The Split Sample That Disagreed: Phase II Data Usability
A Phase II at a former industrial site collected soil from 12 borings. Laboratory A reported TCE at 0.9 mg/kg in one sample, below the state's residential screening level of 1.0 mg/kg. The consultant's QA/QC program included a field duplicate and a matrix spike/matrix spike duplicate (MS/MSD). The duplicate result was 0.2 mg/kg; the MS recovery was 45% (acceptance 70-130%). The consultant's report concludes 'no exceedance' and recommends no further action. The client's attorney asks you to review before closing.
1. Evaluate whether the 'no exceedance' conclusion is defensible given the QA/QC results.
Read the worked answer
It is not defensible.
Three problems converge:
- Low MS recovery (45%). A matrix spike recovery of 45% against a 70-130% acceptance window indicates the analytical method under-recovered TCE in that matrix, likely due to matrix interference or analyte loss during preparation. If the method recovered only about half of the spike, then a reported 0.9 mg/kg may understate the true concentration. The result could plausibly be at or above 1.0 mg/kg.
- Duplicate variability. A duplicate pair of 0.9 and 0.2 mg/kg is a large relative percent difference for a single analyte. Combined with the low MS recovery, this undermines confidence in the single result.
- Single hit, single decision. The entire no-further-action recommendation rests on one sample that is within 10% of the screening level.
The correct conclusion is that the data are not usable as reported for the no-exceedance determination. The report should flag the QA/QC failures and recommend re-sampling and re-analysis before any closure conclusion.
2. What should the consultant do procedurally, and how should the report describe the data?
Read the worked answer
Procedurally:
- Re-sample the affected boring and a nearby location with proper preservation and holding times.
- Re-analyze with dilution or a modified preparation to address matrix interference, or use an alternate method.
- Request a laboratory narrative explaining the low recovery.
- Evaluate whether the data are usable for a screening purpose only, not for a closure decision.
In the report, the consultant should:
- Present all QA/QC results, including the failed MS/MSD and the duplicate.
- State explicitly that the data quality objectives were not met for that sample.
- Qualify the result as estimated (J-flagged) or reject it.
- Avoid a 'no exceedance' statement that reads as a clean bill of health.
The attorney should be told that closing on this record leaves the client exposed to a later finding that the 0.9 mg/kg was a low bias.