Calibration
How to Set Calibration Intervals for Laboratory Equipment
“How often should we calibrate this?” is one of the most consequential questions a lab answers, and one of the most casually answered. The common reasons — “annually, because that’s what we’ve always done” or “whenever the manufacturer says” — are starting points, not justifications. And when an ISO 15189 or CAP assessor asks why this interval, “we’ve always done it this way” is precisely the answer that turns a routine review into a finding.
This post is a practical method for setting calibration intervals you can defend: where to start, how to adjust, how to handle instruments that need more than one schedule, and what the evidence trail should look like.
Why the interval needs a rationale at all
Accreditation standards don’t hand you a table of intervals. ISO 15189 asks that calibration occur on a schedule appropriate to the instrument’s use and stability. CAP and CLIA expect intervals consistent with manufacturer instructions and the lab’s own experience. The common thread is that the interval is a decision the lab owns — and decisions need reasons.
The reason matters because the two failure directions are both real:
- Too long, and the instrument can drift out of specification undetected between calibrations — every patient result in that window is suspect, and you find out at the next calibration or, worse, from an EQA failure.
- Too short, and you’re spending money and instrument downtime calibrating something that was perfectly stable, while the over-calibration itself introduces handling risk.
A defensible interval is the one where the cost of checking is justified by the risk of not checking. That’s a judgement, and judgements get documented.
Step 1: Start with the manufacturer baseline
The manufacturer’s recommended interval is the right default, and the easiest to defend. They tested the instrument’s stability; their number reflects it. Document it as your starting point with a one-line rationale: “Annual calibration per [manufacturer] IFU, section X.”
For many instruments, you can stop here. The baseline is appropriate, the rationale is the manufacturer’s data, and the assessor nods and moves on. Don’t over-engineer intervals that don’t need it.
Step 2: Adjust for risk and use
The baseline assumes typical use. Your use may not be typical. Four factors justify tightening (or, more rarely, extending) the interval:
| Factor | Pushes interval… | Why |
|---|---|---|
| Clinical criticality of results | Shorter | A drifting instrument behind a critical decision point is higher-stakes |
| Usage intensity / duty cycle | Shorter | Heavy use accelerates wear and drift |
| Demonstrated stability history | Longer | An instrument with years of in-spec calibrations has earned trust |
| Environmental stress | Shorter | Temperature swings, vibration, transport all accelerate drift |
The honest version of this is a short, written risk assessment per instrument category — a paragraph, not a thesis. “This analyser runs critical-care chemistry at high volume; despite the 12-month manufacturer baseline we calibrate the critical channels at 6 months, reviewed against EQA performance.” That sentence is worth more to an assessor than a perfect spreadsheet with no reasoning behind it.
Step 3: Use multiple tracks when one schedule isn’t enough
Here’s where a lot of registers fall down: they assume one instrument equals one calibration date. Real instruments often need several schedules running in parallel.
Consider an analytical balance:
- Annual external calibration by an accredited provider — the traceable, certificate-bearing event.
- Monthly internal verification with reference weights — a lighter check that catches drift between the big calibrations.
- Daily function check before first use — recorded as routine QC.
Each of these has a different interval, a different performer, and a different evidence type. Forcing them into one “calibration due date” loses information and hides risk. A proper register lets one instrument carry multiple independent calibration tracks, each with its own interval and its own status — and the instrument’s overall status reflects the worst of them. If the monthly verification is overdue, the instrument is overdue, even if the annual calibration isn’t due for months.
This is exactly the verification-between-calibrations expectation that ISO 15189 assessors probe, covered alongside traceability and acceptance review in ISO 15189 calibration requirements.
Step 4: Decide what happens when a calibration fails
An interval policy is incomplete until it says what a failure does to the schedule. The rule that keeps labs honest: a failed calibration must not advance the schedule.
It sounds obvious, but the common spreadsheet pattern quietly violates it. Someone logs “calibration performed” and rolls the next-due date forward by the interval — regardless of whether it passed. Now a failed instrument shows as “current” and won’t reappear as due until the next cycle, with a year of suspect results in between.
The correct behaviour: only a passing calibration rolls the due date forward. A failed calibration leaves the instrument in a due/overdue state, keeps generating reminders, and triggers a corrective action — investigate, repair, recalibrate, and verify — before it returns to clinical use. The instrument stays visibly broken until it’s actually fixed.
Step 5: Make the interval reviewable
Intervals aren’t set once and frozen. They should be revisited — when an instrument’s stability history accumulates, when usage patterns change, when an EQA trend or a near-miss suggests the current interval is too generous. A note in the record when you change an interval, and why, turns “we adjusted this” into a documented decision instead of an unexplained edit an assessor will ask about.
What the assessor actually wants to see
When a CAP or ISO 15189 assessor probes calibration intervals, they’re checking for three things:
- The interval is defined — there’s an actual schedule, not “when we get to it”.
- The interval has a rationale — manufacturer baseline, risk assessment, or stability history, written down somewhere they can see.
- The schedule is being followed — the calibrations are actually happening on time, and when they don’t, there’s a documented reason and corrective action, not a retrospective shrug.
The third point is where most findings actually land. A perfectly reasoned interval policy means nothing if the calibrations slip and nobody notices until the assessor reconstructs the dates. This is, once again, a reminders-and-tracking problem before it’s a science problem — the same one we keep coming back to in why spreadsheets fail lab compliance.
A working procedure
To set intervals across a register without it becoming a research project:
- Default to the manufacturer baseline for everything, documented in one line.
- Flag the high-criticality, high-use instruments and write a short risk-based rationale for those — usually a small minority of the register.
- Add verification tracks where a single annual calibration leaves too long a gap (balances, pipettes, temperature-controlled equipment).
- Set the failure rule once: only passing calibrations advance the schedule.
- Review intervals annually against accumulated history, adjusting with a documented reason.
Do this and the assessor’s question — why this interval? — has an answer for every instrument. Not always a long answer. But always an answer, which is the whole point.
Related reading:
