Skip to content
Pressure gauge dial with a red over-range zone
  1. Insights
  2. Handling an out-of-tolerance result: from OOT to impact assessment

Audits & quality

Handling an out-of-tolerance result: from OOT to impact assessment

An out-of-tolerance as-found reading is not the end of a calibration. It is the start of a question: what did this instrument tell you since it was last known to be right, and who relied on it?

  • By Nexus Metrology
  • 5 min read

What counts as out of tolerance

An instrument is out of tolerance (OOT) when its as-found reading, taken before any adjustment, falls outside the calibration tolerance set in your program. Two distinctions matter from the start:

  • OOT is not the same as a process failure. The calibration tolerance is usually tighter than the process limit it protects. An instrument can be out of its tolerance while every reading it took was still within the process limit once the error is accounted for. That is what the assessment establishes.
  • An adjustment is not an OOT. Many programs set an adjustment limit inside the tolerance. An instrument found between the two is adjusted but still passed; one found beyond the tolerance is OOT.

21 CFR 211.160(b)(4) expects your calibration program to include provisions for remedial action when accuracy or precision limits are not met, and says that instruments not meeting established specifications shall not be used. The steps below are one way to make those provisions concrete.

Step 1: Contain

  1. Record before you touch. Complete the as-found readings at every test point and write them down. Adjusting first destroys the evidence the rest of the process depends on.
  2. Take the instrument out of use. Label it, remove it from service or switch to a calibrated alternative, so no new data is produced with it.
  3. Protect what is running now. If the instrument monitors a critical condition, such as a room pressure, a chamber temperature or a sterilization parameter, put an interim measurement in place until it is restored.
  4. Notify the owner and quality. Your procedure should say who is told and how quickly.

Step 2: Confirm the result

Before you assess impact, make sure the OOT is real and understand its size. Repeat the measurement. Check the reference standard is in date and working, check connections, tubing and set-up, and for loop-connected instruments check the loop as well as the sensor. Then record the magnitude and direction of the error at each point. An instrument reading 0.8 °C high near the set point tells a very different story from one reading 0.8 °C low.

Step 3: Assess the impact

This is the heart of the process. The question is simple to state: since this instrument was last known to be good, what did it tell us, and what was decided on that basis?

Set the window

The window runs from the last calibration where the instrument was found in tolerance to the moment it was taken out of use. Not from the due date: you don’t know when within the interval the drift began, so assume the earliest possible point unless you have evidence, such as trend data or intermediate checks, that narrows it.

List what relied on it

  • batches manufactured, tested or released using its readings;
  • environmental or utility data used to show rooms and systems were within limits;
  • stability studies, validation runs or qualification data;
  • other instruments calibrated with it, if the OOT instrument is itself a reference standard.

Apply the error

For each use, correct the recorded values by the known error and ask whether the true value could have been outside its process limit or specification. If no corrected reading crosses a limit, the impact is usually low, and the record should show the reasoning. If one could, product or data may be affected, and the assessment widens.

Follow it as far as it goes

21 CFR 211.192 requires unexplained discrepancies and specification failures to be thoroughly investigated, extending to other batches that may have been associated with the failure, whether or not they have been distributed. EU GMP Annex 1 sets a parallel expectation for environmental monitoring: when action limits are exceeded, procedures should prescribe a root cause investigation and an assessment of potential impact on product, including batches produced between monitoring and reporting. The same logic applies to an instrument found OOT: the assessment covers everything inside the window.

ICH Q9(R1) is a useful guide to proportion: the depth and formality of the assessment should match the risk. An OOT on a non-critical utility gauge needs a short, reasoned record; an OOT on a sensor that released product needs a full investigation.

Step 4: Find the cause and act on it

Common causes include normal drift over too long an interval, physical damage or shock, an environment harsher than the instrument is rated for, the wrong range or model for the application, and installation problems such as tubing leaks or loose terminations. Choose corrective actions for this instrument (repair, replace, re-range, relocate) and preventive actions for the program, for example a shorter interval for that model, an intermediate check, or a change to the installation standard. An OOT should shorten the instrument’s interval straight away; see setting risk-based calibration intervals.

Step 5: Document and close

The OOT record should let a reviewer follow every decision without asking anyone:

  • the instrument, its tolerance and the as-found and as-left data;
  • the date of the last in-tolerance calibration and the impact window;
  • what relied on the instrument in that window, and the conclusion for each;
  • the cause, the corrective and preventive actions, and their owners;
  • quality approval, and the date the instrument returned to service.

Link the OOT record from the calibration record, so that anyone reading the certificate sees the investigation behind it. See what auditors look for on a calibration certificate.

Common pitfalls

  • Adjusting first, recording later, which leaves no reliable as-found data.
  • Starting the window at the due date instead of at the last good calibration.
  • Assessing the sensor and forgetting the system: the alarm that didn’t trip or the screen that showed a scaled value.
  • Closing on “no impact” without showing the corrected readings that justify it.
  • Leaving the interval unchanged after the same instrument fails again.

How Nexus Metrology helps

On every on-site calibration, Nexus Metrology records as-found data before any adjustment and flags out-of-tolerance results clearly, so your impact assessment starts from solid evidence. We can also support the assessment itself and the program changes that follow, drawing on 15+ years of GMP experience. See Metrology Consulting & Training or Fractional Leadership.

Sources

Regulations and guidance are paraphrased above; the originals are the reference.

  1. 21 CFR 211.160: Laboratory controls, general requirements (eCFR)
  2. 21 CFR 211.192: Production record review (eCFR)
  3. EudraLex Volume 4, Part I, Chapter 3: Premises and Equipment, paragraph 3.41 (European Commission)
  4. EudraLex Volume 4, Annex 1: Manufacture of Sterile Medicinal Products (2022), paragraph 9.13 (European Commission)
  5. FDA Guidance for Industry: Q9(R1) Quality Risk Management (2023)

Next step

Talk to a metrology expert

Questions about your own instruments, intervals or records? Nexus Metrology brings 15+ years of GMP experience and on-site field calibration to your facility.

Keep reading

All Insights articles