The Lab Labels That Passed the Read-Back
A hospital laboratory's manifest lists a master specimen code: K7M-4Q2-9XB-1R6. A new clerk reads five printed vial labels against it:
L1 K7M-4Q2-9XB-1R6 L2 K7M-4Q2-9XB-1RB L3 K7M-4O2-9XB-1R6 L4 K7M-4Q2-9X8-1R6 L5 K7M-4Q2-9XB-IR6
The clerk reports that four of the five labels match. In the label printer's font the digit 0 and the letter O look alike, as do the digit 1 and the capital I. Codes are always written in four blocks of three characters separated by hyphens. Sixty labels must be verified in a 25-minute window by one verifier, whose average is 22 seconds per label; a complete second pass adds about 40% to that time.
1. How many labels actually match the master code exactly? Identify each mismatch by block position and character.
Read the worked answer
Compare position by position, block by block. The clerk's error is trusting overall shape rather than checking each of the twelve character slots.
| Label | Block 1 | Block 2 | Block 3 | Block 4 | Verdict |
|---|---|---|---|---|---|
| L1 | K7M ✓ | 4Q2 ✓ | 9XB ✓ | 1R6 ✓ | match |
| L2 | K7M ✓ | 4Q2 ✓ | 9XB ✓ | 1R6 vs 1R6 ✓ | final char B ≠ 6 → mismatch |
| L3 | K7M ✓ | 4O2 vs 402 → letter O for digit 0 | 9XB ✓ | 1R6 ✓ | mismatch |
| L4 | K7M ✓ | 4Q2 ✓ | 9X8 vs 9XB | 1R6 ✓ | mismatch |
| L5 | K7M ✓ | 4Q2 ✓ | 9XB ✓ | IR6 vs 1R6 → capital I for digit 1 | mismatch |
Only L1 matches. The clerk's "four of five" inverts the result: four labels are wrong, not one. Notably, both look-alike substitutions (L3 and L5) sit inside otherwise perfect codes, which is exactly why they survive a glance.
2. The clerk's errors are all single-character substitutions. Explain why two of them are far more likely to be missed than the other two, and describe a checking procedure that would catch them.
Read the worked answer
Why L3 and L5 are the dangerous ones: they are glyph confusions (O/0 and I/1), where the printed shapes are near-identical in this font. A reader comparing by pattern recognition sees the same squiggle and registers agreement.
Why L2 and L4 are safer: B/6 and 8/B differ in curvature and closure, so they look different even at a glance; they are likelier to be caught (the clerk did flag one of them).
Procedure that removes the failure mode:
- Decode, don't scan - read the label aloud character by character ("one, R, six"), forcing the reader to name the symbol type rather than its shape.
- Reverse direction on the second pass (right to left), which breaks the pattern-matching habit.
- Ambiguous glyph ban - codes should never use 0/O or 1/I in the same alphabet; that is a design fix, not a vigilance fix.
- Independent verification for any label where a single character differs, since one substituted character in a specimen ID is a patient-safety event, not a typo.
3. Does the 25-minute window hold for one verifier? Show the arithmetic for a single pass and for a full second pass, and state the minimum staffing.
Read the worked answer
Single pass: 60 labels × 22 s = 1,320 s = 22.0 min. Against a 25-minute window this fits, with 3.0 min spare (about 8 labels of slack).
With a full second pass: 1,320 × 1.40 = 1,848 s = 30.8 min. That exceeds the window by 5.8 min - the plan fails.
Staffing options:
- Two verifiers working in parallel on different labels: the second pass is not independent (a verifier re-reading their own work), so this is weaker than it looks.
- Split labels between two verifiers and cross-check: 30 labels each, 22 min single pass per person plus a 40% cross-check on 30 labels ≈ 22 + 6.2 ≈ 28 min - still over.
- A dedicated second verifier covering all 60: the first pass takes 22 min, the second 8.8 min of added work, and the two run concurrently, so the wall-clock requirement is roughly max(22, 8.8) + handoff ≈ 25 min - only just feasible, with no margin.
Judgement: the honest recommendation is that the window is one verifier short for a two-pass process. Either extend the window or accept a single pass with mandatory character-by-character decoding, plus the code-alphabet fix that removes the O/0 and I/1 confusions permanently.