Reading your hospital’s antibiogram —
and thinking like a steward
Empiric antibiotics are among the commonest consequential decisions an intern makes — and a reference that should inform them — beside the syndrome, the severity, the source, allergies, renal function, and the patient’s own culture history — is a document most interns have never opened: the hospital’s own antibiogram, the yearly map of what actually grows here and which agents remain active in vitro. Activity on paper, not a promise of cure. This session teaches the map’s grammar once, nationally; the map itself is irreducibly yours.
Why this session
Resistance is local. The organism your patient grew last night lives in this building’s ecology, shaped by this region’s prescribing — which is why the same national guideline carries a footnote the textbook cannot fill in: adjust to local susceptibility patterns. The antibiogram is that adjustment, compiled annually by your microbiology laboratory under national laboratory standards, and the intern who can read it makes better empiric choices on night one than the intern quoting a chapter written for the whole country. The session’s second purpose is the habit set around the choice — stewardship — taught not as a compliance lecture but as what it is: the craft of treating this patient well while keeping the drugs working for the next one.
What interns leave able to do
- Read the antibiogram’s grid: organisms, agents, percent-susceptible, isolate counts — and its caveats.
- Use it for the empiric question: “given this syndrome and this hospital, what covers the likely organisms well enough tonight?”
- Explain the inferred-susceptibility idea — why the lab’s tested agents speak for untested relatives, and where that translation lives.
- Run the steward’s loop: cultures before antibiotics — never delaying them — narrow on data, name the duration, review everything at 48–72 hours.
- Know the local machinery: where the antibiogram lives, what requires stewardship or ID approval, and who to call.
The case
Night float. Two patients need antibiotics started now.
Patient one: a 62-year-old woman with pyelonephritis — febrile, flank pain, nitrite-positive urine, cultures drawn. Your question bank’s answer and half the internet reach for a fluoroquinolone.
Patient two: a 58-year-old man with community-acquired pneumonia, moderately ill, admitted to the floor. The national default is familiar. His chart shows a hospitalization here four months ago.
On the workroom wall — or three clicks deep in the intranet — is the hospital’s antibiogram. Suppose it shows what many hospitals’ now show: E. coli susceptibility to fluoroquinolones down in the 70s, while other first-line options hold in the 90s.
Does the textbook answer survive contact with the local grid? What do you start, why — and what do you do differently because cultures are already cooking?
The teaching beat inside the case: the question bank is calibrated to the nation; your patient is admitted to a building. Board answers and local empiric choices can legitimately differ — which is the boards session’s two-registers rule running in the opposite direction: know the national answer for the exam, and the local map for tonight’s order.
Reading the grid
- The anatomy: organisms down the rows, antimicrobials across the columns, and in each cell the percentage of this hospital’s isolates — usually last calendar year’s — that tested susceptible. An isolate count rides with each organism; treat it like the denominator it is.
- The empiric read: for a syndrome, shortlist the likely organisms, then scan their columns — you are asking which regimen gives high enough coverage for this severity. A percentage in the 70s may be tolerable for a stable patient with cultures pending and a follow-up plan; it is not where you park a septic one. Severity buys breadth; stability buys patience.
- The caveats, which are half the literacy: small isolate counts make noisy percentages — be humble below the threshold your lab flags. The grid summarizes everyone’s isolates: recently hospitalized, recently treated, and chronically colonized patients (your pneumonia patient with the recent admission) sit in a worse-than-average corner of it. Some hospitals publish unit-specific or ICU antibiograms because the ICU’s ecology is its own — know whether yours does. And the antibiogram is a probability tool for the empiric window only: the moment your patient’s own susceptibilities return, the grid retires from the case.
- Inferred susceptibility — the translation layer: the lab does not test every drug against every isolate; it tests representatives, and standardized rules let results for tested agents speak for untested relatives in the class. Your microbiology lab publishes its version of this translation — often as a companion sheet to the antibiogram — and the practical intern skill is knowing it exists: when the drug you want is not on the report, the answer may already be there under a cousin’s name, and the stewardship pharmacist can read it with you in thirty seconds.
The steward’s habits — the loop that travels everywhere
- Cultures before antibiotics — without delaying the antibiotics. The rapid-response session’s sepsis reflex, generalized: the culture you send tonight is what lets you be precise on Thursday.
- Empiric breadth is a loan, not a gift. Start as broad as the severity and the local map genuinely require — then narrow the moment the data allow. De-escalation on day two is not undoing your choice; it is completing it.
- Every antibiotic order carries three named things: the suspected source, the drug’s plan, and the intended duration. “Continue antibiotics” with no stop date is how seven-day courses become seventeen.
- The 48–72-hour review, every patient, every time: cultures back? Narrow. No infection after all? Stop — stopping is a stewardship act, and often the hardest one. Not improving? Rethink the source before broadening the net.
- The allies: the stewardship pharmacist and the ID service are collaborators, not gatekeepers — and the restricted-antimicrobial list (most hospitals have one) is the formal version of a good habit: the biggest guns get a conversation first.
The local organ — what your program plugs in
This page is deliberately incomplete. To run the session, bring: your hospital’s current antibiogram (and the unit-specific ones if they exist), the stewardship program’s empiric-therapy guidance — many publish a one-page memo pairing common syndromes with locally preferred regimens, which is the antibiogram pre-digested — the restricted-antimicrobial list and its approval pathway, and the micro lab’s inferred-susceptibility reference. Walk the actual documents, project the actual grid, and have the interns answer the case from it. Where the antibiogram lives — the intranet path, the workroom wall, the pharmacy page — is itself a required localization fact: a map nobody can find protects nobody.
Pocket card
- The textbook is national; your patient is admitted to a building. Read the local map.
- Severity buys breadth; stability buys patience. Check the isolate count.
- Cultures first — then narrow the moment data allow. De-escalation completes the choice.
- Every antibiotic order: source + drug + duration. No orphan “continue.”
- 48–72-hour review, every patient. Stopping is a stewardship act.
- Know where the antibiogram lives — and the stewardship pharmacist’s name.
Notes
This page is a teaching framework, not clinical guidance: it names no empiric regimen for any condition, and the illustrative percentages in the case are hypothetical. Antibiotic choices belong to current national guidelines interpreted through your hospital’s antibiogram, stewardship guidance, and ID consultation. Last reviewed July 2026.