Internal Quality Control (IQC) Record and Levey–Jennings Chart
Required by: ISO 15189:2022 Cl. 7.3.7.2 — Internal quality control
Daily proof that each analytical system was in control on the day the patient result was produced.
Checked against NABL 112A (Issue 01, 18-Dec-2024) · page updated 23 September 2026
How often it must be recorded
Two levels of QC on the day of testing, then one level every shift / 8 hours, run at fixed times during operational hours.
Every field this record must carry
An assessor reads the record, not the intention behind it. A field left off the form is a finding, however well the work was actually done.
- Analyte and instrument
- QC lot and level
- Date, shift and fixed run time
- Observed value
- Laboratory's own mean and SD
- LJ plot point
- Accept / reject decision against defined criteria
- Outlier, trend and shift analysis
- Corrective action taken
- Monthly mean, SD and %CV
What NABL 112A actually says
NABL 112A: two levels of QC shall be included on the day of performing the test and subsequently one level every shift/8 hour; controls shall be run at fixed timings during operational hours. The laboratory shall derive its OWN mean and SD using a minimum of 20 data points to plot an LJ chart, shall review LJ charts daily, and shall calculate monthly mean, SD and %CV. Where a control has a short shelf life (the document gives CBC as the example) the manufacturer's assigned mean and SD may be used, but the laboratory must still calculate its own imprecision as %CV and keep it within the acceptable range. A minimum of six months of IQC data is recommended to calculate routine imprecision, updated annually where possible.
Source: NABL 112A — Specific Criteria for Accreditation of Medical Laboratories, Issue 01, 18-Dec-2024, 108pp. Downloaded from nabl-india.org 2026-09-23. Paraphrased with attribution. ISO 15189:2022 clause numbers and titles are cited as references; the standard's text is copyrighted and is not reproduced here.
What an assessor actually asks for
Reasoning from the clause, not from a published checklist — NABL does not publish assessor scripts. Treat this as how the requirement is normally evidenced.
- The record itself, for a period they choose — not a blank template.
- Completeness across the period. Gaps are the finding; a perfect January and an empty March reads worse than a consistently modest record.
- Who recorded and who verified, with dates. An unsigned record is not evidence that the work was checked.
- The linked procedure — the SOP that says this record must exist, at this frequency, in this format.
- What happened when it went wrong. Every record above has a failure path, and the action taken is usually the part that is missing.
Records this one depends on
- EQA / Proficiency Testing Participation Record Cl. 7.3.7.3
- Measurement Uncertainty (MU) Record Cl. 7.3.4
- Comparability of Results Record Cl. 7.3.7.4
Where the software fits
HealthFlow is lab software, not an accreditation body. It is not NABL accredited and using it does not make a laboratory accredited or compliant — the records above are the laboratory's own responsibility.
What it does do is produce some of this evidence as a by-product of normal work: turnaround time is measured per test on the owner dashboard, reports are versioned and locked on authorisation with dual signatures, every report access is written to an append-only audit trail enforced at the database, retained two years, and all of it exports at any time.