How this instrument works
In clinical laboratory quality control, labs regularly run the same control material as hundreds of other labs and compare their own average result against the peer group's — the wider consensus of labs using the same method or instrument. The Standard Deviation Index (SDI) turns that comparison into a single number: how many peer-group standard deviations your lab's mean sits away from the peer-group mean.
The formula is SDI = (your lab's mean − peer-group mean) / peer-group standard deviation. An SDI of 0 means your lab's average exactly matches the peer group's; an SDI of +1.0 means your lab runs one peer-group standard deviation above the group; an SDI of -2.0 means two standard deviations below. Because it's standardized, an SDI can be compared meaningfully across completely different analytes and instruments, even though a glucose result and a cholesterol result are measured in different units on different scales.
SDI is a bias check, not a precision check — it tells you whether your lab is running systematically high or low relative to its peers, not how much your own results scatter from run to run. Labs commonly use SDI thresholds (often ±2 or ±3) as a trigger to investigate calibration, reagent lots, or instrument drift, which is why this figure shows up constantly in proficiency-testing and peer-comparison reports.
- Enter your laboratory's own average result for the control material into Your lab's mean result.
- Enter the peer group's (or target's) average result into Peer-group (target) mean.
- Enter the peer group's standard deviation into Peer-group standard deviation — must be greater than zero.
- Read Standard Deviation Index (SDI) — a positive value means your lab runs high relative to the peer group; negative means low; values near zero indicate good agreement.
Worked example — a lab mean of 102 against a peer mean of 100
Enter 102 into Your lab's mean result, 100 into Peer-group (target) mean, and 2 into Peer-group standard deviation. The instrument subtracts the means (102 − 100 = 2) and divides by the peer SD.
Standard Deviation Index (SDI) reads 1.0000 exactly — your lab's mean sits exactly one peer-group standard deviation above the target. Many labs use |SDI| ≥ 2 or ≥ 3 as a warning threshold to investigate, so an SDI of 1.0 here would typically be considered acceptable, though worth watching if it persists across multiple QC cycles.
Questions
What does an SDI of 1.0 actually mean?
It means your lab's average result on that control material is exactly one peer-group standard deviation higher than the peer group's own average — a moderate positive bias. Since SDI is standardized, that same 1.0 value means the same thing whether you're comparing glucose results, cholesterol results, or any other analyte, which is what makes it useful for comparing bias across very different tests.
What SDI value should trigger investigation?
There's no single mandated cutoff, but many labs follow the common convention of flagging |SDI| ≥ 2 for closer review and |SDI| ≥ 3 as a more serious concern warranting immediate investigation of calibration, reagent lots, or instrument performance. Individual laboratory policies and accreditation requirements may set stricter or more specific thresholds.
Is a negative SDI worse than a positive one?
Not inherently — the sign just indicates direction, not severity. A negative SDI means your lab's mean runs below the peer group; a positive SDI means it runs above. What matters for flagging a problem is the magnitude, |SDI|, not which direction the bias points, unless the specific analyte has a clinical reason to care more about high or low bias specifically.
How is SDI different from just looking at my lab's own control chart?
Your own control chart (like a Levey-Jennings chart) tracks your lab's results against your own historical mean and standard deviation, which is good for catching a sudden shift or drift within your own lab. SDI instead compares your lab against the wider peer group using the same method, which catches a different problem: a lab that's internally stable and consistent, but has drifted away from — or was never aligned with — the broader consensus of similar labs.
Does SDI tell you anything about precision, not just bias?
No — SDI is purely a measure of bias (how far your average sits from the peer average). It says nothing about your own lab's precision, meaning how tightly your individual results cluster around your own mean. A lab can have an excellent (near-zero) SDI while still having poor precision, or vice versa; the two are evaluated with separate statistics.