ANTIBIOGRAM_MISREAD_R medical data_error ai_generated true

AI misreads an antibiogram report, interpreting 'R' as 'Resistant' for an organism that is actually susceptible to the antibiotic, leading to inappropriate antibiotic selection

ID: medical/antibiotic-sensitivity-misread

Also available as: JSON · Markdown · 中文
80%Fix Rate
82%Confidence
1Evidence
2023-12-05First Seen

Root Cause

Antibiogram reports use standard abbreviations: S (susceptible), I (intermediate), R (resistant). AI may misinterpret 'R' as 'Resistant' when the report uses a different abbreviation scheme (e.g., 'R' for 'Reduced susceptibility' in some systems, or the organism is actually susceptible but the AI misaligns columns).

generic

中文

抗菌药物敏感性报告使用标准缩写：S（敏感）、I（中介）、R（耐药）。AI可能将‘R’误解为‘耐药’，而报告使用了不同的缩写方案（例如，某些系统中‘R’表示‘敏感性降低’，或者菌株实际上敏感但AI对齐了错误列）。

Official Documentation

https://clsi.org/standards/products/microbiology/documents/m100/

Workarounds

  1. 90% success Cross-check the antibiogram with the organism's known susceptibility patterns. Use a standard reference (e.g., CLSI M100 or EUCAST breakpoints) to verify. Example: if the report shows 'E. coli: Ampicillin R' but the organism is known to be susceptible, re-read the report carefully for column alignment or consult the lab for clarification. 
    Cross-check the antibiogram with the organism's known susceptibility patterns. Use a standard reference (e.g., CLSI M100 or EUCAST breakpoints) to verify. Example: if the report shows 'E. coli: Ampicillin R' but the organism is known to be susceptible, re-read the report carefully for column alignment or consult the lab for clarification.
  2. 85% success If the report uses non-standard abbreviations (e.g., 'R' for 'Reduced'), re-interpret using the lab's specific key. Contact the microbiology lab if unsure. 
    If the report uses non-standard abbreviations (e.g., 'R' for 'Reduced'), re-interpret using the lab's specific key. Contact the microbiology lab if unsure.

中文步骤

  1. Cross-check the antibiogram with the organism's known susceptibility patterns. Use a standard reference (e.g., CLSI M100 or EUCAST breakpoints) to verify. Example: if the report shows 'E. coli: Ampicillin R' but the organism is known to be susceptible, re-read the report carefully for column alignment or consult the lab for clarification.
  2. If the report uses non-standard abbreviations (e.g., 'R' for 'Reduced'), re-interpret using the lab's specific key. Contact the microbiology lab if unsure.

Dead Ends

Common approaches that don't work:

  1. 90% fail

    In some systems (e.g., EUCAST), 'R' may indicate 'Reduced susceptibility' or the report may have a different column order; misinterpreting leads to using broader-spectrum antibiotics unnecessarily, promoting resistance.

  2. 85% fail

    Empiric therapy may be inappropriate for the specific organism, leading to treatment failure or adverse effects.

  3. 70% fail

    Delays treatment unnecessarily; the original report may be correct but misread.