Notice. Illustration and research prototype. Not clinically validated, not a medical device, not for patient care. Example data are fictional.
Live demonstration. The sliders and test cases work — every score recomputes in your browser. The data is generated, no patient is involved, and nothing here has been verified by a clinician.

stewardship.yaml

Configuration · stewardship.yaml · 200 lines · editable in place

ANTIMICROBIAL DE-ESCALATION ============================================================================ Idea 38. Nobody disagrees that day-3 review should happen. The patient started on meropenem at 2 AM on Monday is still on it on Thursday because nobody revisited it — the failure is temporal, not clinical.

THREE QUESTIONS, ASKED ON A CLOCK --------------------------------- Day 3 cultures are back. Can this be narrowed? Day 7 most indications are complete. Can this stop? Always is the spectrum still justified by what grew?

THE ASYMMETRY THAT MATTERS -------------------------- Procalcitonin may support a decision to STOP. It must never be used to justify CONTINUING. If the clinical picture says stop and the procalcitonin is still raised, stop — a biomarker that only ever extends therapy is worse than no biomarker.

ISCCM rated procalcitonin at 6.7% feasibility in Indian settings, so it cannot be a gate here. CRP scored 93.3% and is the practical alternative.

Sources: Procalcitonin-guided stewardship trials; ISCCM Position Statement on Sepsis in Resource-limited Settings, IJCCM 2024; standard stewardship programmes.

meta
version0.1.0
ownerUNASSIGNED — set to the responsible clinician or microbiologist
verified bynot set
reviews
- idday_3
nameDay 3 review
day3
why
Cultures are back. 64% of de-escalation that happens at all happens by
day 3 — the rest is drift.
asks
"What grew, and does the current agent still fit it?"
"If nothing grew, is there still a reason to treat?"
"Can the spectrum be narrowed today?"
- idday_7
nameDay 7 stop review
day7
why
Most indications are complete by seven days. Longer courses need a
stated reason, not momentum.
asks
"Is the source controlled?"
"What is the planned total duration, and why?"
spectrum
very broad
tier1
drugs[meropenem, imipenem, colistin, polymyxin-b, tigecycline,
linezolid, daptomycin, ceftazidime-avibactam]
noteReserve agents. Continuing past day 3 without an organism needs a reason.
broad
tier2
drugs[piperacillin-tazobactam, cefepime, vancomycin, teicoplanin,
amikacin, levofloxacin, ciprofloxacin]
moderate
tier3
drugs[ceftriaxone, cefotaxime, ceftazidime, gentamicin, clindamycin,
azithromycin, doxycycline]
narrow
tier4
drugs[amoxicillin, ampicillin, amoxicillin-clavulanate, cloxacillin,
flucloxacillin, cefazolin, metronidazole, cotrimoxazole,
nitrofurantoin]
durations
- indicationCommunity-acquired pneumonia
days5
- indicationVentilator-associated pneumonia
days7
- indicationIntra-abdominal infection with source control
days4
- indicationUncomplicated bloodstream infection, gram-negative
days7
- indicationStaphylococcus aureus bacteraemia
days14
noteLonger by design. 14 days is a minimum, not a target.
- indicationUrinary tract infection
days5
- indicationSkin and soft tissue
days5
- indicationUndifferentiated sepsis, cultures negative, patient improved
days5
noteThe commonest reason for a long course is that nobody stopped it.
default duration days7
biomarkers
procalcitonin
availablefalse
ISCCM feasibility 6.7% in Indian settings
stop below0.25
stop on drop percent80
asymmetry
May support stopping. Must never be used to justify continuing. If the
clinical picture says stop and the procalcitonin is still raised, stop.
noteISCCM rated feasibility at 6.7%. Not a gate.
crp
availabletrue
note
ISCCM rated CRP the most feasible biomarker at 93.3%. A falling trend
supports stopping; a raised value alone does not justify continuing.
checks
- idday_3_review_due
severitywarn
textDay 3 antimicrobial review due
whyCultures are back. This is where de-escalation happens or does not.
- idbroad_without_organism
severitywarn
textBroad-spectrum continuing with no organism identified
why
Past day 3 with negative cultures, a reserve agent needs a stated reason
rather than momentum.
- idnot_narrowed_to_organism
severitywarn
textSpectrum broader than the organism requires
why
An organism is known and a narrower agent would cover it. This is the
definition of de-escalation and the commonest missed opportunity.
- idduration_exceeded
severitywarn
textPast the typical duration for this indication
whyLonger courses need a stated reason, not momentum.
- idno_indication_recorded
severitywarn
textNo indication recorded
why
Without an indication there is no duration to measure against and no
basis for stopping.
- idno_cultures_sent
severitywarn
textAntimicrobials started without cultures
why
ISCCMcultures before antibiotics, but do not delay antibiotics for
them. Sent late is better than not sent — without them de-escalation is
guesswork.
- idredundant_cover
severitywarn
textOverlapping spectrum
whyTwo agents covering the same organisms adds toxicity, not cover.
redundant
- drugs[meropenem, piperacillin-tazobactam]
whyBoth cover gram-negatives and anaerobes. One is enough.
- drugs[meropenem, metronidazole]
whyMeropenem already covers anaerobes.
- drugs[piperacillin-tazobactam, metronidazole]
whyPiperacillin-tazobactam already covers anaerobes.
- drugs[vancomycin, linezolid]
whyBoth cover gram-positives including MRSA.
- drugs[vancomycin, teicoplanin]
whySame class.