STATISTICS
Detection capability at the low end: three related but different limits
In this fifth article in his latest series, Stephen MacDonald explores the intricacies of dealing with three limits – blank, detection and quantitation – and common misconceptions, limitations and cautions to their interpretation, and offers readers some practical take home points.
At the low end of a quantitative measurement procedure, three questions are often collapsed into one. What result could a blank sample give? What concentration can be detected reliably? At what concentration is the numerical result good enough to use? The limit of blank (LoB), limit of detection (LoD) and limit of quantitation (LoQ) answer these questions, respectively. Treating them as interchangeable can turn an analytical threshold into an unjustified clinical claim.
This article explains the statistical
logic, practical study design and reporting consequences of these limits. It concentrates on continuous quantitative procedures. Qualitative and molecular methods can use related probability based ideas, but may need method-specific models. Linearity and the wider reportable range are addressed in the previous article (Part 4, September issue) in this series, and are not repeated here.
Why the distinction matters A low numerical result may be analytically distinguishable from a blank but still too imprecise or biased for interpretation. Conversely, a laboratory can truncate a result below its LoQ even though the procedure detects analyte there with high probability. These are not contradictions. Detection and quantitation are different claims.
The distinction affects verification
plans, laboratory information system rules, comments atached to results and
Verification should account for lot-to-lot or site-to-site variation if such conditions mater to the intended claim.
October 2026
WWW.PATHOLOGYINPRACTICE.COM 17
the wording used in specifications. A statement such as “analytical sensitivity 0.2” is incomplete unless it identifies the limit, units, calculation or verification approach, and the probability or performance criterion. The same number cannot safely be used as an LoB, LoD and LoQ simply because it appears near the botom of a kit insert. Further information is shown in Table 1.
Three related but different limits Limit of blank
The LoB describes the upper tail of results produced by samples that contain no measurand (see Fig 1). It is a decision boundary for the blank distribution. If the one-sided false-positive probability is set at 5%, approximately 95% of genuine blank results should fall at or below the
CC BY-SA 4.0 Wikimedia Commons
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56
