repeat the control, investigate a trend, hold results until resolved | | Susceptibility MICs shift by two or more dilutions | Inoculum density, incubation temperature, or a new lot | Repeat with a fresh 0.5 McFarland suspension and a calibrated incubator | Correct the inoculum, verify the temperature, retest | | Zone diameters shrink across all drugs | Agar depth, pH, or inoculum | Measure plate depth with a ruler | Remake plates to 4 mm depth, verify pH, retest | | Incubator temperature alarm | Door left open, thermostat fault, or power interruption | Compare with an independent thermometer | Move cultures, service the unit, document the excursion | | Gram stain controls faint or decolorized | Aged stain, timing, or technique | Stain known positive and negative controls | Replace the stain, retrain, review the procedure | | External QC outside the assigned limits | Systematic bias or a matrix effect | Compare with the recent QC trend and kit lot | Investigate calibration and lot, take corrective action | | QC fails for one analyst only | Technique or reading error | Observe the analyst and split a sample | Retrain and reassess competency | ## Work From the Most Common Cause to the Rarest ### 1. Degraded, Expired, or Mishandled Control Material A control material is a sample with an assigned or established expected result. It can be a lyophilized kit control, a frozen aliquot, a swab seeded with a known organism, or a characterized patient pool. Every control has a stability window that begins when it is reconstituted or thawed, not when it is manufactured. The window is shorter than most benches assume. Lyophilized controls absorb moisture from the air every time the vial is opened, and reconstituted controls lose activity at refrigerator temperature over days to weeks. Freeze-thaw cycles are a common and undercounted cause of failure. Ice crystal formation concentrates solutes and can denature proteins or disrupt cells. A control that was aliquoted into single-use volumes and frozen once behaves differently from a control that was thawed, refrozen, and thawed again. Frost-free freezers compound the problem because they cycle above freezing to clear ice. Store critical controls in a non-frost-free unit and record every freeze-thaw cycle on the vial. Contamination of an open control vial is another path. A pipette tip that touched a patient sample, a vial left open on the bench, or a cap swapped between tubes can introduce organisms or nucleic acid. Molecular controls are especially vulnerable because a single amplicon can overwhelm a clean reaction. Handle controls in a separate area from patient material whenever the workflow allows. The confirmation is simple. Check the expiry date, the storage log, the freeze-thaw count, the reconstitution date and initials, and the physical appearance. Cloudiness, precipitate, a color change, or a volume that is lower than expected all point to the control. Repeat the test with a fresh aliquot of the same lot. If the fresh aliquot passes, the original material was the problem. If both fail, move to the reagent or the instrument. Some controls contain human-derived material and must be handled as potentially infectious. Wear gloves and a gown, work in a biosafety cabinet when the procedure calls for one, and discard the material as biohazard waste. ### 2. Reagent and Media Preparation Errors Media QC has four parts: sterility, growth promotion, selectivity or inhibition, and physical characteristics. Each part can fail for a different reason. Sterility fails when the autoclave cycle is inadequate, when a container cracks during handling, or when the medium is contaminated after sterilization. Growth promotion fails when a dehydrated base is expired, when a supplement is added at the wrong temperature, or when a selective agent suppresses the test organism more than expected. Water quality affects every prepared reagent. Deionized water with a high bioburden or residual disinfectant can inhibit growth and shift pH. Review the water system records before blaming the dehydrated medium. Weighing errors and incomplete dissolution cause pH drift, and pH measured before autoclaving is not the same as pH after cooling. Agar depth matters for disk diffusion, where the standard is 4 mm. A plate poured too thin gives falsely large zones and a plate poured too thick gives falsely small zones. Blood and supplements are heat sensitive. Blood added to agar that is still too hot will lyse and change the color and performance of the medium. Selective agents added above their tolerance temperature lose potency. Both errors look like a bad lot from the manufacturer when the real cause is on the bench. Review the preparation record for lot numbers, weights, water source, pH before and after sterilization, and the initials of the preparer. Measure the depth of a representative plate. Repeat growth promotion with fresh cultures of the required organisms. Remake the medium from a fresh dehydrated base and verify pH after cooling. Record the correction and the new lot. Autoclave operation carries burn and pressure hazards. Follow the manufacturer's loading and cycle instructions, never open a cycle before the pressure has returned to zero, and verify performance with biological indicators on the schedule your procedure requires. ### 3. Operator Technique and Inoculum Errors Inoculum density is the most common technique error in susceptibility testing. A 0.5 McFarland standard approximates 1.5 x 10^8 colony-forming units per milliliter, and small differences in how it is prepared and diluted change the endpoint. Colonies taken directly from an overnight plate behave differently from colonies suspended from broth. A suspension that sits too long before inoculation settles. A swab that is too wet leaves excess fluid on the plate. Timing errors are equally common. Reading a plate before the incubation window is complete gives endpoints that are one or more dilutions off. Reading too late allows trailing endpoints and overgrowth. Gram stain decolorization is a timed step, and a few seconds of over-decolorization turns a positive control pink. Pipetting errors in a dilution series multiply down the series and can make a single control fail while the rest of the run looks acceptable. The confirmation is direct observation. Watch the analyst perform the step and compare the result with a split sample run by a second analyst. Review the competency record and the last proficiency test result. If the failure follows the analyst rather than the material, the cause is technique. Correct with retraining, a competency reassessment, and a written procedure that specifies the inoculum method, the timing, and the reading criteria. Do not treat a single retraining as a permanent fix. Recheck competency on the schedule the laboratory has set. ### 4. Equipment, Temperature, and Environmental Drift Incubators, water baths, refrigerators, freezers, biosafety cabinets, autoclaves, pH meters, pipettes, and centrifuges all drift. An incubator that runs warm at the top shelf and cool at the bottom changes growth rate and susceptibility endpoints across the same run. A refrigerator that cycles above its set point shortens the life of every reagent stored inside it. A pH meter calibrated with stale buffers reports a pH that is wrong by enough to change medium performance. Temperature monitoring is the first check for any unexplained shift in growth or susceptibility. Compare the displayed temperature with an independent thermometer or a calibrated data logger. Review the alarm history and the door-opening log. For an autoclave, review the biological indicator and the physical cycle record. For a biosafety cabinet, review the certification date and the airflow alarm history. For a molecular bench, review the environmental monitoring results, including surface swabs and air plates. Correct by servicing or recalibrating the unit, moving cultures to a verified backup, and documenting the excursion and its effect on patient results. Equipment failures that occur outside working hours are the ones that damage QC most often, so the monitoring system should alarm to a phone, not just to a display on the door. Ultraviolet lamps in biosafety cabinets and laminar flow hoods are a hazard to eyes and skin. Never work with the lamp on, and follow the manufacturer's schedule for lamp replacement and cabinet certification. ### 5. Acceptance Limits Set Too Tight (False Rejection) A false rejection is a QC result that falls outside the acceptance limits even though the test system is performing correctly. False rejections cost staff time, reagents, controls, and calibrators, and they delay patient results. The traditional approach sets acceptance criteria from the mean and standard deviation of 20 external QC results [1]. That approach has been widely accepted in medical pathology, and Westgard rules, developed for clinical chemistry, are now applied to infectious disease testing [1]. The rules create frequent false rejections that consume resources [1]. A five-month cost mapping in a single laboratory logged every QC result that failed Westgard rules using Bio-Rad Unity software and counted the reagents, external QC material, kit controls, calibrators, and staff time involved [1]. The same laboratory compared its traditional external QC review with QConnect limits developed by the National Serology Reference Laboratory [1]. The lesson is that some apparent failures are statistical artifacts, not assay problems. Limits set from too few data points are unstable. A mean and standard deviation from 20 results has wide [confidence intervals](/blog/data-analysis/confidence-interval-formula-examples), and a new lot or a new instrument can shift the mean enough to put every subsequent result outside the old limits. Multirule schemes increase the chance that at least one rule fires on a normal run, which is why a single point outside 2 SD is weaker evidence than a consistent trend. The confirmation is a review of the Levey-Jennings chart, the number of recent false rejections, and the external QC history. If the system agrees with external QC and peer results, the limits are the problem. Correct by establishing limits from a larger data set, by using robust or percentile-based limits, and by verifying the new limits against external QC before adopting them. Never widen limits simply to make a failing result pass. That converts a real problem into a silent one. ### 6. Lot-to-Lot Variation Manufacturers change raw materials, production lines, and minor formulation details between lots. Most changes are invisible, and a few shift the performance enough to move a control outside its limits. A new kit lot that fails on first use is not proof of a bad lot. It is a signal to verify the new lot against the old one before accepting it. The confirmation is a parallel comparison. Run the old lot and the new lot side by side with the same control and the same instrument. If the new lot reads consistently higher or lower, the difference is lot related. If both lots fail in the same direction, the cause is upstream in the instrument, the control, or the operator. Correct by documenting the verification, accepting the lot only when it meets the predefined criteria, and updating the QC target only when the shift is real and justified. Adjusting the target to fit a new lot without evidence is a form of normalization of deviance. It hides drift and defeats the purpose of laboratory quality management. ### 7. Real Assay Degradation or Contamination Sometimes the assay really is failing. Reagent degradation shows up as a gradual loss of signal or a slow drift in the positive control, not a sudden jump. Lamp aging, detector drift, and column or cartridge wear produce the same pattern. Carryover in an automated platform shows up as a positive result in the negative control that follows a strongly positive sample. Environmental contamination in a molecular laboratory shows up as a negative control that amplifies intermittently, often after a busy day. The confirmation is a swap test. Replace one component at a time with a known good component and repeat the control. A new reagent lot, a new extraction kit, a new amplification plate, or a new lamp will isolate the failing part. For molecular contamination, repeat the negative control with new reagents in a clean room and swab the work surfaces. If the clean-room repeat is negative and the original bench repeat is positive, the contamination is local. Correct by decontaminating the area, replacing the failing component, and repeating the QC until it passes on consecutive runs. Document the cleaning and the replacement. For carryover, review the wash cycle and the sample order on the instrument. Sodium hypochlorite and other disinfectants used for decontamination are irritants. Dilute them according to the manufacturer's instructions, use them in a ventilated area, and wear gloves and eye protection. ### 8. Data Handling and LIS Errors Some QC failures are not failures at all. The result was entered against the wrong control level, the wrong lot number was selected in the software, a unit conversion was applied twice, or the control was recorded on the wrong date. Transcription errors are more common when the bench is busy and the result is entered by hand. The confirmation is the audit trail and the raw data. Compare the printed instrument output with the value in the laboratory information system, check the lot and level mapping, and review who entered the result and when. If the raw data pass and the recorded result fails, the error is in the data path. Correct by fixing the entry, retraining on the LIS workflow, and adding a barcode or interface check where manual entry is unavoidable. Report the event through the laboratory's incident system so the pattern is visible if it recurs. ## A Step-by-Step Microbiology QC Failure Investigation Workflow A microbiology QC failure investigation should follow a fixed order so that the cheapest and most common causes are excluded first. The sequence below works for media, stains, susceptibility, serology, and molecular assays. Adapt the specific checks to the method, but keep the order. 1. Stop release. Hold patient results that depend on the failed QC. Do not report, do not verify, and do not let the bench talk itself into releasing because the result looks plausible. 2. Confirm the failure is real. Recheck the reading, the calculation, the control level, the lot number, and the data entry. A surprising number of failures end here. 3. Check the control. Expiry, storage, freeze-thaw count, reconstitution volume, appearance, and aliquot date. Repeat with a fresh aliquot. 4. Check the reagent and the medium. Expiry, lot, storage, preparation record, pH, sterility, and growth promotion. Repeat with a new lot if one is available. 5. Check the equipment. Temperature logs, calibration dates, maintenance records, and independent verification with a second device. 6. Check the operator. Observe the technique, review competency, and split a sample with a second analyst. 7. Check the acceptance criteria. Review the Levey-Jennings chart, the false rejection history, and the external QC agreement. 8. Escalate to the manufacturer when the failure follows a lot or a kit and the laboratory has excluded its own processes. Provide the lot number, the control results, and the raw data. 9. Perform root cause analysis and open a corrective and preventive action (CAPA) record if the cause is systemic. 10. Verify effectiveness. Repeat QC on consecutive runs, review the trend for a defined period, and close the CAPA only when the correction is shown to hold. The workflow also defines what not to do. Do not repeat the control until it passes without a documented reason. Do not release patient results on the basis of a single passing repeat. Do not change the acceptance limits before the cause is known. Each of those shortcuts converts a visible failure into an invisible one. ## Root Cause Analysis and Corrective Action Root cause analysis is the structured search for the underlying condition that allowed the failure. It is not a search for someone to blame. Two tools cover most microbiology events. The first is a cause-and-effect diagram organized around people, machine, method, material, measurement, and environment. The second is the five whys, which asks why the failure occurred and then why that cause occurred, until the answer reaches a process that can be changed. A correction is the immediate fix. A corrective action is the change that prevents the same failure from recurring. A preventive action is the change that prevents a similar failure elsewhere. The three are different records and should not be collapsed into one line in the quality system. Effectiveness verification is the step most often skipped. The CAPA is not complete when the form is signed. It is complete when the QC results after the change show that the problem is gone. Define the verification period and the acceptance criteria before closing the record. ## Common Mistakes and Limitations 1. Adjusting limits to fit a failing result. This is the most damaging mistake in laboratory quality management. It hides real drift and converts a controlled failure into an uncontrolled one. 2. Repeating the control until it passes. A passing repeat without a documented cause proves only that the failure is intermittent. Intermittent failures are the ones that reach patients. 3. Blaming the analyst first. Technique is a real cause, but it is rarely the first one. Check the material and the equipment before the person. 4. Using an expired control just this once. Expired material has a reason for its expiry date. A single use can produce a result that looks acceptable and is not. 5. Not documenting transient failures. A failure that resolves on repeat still belongs in the record. Trends are built from events that were written down. 6. Treating every point outside 2 SD as a failure. A single point is weak evidence. Apply the rules, look at the trend, and check the external QC before holding a line. 7. Failing to verify a new lot. A new lot is a new variable. Verify it in parallel before accepting it, even when the manufacturer's data look convincing. 8. Looking past the data path. Transcription and LIS mapping errors are common and easy to miss. Compare the raw output with the recorded result. 9. Assuming molecular contamination is impossible. Every amplification laboratory has contamination risk. The negative control exists because the risk is real. 10. Treating a false rejection as a free lesson. False rejections cost reagents, controls, calibrators, and staff time, and a laboratory that ignores the cost will eventually ignore the limits. A single out-of-range result does not diagnose a systemic problem on its own. The laboratory director or quality manager has to weigh the trend, the clinical impact, and the requirements that apply to the specific assay. Judgment is part of the process, not evidence that the process failed. ## Prevention: QC Systems That Rarely Fail Prevention is cheaper than investigation. The practices below reduce both the number of failures and the time needed to resolve them. - Write a procedure for every QC activity. The procedure should state the control material, the frequency, the acceptance limits, the action to take on failure, and the records to keep. - Train and reassess competency. Initial training is not enough. Reassess on a defined schedule and after any QC failure that points to technique. - Verify every new lot in parallel. Document the comparison and accept the lot only when it meets the predefined criteria. - Monitor temperatures continuously. Use calibrated data loggers with alarms that reach an on-call phone. - Maintain and calibrate equipment on schedule. Keep the certificates, the service reports, and the corrective maintenance records together. - Run external QC and review it against peer limits. The QConnect approach described in the cost study is one model for comparing performance with a reference group [1]. - Review Levey-Jennings charts and trend reports at a fixed interval. A trend caught early is a minor adjustment. A trend caught late is a recall. - Audit the data path. Check the LIS mapping, the unit settings, and the manual entry steps for the assays with the highest volume. - Keep the environmental monitoring program current. Surface swabs, air plates, and water testing catch contamination before the negative control does. - Track CAPA effectiveness. A closed record with no follow-up data is an open risk. The cost of prevention is visible in the budget. The cost of investigation is visible in the same budget, spread across reagents, controls, calibrators, and staff time. A laboratory that maps its QC failures, as the five-month cost study did, can see which limit changes would reduce false rejections without weakening the ability to detect real shifts [1]. ## Frequently Asked Questions ### What is the most common cause of a failed QC result in microbiology? A degraded, expired, or mishandled control material is the most common cause. Check the expiry date, storage temperature, freeze-thaw history, and reconstitution record before looking at the instrument or the analyst. ### Should patient results be held when QC fails? Yes. Patient results that depend on the failed QC should be held until the failure is resolved or the laboratory director documents a justified exception. Releasing results during an unresolved failure removes the protection that QC exists to provide. ### What is a false rejection and why does it matter? A false rejection is an out-of-range QC result on a system that is actually performing correctly. It matters because it consumes reagents, controls, calibrators, and staff time, and it delays patient results. ### How do I tell random error from systematic error? Random error scatters on both sides of the mean and does not repeat. Systematic error shifts consistently in one direction across consecutive points. Plot the Levey-Jennings chart and look at the pattern before deciding which one you have. ### How do I investigate a failed negative control in a molecular assay? Repeat the negative control with new reagents and new pipettes in a clean area, and swab the work surfaces. If the clean-area repeat is negative, the contamination is local to the original bench. ### Is it acceptable to adjust QC limits after a new lot fails? No, not as a first response. Verify the new lot against the old lot in parallel, confirm the shift with external QC, and change the limits only when the evidence shows the shift is real. ### What is the difference between correction and corrective action? A correction fixes the immediate problem, such as replacing a bad control vial. A corrective action changes the process so the same problem does not recur, such as adding a freeze-thaw log to the control handling procedure. ### How often should QC performance be reviewed? Review QC results at every run for immediate acceptance and on a fixed schedule for trends. The scheduled review should include Levey-Jennings charts, external QC, and any CAPA that is still open. ## Sources 1. [Cost-benefit analysis of two quality control approaches for infectious disease testing.](https://pubmed.ncbi.nlm.nih.gov/40157876/)