Evaluate proficiency testing performance using standardized deviation from expected values.
Last updated: March 2026
Measurement result from lab
Certified reference material
Known proficiency SD
The Standard Deviation Index (SDI) is a statistical metric used in proficiency testing to evaluate how well a laboratory performs relative to expected values. It measures the deviation of a measurement from a reference value in terms of standard deviations. SDI standardizes performance across different measurement scales and uncertainty levels, allowing consistent comparison of lab performance.
SDI is calculated as: SDI = (Observed − Expected) / SD. An SDI of 0 means perfect agreement with the reference. Positive SDI indicates the measurement is higher than expected; negative indicates lower. The absolute value |SDI| determines performance: |SDI| ≤ 1 is acceptable, 1 < |SDI| ≤ 2 is marginal (warrants investigation), and |SDI| > 2 is unacceptable (corrective action needed).
SDI is widely used in clinical laboratories, environmental testing, and food safety programs for ongoing quality assurance and proficiency assessment.
Clinical Lab: Glucose Measurement
Perfect agreement between observed and expected value. Indicates the measurement exactly matches the reference material. In practice, SDI = 0 is rare due to natural measurement variation.
Yes. Negative SDI means the observed value is below the expected value (measurement is too low). The interpretation focuses on absolute value {'|SDI|'}, not the sign.
SD normalizes the deviation to account for measurement uncertainty. Different parameters have different acceptable variations. Using SD makes SDI comparable across different analytes and methods.
Frequency depends on regulatory requirements and risk level. Clinical labs typically perform it monthly per CLIA. Environmental and food labs follow EPA/AOAC guidelines, often quarterly or semi-annually.
Investigate the cause: calibration drift, reagent problems, operator error, or instrument maintenance needs. Document corrective actions and retest with fresh QC material.
They're very similar. Both standardize deviation by dividing by SD. The main difference: SDI uses proficiency program SD; Z-scores use dataset SD. SDI is preferred for lab quality assessment.
Related Tools
Evaluate classification models.
Classification metrics.
Diagnostic test accuracy.
Diagnostic test performance.
Updated diagnostic probability.
Risk ratio comparison.