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.

infection.yaml

Configuration · infection.yaml · 288 lines · editable in place

INFECTION SURVEILLANCE ============================================================================ Idea 75. Line infections and VAP counted continuously rather than by retrospective chart audit.

WHY CONTINUOUS COUNTING CHANGES THE NUMBER ------------------------------------------ Retrospective audit undercounts, and it undercounts unevenly. A busy month is audited less thoroughly than a quiet one, which means the surveillance figure is lowest exactly when the true rate is highest. Continuous counting from what is already recorded removes that particular distortion.

It does not remove all of them. See the denominator note below.

THE DENOMINATOR IS WHERE THE DISHONESTY USUALLY LIVES ----------------------------------------------------- Infection rates are reported per 1000 device-days. The numerator is watched closely because it is the number people are judged on. The denominator is where a unit can quietly improve its rate without preventing a single infection — by counting device-days generously, or by removing lines late so the denominator grows.

So this engine computes and reports BOTH, always, and never emits a rate without the counts that produced it.

SURVEILLANCE DEFINITIONS ARE NOT CLINICAL DIAGNOSES --------------------------------------------------- A patient can meet the CLABSI surveillance definition without having a line infection, and can have one without meeting it. The definitions exist to be counted consistently across units, not to be right about individuals.

The engine says so on every finding, because a surveillance flag read as a diagnosis leads to antibiotics, and that is a real harm arising from a counting tool.

Sources: CDC/NHSN surveillance definitions for CLABSI, CAUTI, and VAE. Structure only — every definition must be checked against the version the unit reports under before use.

meta
version0.1.0
ownerUNASSIGNED — set to the infection control lead
verified bynot set
boundary
Surveillance definitions count consistently across units. They are not
clinical diagnoses and must not be treated as one.
denominator note
A rate is never emitted without the numerator and denominator that
produced it. The denominator is where a unit can improve its reported
rate without preventing any infection.
definitions
- idclabsi
nameCentral line-associated bloodstream infection
devicecentral_line
minimum device days2
note
The line must have been in place for more than 2 calendar days, with
day 1 the day of insertion. The first attributable day is therefore
day 3.
pathways
NHSN defines TWO ALTERNATIVE PATHWAYS, not a list of co-requirements. LCBI-1 and LCBI-2 are each sufficient on their own. An earlier version of this file listed them flat and required all of them, which scored a patient with Klebsiella in one bottle as NOT a CLABSI — the commonest true event, silently dropped. Alternatives must stay alternatives.
- idlcbi_1
nameLCBI-1 — recognised pathogen
all of
- idpositive_culture
detailRecognised pathogen cultured from one or more blood cultures
note
One bottle is enough for a recognised pathogen. No symptom criterion
is required.
- idlcbi_2
nameLCBI-2 — common commensal
all of
- idcommensal_two_cultures
detail
Common commensal from two or more blood cultures drawn on
separate occasions. One positive commensal culture is
contamination until proven otherwise.
- idinfection_signs
detail
Fever above 38C, chills, or hypotension — required for this
pathway and not for LCBI-1
exclusions
- idsecondary_bloodstream
detailBloodstream infection secondary to another documented site
why it matters
The commonest source of over-counting. A pneumonia seeding the
blood is not a line infection, and counting it as one both inflates
the rate and misdirects the response.
- idmucosal_barrier
detailMucosal barrier injury in neutropenic patients
why it mattersGut translocation, not line contamination
known to get wrong
Under-counts in units that culture rarely, because an uncultured
infection is not an infection by this definition. A falling rate can
mean fewer infections or fewer cultures, and the definition cannot tell
them apart.
- idcauti
nameCatheter-associated urinary tract infection
deviceurinary_catheter
minimum device days2
note
Catheter in place more than 2 calendar days, day 1 being the day of
insertion. The first attributable day is day 3. Symptoms assessable only
after removal are marked below, because they cannot be elicited from a
catheterised patient.
pathways
- idsuti_1a
nameCatheter in place, or removed the day of or day before
all of
- idpositive_urine_culture
detail
10^5 CFU/mL or more, with no more than two species. NHSN also
recognises a 10^3 to under 10^5 pathway in defined
circumstances; confirm which version the unit reports under.
- idsymptoms_catheterised
detail
Fever above 38C, suprapubic tenderness, or costovertebral angle
pain or tenderness. Urgency, frequency and dysuria are NOT
usable while the catheter is in place.
- idsuti_1b
nameCatheter removed more than one day ago
all of
- idpositive_urine_culture
detail10^5 CFU/mL or more, no more than two species
- idsymptoms_post_removal
detail
Fever above 38C, suprapubic tenderness, costovertebral angle
pain, OR urgency, frequency, or dysuria — the last three become
usable only once the catheter is out.
exclusions
- idasymptomatic_bacteriuria
detailPositive culture with no symptoms
why it matters
Not a CAUTI, and treating it drives resistance without benefit.
The single most consequential exclusion in this whole config.
known to get wrong
Fever in a catheterised ICU patient has many causes. The definition
attributes it to the catheter whenever a culture is positive, which
over-attributes in exactly the population with the most competing
sources of fever.
- idvae
nameVentilator-associated event
deviceventilator
minimum device days2
note
Ventilated more than 2 calendar days; the earliest event date is day 3.
Requires 2 days of stable or decreasing daily minimum FiO2 and PEEP,
followed by 2 days of worsening. The baseline period is what makes it
objective.
tiers
VAE IS A HIERARCHY, NOT A SET OF CO-REQUIREMENTS. VAC is the base tier; IVAC adds infection signs; PVAP adds microbiological evidence. Each is nested inside the last. Requiring all three at once would report only PVAP — the narrowest tier — while labelling the output "ventilator-associated event", which under-counts by a large multiple. The tier reached is reported explicitly.
- idvac
nameVAC — ventilator-associated condition
level1
all of
- idoxygenation_worsening
detail
A rise in DAILY MINIMUM FiO2 of 0.20 (20 points) or more, OR a
rise in DAILY MINIMUM PEEP of 3 cmH2O or more, sustained 2 or
more calendar days, after 2 days of stable or decreasing values.
why daily minimum
"Daily minimum" is load-bearing. Using the set FiO2 or any
transient value over-calls VAC substantially, because brief
escalations for suctioning and procedures are routine.
- idivac
nameIVAC — infection-related ventilator-associated complication
level2
requires tiervac
all of
- idinfection_signs
detail
Temperature above 38C or below 36C, or white cell count 12,000
or above or 4,000 or below, AND a new antimicrobial started and
continued 4 or more days.
- idpvap
namePVAP — possible ventilator-associated pneumonia
level3
requires tierivac
all of
- idmicrobiology
detail
Purulent respiratory secretions or a qualifying positive
respiratory culture.
exclusions
- idno_stable_baseline
detailNo 2-day stable period to measure against
why it matters
A patient deteriorating from the moment of intubation has no
baseline, and the definition cannot be applied — which is not the
same as the definition being met and should never be recorded as
a negative.
known to get wrong
VAE was designed to be objective and countable, and it succeeded. It
captures fluid overload, atelectasis, and ARDS progression alongside
pneumonia. It is a good surveillance measure and a poor diagnosis.
bundles
- idclabsi_prevention
devicecentral_line
elements
- iddaily_necessity_review
elementAsk daily whether the line is still needed
why skipped
The single most effective element and the easiest to skip, because
nothing goes wrong on the day it is skipped.
- idchlorhexidine_dressing
elementChlorhexidine-impregnated dressing, intact
- idhub_disinfection
elementDisinfect the hub before every access
why skippedTakes 15 seconds, happens 20 times a day
- iddressing_intact
elementDressing clean, dry, and dated
- idcauti_prevention
deviceurinary_catheter
elements
- iddaily_necessity_review
elementAsk daily whether the catheter is still needed
why skipped
Convenience for staff is an unstated indication in most units, and
it is never written in the notes.
- idclosed_system
elementClosed drainage system, unbroken
- idbag_below_bladder
elementBag below bladder level, off the floor
- idvap_prevention
deviceventilator
elements
- idhead_elevation
elementHead of bed at 30-45 degrees
why skippedLowered for a procedure and not raised again
- idoral_care
elementChlorhexidine oral care
- idsedation_interruption
elementDaily sedation interruption
- idcuff_pressure
elementCuff pressure 20-30 cmH2O, checked
- idsubglottic_drainage
elementSubglottic secretion drainage where the tube allows
reporting
rate per1000
rate unitdevice-days
minimum device days for rate50
small number note
Below 50 device-days a rate is arithmetic, not information. One infection
in 30 device-days reads as 33 per 1000 and means almost nothing. The
engine reports the counts and withholds the rate.