Reading antibiograms in context: what resistance data can and cannot show

Ask who is represented

WHO GLASS supports standardised collection, analysis and interpretation of resistance data with broader context.^[1] A hospital antibiogram should make its specimen population, period and setting visible. Frequently sampled patients should not automatically be treated as representative of every patient receiving care.

Recognise important limitations

The GLASS manual explains biases in routine clinical sampling, including greater testing of severe illness or treatment failure and sampling after antibiotics begin.^[2] Interpretation also requires attention to infection versus colonisation or contamination, and to the difference between a complete testing database and selectively reported results.

Make the report interpretable

ICORP suggests that microbiology and stewardship teams jointly add a short methods note: data selection, repeated-isolate handling, testing limitations and changes since the previous report. Avoid ranking units using unadjusted percentages. Before comparison, ask whether patients and access to testing are reasonably comparable.

Inform decisions without replacing assessment

Suggested use is to identify review priorities alongside antibiotic use and infection events, then improve the data behind them. An aggregate table is not an individual isolate susceptibility result or an automatic prescribing instruction. Treatment requires the clinical team to consider the patient, infection site, history and actual results.