Troubleshooting / Common Issues
Paired difference test fails — observed difference exceeds allowable limit
Possible causes
- –The interferent at the tested concentration causes a real, clinically significant bias in the measurement procedure
- –The allowable difference is set too tight relative to the magnitude of the interference effect
- –High within-group variability inflates the estimated difference
Recommended action
Confirm that the allowable difference reflects the clinically relevant acceptance criterion for this analyte and decision point. If the result correctly identifies significant interference, proceed to a dose-response experiment to characterize the relationship and identify the maximum interferent concentration that remains within acceptable limits. If within-group variability is high, verify instrument precision and sample preparation consistency before repeating the study.
Dose-response regression model fits poorly — measured values do not follow the curve
Possible causes
- –Non-monotonic interference relationship — the interferent has both positive and negative effects at different concentrations
- –One or more concentration levels have high measurement variability that distorts the fit
- –Sample preparation at intermediate concentrations introduced error (e.g., dilution error when preparing 25%, 50%, 75% levels)
Recommended action
Inspect the scatter plot for the pattern of residuals. If one intermediate level appears to be an outlier from the overall trend, review the sample preparation for that level. If the relationship is genuinely non-monotonic, a simple linear or quadratic model may not be appropriate — consult CLSI EP07 for guidance on characterizing non-standard interference relationships.
Interference Summary shows a very low maximum safe interferent concentration
Possible causes
- –The measurement procedure is highly sensitive to this interferent — even moderate concentrations cause significant bias
- –The allowable difference is very tight relative to the interference effect magnitude
- –The regression model's 95% CI is wide due to high within-condition variability
Recommended action
Compare the maximum safe interferent concentration against the concentrations typically found in the laboratory's patient population (consult CLSI EP37 for reference ranges of common interferents). If the safe concentration is below the typical range seen in patients, document the interference in the method's limitations, establish a specimen rejection criteria or dilution protocol for flagged specimens, and notify clinical staff of the potential for erroneously biased results in affected patients.
Control group SD is high, making it difficult to detect interference
Possible causes
- –The measurement procedure's repeatability at this analyte concentration is inherently variable
- –Sample matrix instability — the specimen pool used for the control group may have degraded between replicates
- –Instrument instability during the measurement session
Recommended action
High control variability directly inflates the confidence interval width and may mask true interference effects. Verify instrument precision before the study using recent precision data. Use fresh, stable sample materials and run all replicates within a single session to minimize between-replicate drift. If repeatability is genuinely poor at this concentration, more replicates will be required — use the replicate calculator with the observed repeatability SD to determine the minimum needed for adequate statistical power.