Glucose QC: two levels, one month
Review a month of glucose controls across two levels.
Levey-Jennings Chart Maker
FigranPaste QC results for each control level to plot SD limits and flag Westgard rule violations.
- Several control levels
- Westgard multirule flags
- SVG, PNG and flag table
2 levels · 30 runs · 60 results
Options
8 runs rejected
of 30 runs · 2 warnings · 2 levels · target mean and SD
Flag table
| Run number | Level | Result | z | Rule(s) | Status |
|---|---|---|---|---|---|
| 5 | Level 1 | 5.845 | 2.30 | 1-2s | Warning |
| 8 | Level 1 | 5.980 | 3.20 | 1-2s, 1-3s | Reject |
| 10 | Level 1 | 5.830 | 2.20 | 1-2s | Warning |
| 11 | Level 1 | 5.860 | 2.40 | 1-2s, 2-2s | Reject |
| 14 | Level 1 | 5.815 | 2.10 | 1-2s, R-4s | Reject |
| 14 | Level 2 | 15.12 | −2.20 | 1-2s, R-4s | Reject |
| 19 | Level 1 | 5.665 | 1.10 | 4-1s | Reject |
| 24 | Level 1 | 5.185 | −2.10 | 1-2s, 2-2s | Reject |
| 24 | Level 2 | 15.04 | −2.40 | 1-2s, 2-2s | Reject |
| 28 | Level 2 | 15.56 | −1.10 | 10x | Reject |
| 29 | Level 2 | 15.72 | −0.70 | 10x | Reject |
| 30 | Level 2 | 15.84 | −0.40 | 10x | Reject |
Control level summary
| Level | Unit | n | Mean | SD | CV (%) | Target mean | Target SD | Bias (%) | Rejected runs |
|---|---|---|---|---|---|---|---|---|---|
| Level 1 | mmol/L | 30 | 5.584 | 0.1774 | 3.18 | 5.500 | 0.1500 | 1.53 | 5 |
| Level 2 | mmol/L | 30 | 15.88 | 0.338 | 2.13 | 16.00 | 0.400 | −0.76 | 5 |
Report
QC results for 2 control levels over 30 runs were evaluated against target mean and SD with Westgard rules 1-2s/1-3s/2-2s/R-4s/4-1s/10x (1-2s as warning); 8 runs violated a rejection rule.
Levey-Jennings chart examples
Explore common control patterns and load an example into the tool.
Imprecision rises after run 15
Explore opposing excursions and increased scatter.
Three-level TSH immunoassay
Apply a rule set for three control levels.
Calibration drift on a z-score chart
Compare drift across two levels on a common SD scale.
Setting limits from 20 baseline values
Estimate mean and SD from a published set of values.
A shift after a new reagent lot
Mark a reagent change and follow the subsequent shift.
ELISA positive control across 26 plates
Set limits from the first 20 plates and monitor later plates.
qPCR standard Cq: in control
See a single warning in a stable standard series.
How to make a Levey-Jennings chart
Turn control results into a chart and flag table in four steps.
1Paste QC results
Paste one column per control level, or import a CSV.
2Choose mean and SD
Enter target values or calculate them from a baseline.
3Choose rules
Select a rule set and inspect flagged runs.
4Export your chart
Download the chart or copy the flag table.
Westgard rules
Westgard QC rules describe patterns in control measurements. These Levey-Jennings chart rules can flag individual excursions and repeated shifts.
| Rule(s) | When it fires | Applied | What it usually means | Example |
|---|---|---|---|---|
| 1-2s | One result beyond ±2 SD | Single result | Random error | |
| 1-3s | One result beyond ±3 SD | Single result | Random error | |
| 2-2s | Two results beyond the same ±2 SD limit | Same level across runs or same run across levels | Systematic error | |
| 2of3-2s | Two of three results beyond the same ±2 SD limit | Same level across runs or same run across levels | Systematic error | |
| R-4s | One result above +2 SD and another below −2 SD | Same run across levels | Random error | |
| 3-1s | Three consecutive results beyond the same ±1 SD limit | Same level across runs | Systematic error | |
| 4-1s | Four consecutive results beyond the same ±1 SD limit | Same level across runs | Systematic error | |
| 6x | Consecutive results on one side of the mean | Same level across runs | Systematic error | |
| 8x | Consecutive results on one side of the mean | Same level across runs | Systematic error | |
| 10x | Consecutive results on one side of the mean | Same level across runs | Systematic error | |
| 12x | Consecutive results on one side of the mean | Same level across runs | Systematic error |
Warning or rejection
1-2s is a warning by default. Rejection rules are evaluated independently; follow your laboratory’s procedure when investigating a flag.
Rule sets for two and three control levels
The two-level set uses 1-3s, 2-2s, R-4s, 4-1s and 10x. The three-level set uses 1-3s, 2of3-2s, R-4s, 3-1s and 6x. Across-run rules are checked separately for each level; 2of3-2s is also checked within the last three runs of a level.
References: Westgard et al. (1981) · Westgard QC: Westgard Rules
Mean, SD and control limits
A Levey-Jennings chart displays control measurements against a mean and SD limits for internal quality control.
x̄ = Σx / ns = √[Σ(x − x̄)² / (n − 1)]
x̄ ± k · sz = (x − x̄) / sCV = 100 · s / |x̄|
For 20 baseline values, 199.55 ± 2 × 4.18613 gives approximately 191.18–207.92.
Target values or your own baseline
Use the values established by your laboratory when available. A baseline estimated from early runs remains fixed while later runs are assessed. Using all data is descriptive and may obscure shifts.
References: Levey & Jennings (1950) · CLSI C24 · Westgard QC: The Calculations
Levey-Jennings chart FAQ
What is the difference between a Levey-Jennings chart and a Shewhart chart?
A Levey-Jennings chart applies the individual control chart to laboratory QC, usually using a long-term SD. Individual Shewhart charts may estimate short-term variation using a moving range.
How many QC results do I need to set the mean and SD?
A common starting point is at least 20 measurements over at least 10 days, preferably 20 days. Use established target values while a suitable baseline is being collected.
Should I use the manufacturer's target values or my own?
Use values established for your own laboratory when available; manufacturer values can be a starting point for a new control lot.
Why not use all the data to calculate the limits?
Estimating limits from the data being assessed can widen them and hide a shift or outlier. Use a separate baseline for monitoring.
What should I do when a Westgard rule is violated?
Investigate rejection flags according to your laboratory procedure, correct the cause, repeat the control and record the action before reporting affected results. 1-3s and R-4s often indicate random error; 2-2s, 4-1s and 10x can indicate systematic error.
Which Westgard rules should I use?
The choice depends on your control design and laboratory procedure. Westgard Sigma Rules use measures such as allowable error, bias and CV to guide the choice; this tool does not calculate Sigma.
Can I make a Levey-Jennings chart in Excel?
Yes. Calculate AVERAGE and STDEV.S, then plot results with constant mean and SD lines; rule checks need separate formulas. The Levey-Jennings chart generator above also acts as a Westgard rules calculator for the selected rules.
Can I print a blank Levey-Jennings chart?
Yes. Enter target mean and SD values without results, then download the chart for printing.
How do I report the CV of my QC material?
The level summary reports CV as sample SD divided by the observed mean. An RSD calculator can help with other relative standard deviation calculations.
Chart this month's QC results
Turn your control results into a chart and review the flags.
Make a Levey-Jennings chart