All quick-reference guides
Download Letter PDFA4 PDF

Print in portrait on US Letter or A4 at 100% scale, with 10 mm (about 0.4 in) margins. Turn off browser headers and footers. Background graphics are optional.

LatticeMLSLaboratory operations · QUICK REFERENCE 02 / 09

Common QC rule violations

Six statistical rules and the limits of their interpretation

Quality control (QC) compares a control result with its established mean and standard deviation (SD), a measure of usual variation. Random error increases scatter; systematic error shifts results toward one side. The patterns below suggest which type to investigate.1

Vertical position = (result − mean) ÷ SD, using each level's own target. Lines: mean (0), ±1, ±2, ±3 SD. Squares meet the named rule; circles give context.

12s

One result beyond ±2 SD.

1-2s example−3−2−10+1+2+3SD1234Run number →

Classic warning. Inspect the selected rejection rules.

13s

One result beyond ±3 SD.

1-3s example−3−2−10+1+2+3SD1234Run number →

Rejection. May reflect random or systematic error.

22s

Two consecutive beyond the same 2 SD limit.

2-2s example−3−2−10+1+2+3SD1234Run number →

Rejection. Suggests systematic error.

R4s

One > +2 SD, one < −2 SD, in one run.

R-4s example−3−2−10+1+2+3SDABLevel (one run)

Rejection: possible random error. Spread = 4.8 SD.

41s

Four consecutive beyond the same 1 SD limit.

4-1s example−3−2−10+1+2+3SD1234Run number →

Rejection. Suggests systematic error.

10x

Ten consecutive on one side of the mean.

10-x example−3−2−10+1+2+3SD12345678910Run number →

Rejection. Suggests systematic error.

Examples use one level across runs, except R4s (two levels, one run). The 22s rule also applies across levels within a run; 41s and 10x can span levels and runs as the QC plan specifies.2

Warning and rejection

In the classic Westgard scheme, 12s is a warning to check the rejection rules. Plans that evaluate every rule directly can detect the 41s and 10x examples here even though no result exceeds 2 SD. The adopted plan defines rules, levels, runs, and look-back periods.1,2

Respond to a rejection

Hold affected patient results. Use the adopted QC plan to investigate controls, reagents, calibration, and instrument performance. After correction and acceptable QC, assess which patient results need review or repeat testing. A passing repeat alone does not explain the original failure.2