Required section · Section 4 of 6

Working the five-level reagent-lot verification

Five-level commercial calibration-verification material is tested under this case’s approved plan before the new reagent lot is released. The levels span low through high concentrations, including the clinical decision zone. The case acceptance criterion is ±8% or ±6 mg/dL, whichever is greater.

Level 1 is the one worth slowing down for. The assigned value is 5 mg/dL and the observed mean is 6 mg/dL, a difference of only 1 mg/dL but a striking-looking +20% difference. Read as a percentage alone, that looks like a failure. But the acceptance rule takes whichever limit is greater, and 8% of 5 mg/dL is only 0.4 mg/dL, so the governing limit here is the absolute 6 mg/dL term. A 1 mg/dL difference is well inside 6 mg/dL, so level 1 passes. This is why a criterion expressed only as a percentage misleads near the low end of a method's range, and why the low-concentration challenge point needs its own reading, not just a glance at percent recovery.

The remaining four levels are read the same way, difference and percent difference each compared against the calculated limit for that level. Level 2 (assigned 112, observed 105) differs by -7, well inside its 9.0 mg/dL limit. Level 3, inside the ADA decision zone (assigned 124, observed 130), differs by +6, inside its 9.9 mg/dL limit. Level 4 (assigned 410, observed 393) differs by -17, inside its 32.8 mg/dL limit. Level 5, near the AMR ceiling (assigned 705, observed 668), differs by -37, inside its 56.4 mg/dL limit. All five levels pass.

The residual plot shows the observed residuals at these five levels, including the failed high point in the separate post-maintenance scenario. These five verification levels neither establish linearity nor justify choosing a polynomial; they identify observations that need review under the approved plan.

Read the low-concentration point against the absolute term of the limit before concluding a large percent difference is a failure.

Scatter chart of five observed versus assigned glucose values. The light-blue case acceptance region is ±6 mg/dL through 75 mg/dL, then ±8%; Level 1 at 5 to 6 mg/dL is inside the absolute allowance. — Figure 1Observed versus assigned glucose values with the case acceptance region: ±6 mg/dL at low concentrations, changing to ±8% above 75 mg/dL.

Chart of five observed residuals from plus 6 to minus 37 mg/dL, with a zero reference line. The level 5 residual of minus 37 mg/dL is highlighted as the largest and still passes the case criterion; five verification points do not establish linearity. — Figure 2Observed residuals (observed minus assigned) across five verification levels; this set does not establish linearity or select a polynomial.

Five-level calibration verification, new glucose reagent lot (mg/dL)
Level	Assigned	Observed mean	Difference	% Difference	Acceptance limit	Result
1 (low)	5	6	+1	+20.0%	±6 mg/dL (absolute term governs)	Pass
2 (low-mid)	112	105	-7	-6.3%	±9.0 mg/dL	Pass
3 (mid, decision zone)	124	130	+6	+4.8%	±9.9 mg/dL	Pass
4 (mid-high)	410	393	-17	-4.1%	±32.8 mg/dL	Pass
5 (high)	705	668	-37	-5.2%	±56.4 mg/dL	Pass

Residuals (observed minus assigned) by level
Level	Assigned (mg/dL)	Residual (mg/dL)
1	5	+1
2	112	-7
3	124	+6
4	410	-17
5	705	-37

Knowledge checks

Reading and checks are open. Sign in only to save.

Knowledge check 1

Which analyses does a calibration-verification/linearity evaluation use to interpret the observed-versus-assigned data? Select all that apply.

Choose at least 3 options.

Recovery, expressed as percent or absolute difference from the assigned value

Linear regression of observed versus assigned value

Residual analysis (observed minus predicted or assigned)

A single t-test comparing the mean of all observed values to the mean of all assigned values

Discarding any point that looks unusual before analysis

Section status

Finish this section

Reading and checks are open. Sign in only to save.

The module finishes after every required section is marked done and every check in those sections is correct.