An Instrument Out of Calibration Should Not Be Bookable

Why is "overdue" a warning and not a control?#

Because a warning requires somebody to read it, and a control does not. A calibration schedule, whether it is a spreadsheet, a wall chart or a module in the quality system, records when each instrument is next due and shows which dates have passed. The information is correct. What it lacks is any connection to the moment of use. The analyst booking the balance for Tuesday does not look at the schedule; they look at the booking calendar, and the calendar does not know the balance is overdue.

The gap is structural, not a matter of diligence. The schedule is owned by the quality function and reviewed monthly. The calendar is used by the bench and consulted hourly. Two systems, two rhythms, and the instrument sits between them. In a small lab with one quality manager who also runs the bench, the gap is bridged by one person’s memory. In any lab larger than that, it is bridged by luck.

A control closes the gap at the point of use. An instrument whose calibration has lapsed, whose service is overdue or which has been flagged after a failed check should not appear as available in the calendar the lab already uses. Anyone who tries to book it should see that it is locked and why, and the lock should lift only when a person with the authority to clear it does so, with their name on the clearance. That is the difference between an accreditation assessor finding a red cell on a schedule and finding a result signed on a locked instrument. The first is a system working. The second is a nonconformance.

What does as-found versus as-left tell you?#

A calibration certificate carries two sets of readings for each test point, and most people read the wrong one. As-left is the instrument’s reading after any adjustment, and it is the number that says the instrument is fine from now on. As-found is the reading before adjustment, and it is the number that says what the instrument was doing for the months before the technician arrived.

For the future, as-left matters. For the past, only as-found does. An instrument found within tolerance and adjusted slightly closer to nominal has been right all along; every result it produced stands. An instrument found outside tolerance and adjusted back has been wrong for an unknown period ending today; every result it produced since the last time it was known to be right is now in question. The as-left reading is identical in both cases. The consequences are not.

The comparison should be arithmetic the system does, not a judgement the technician writes in a comment box. Each test point has a nominal value and a tolerance. Each as-found reading is inside or outside it. An instrument with any point outside is out of tolerance as found, and that verdict should be recorded on the instrument, dated, and visible to anyone who looks at the instrument’s history, regardless of what the as-left readings say.

Which results are suspect when an instrument fails as found?#

All of them since the last calibration at which the instrument was found within tolerance. The window is often wider than people expect: an annual calibration that fails means twelve months of results, not the few weeks since someone noticed a drift. The lab’s obligation is to assess the impact of the out-of-tolerance condition on those results, and the first step of that assessment is knowing which results they are.

This is where most systems fail, because they do not connect the instrument to the results it produced. The calibration record lives in the quality system. The results live in the LIMS or the notebook. The connection between them is a field in which the analyst typed the instrument’s identifier, if they did. The list of suspect results is then a search of that field across a year of records, and its confidence is the confidence in a year of typing.

The list should come from the record, not from memory or from a search of free text. Every result should carry a link to the instrument that produced it, as a real reference rather than a typed name, so that the question “what did this instrument produce between these two dates” is a query with a complete answer. Then the impact assessment begins from a list rather than an estimate, affected customers can be identified, and the assessor who asks how the lab knew the scope can be shown the record.

What happens downstream when a shared instrument goes offline?#

Instruments depend on each other more than schedules admit. A gas chromatograph depends on a gas generator. A microbiology suite depends on an incubator and an autoclave. A tensile test rig depends on the environmental chamber it sits in and the reference load cell it is checked against. When the upstream item is locked, the downstream items are unusable, but their own calibration dates are fine and their own calendar entries look available.

So an analyst books the chromatograph, arrives, and finds no carrier gas. Or worse, does not find out, because the generator’s fault is intermittent and the run completes with a result nobody knows to doubt. The schedule shows one instrument down. The lab has three.

A system that knows the dependencies can show them. When the upstream instrument is flagged, the downstream instruments that rely on it should appear as unusable too, with the reason pointing back at the upstream item. The propagation should follow the chain as far as it goes, because a reference standard at the top of a hierarchy can take a dozen instruments with it. The alternative is the quality manager holding the dependency map in their head and walking the lab telling people what not to use, which works until the quality manager is on leave.

Should the calibration record and the booking calendar be the same system?#

Yes, or one of them will lie. Everything above depends on the moment of booking knowing the state of calibration, and two systems maintained by two functions on two rhythms cannot guarantee that. A synchronisation between them is a third system with its own failure modes, and when it fails, the calendar shows an instrument available that the quality record shows locked, and nobody can tell which is right.

When the calibration record, the service schedule and the booking calendar are one record on the instrument, the lock is not a synchronisation. It is a property of the instrument that the calendar reads. The schedule can be driven by the calendar as well as the date, so an instrument approaching its service interval by usage count is flagged before the date arrives, which matters for anything whose wear depends on how much it is used rather than how long it has sat. And the history of every booking, every calibration, every lock and every clearance sits in one place, attributed and dated, which is what the assessor asks to see.

The practical test is simple. Name one instrument in your lab that is overdue today, and find out whether anyone booked it this week. If answering that takes two systems and a conversation, the calendar and the calibration record are not the same system, and one of them is wrong about something right now. The accreditation standard asks the lab to ensure equipment is fit for use before it is used. A due date on a schedule does not ensure anything. A lock in the calendar does.

What instrument management carries in this work#

LabLynx builds instrument and asset management with scheduling for laboratories whose accreditation depends on proving equipment was fit for use, alongside inventory management for the reference materials and consumables those instruments consume, and instrument integration so the result carries a real link to the instrument that produced it.