NAP · Evidence & methodology
How NAP is grounded.
NAP is a browser-based teaching simulator, calibrated to what makes a lesson land rather than to perfect realism. The professional sources on this page informed the design of its simulated physiology, teaching defaults, dynamics, morphologies and monitor behaviours — and the simulation-education evidence below explains why calibrating fidelity to the learning objective is the right target. This is the honest evidence base behind those choices, presented so educators and institutions can see exactly what each source did, and did not, establish.
Scope and claim boundary
These sources do not establish that NAP is clinically validated, diagnostically accurate, equivalent to a real monitor, or suitable for patient care. NAP is a teaching simulator, not a medical device or clinical decision-support system.
A citation means a source informed a design choice — it does not mean the source endorses NAP. Every module-level fidelity claim remains gated on independent external clinical face-validation, a separate and still-open step.
Fidelity serves learning.
NAP treats fidelity as a means, not an end. Its target is functional correspondence with the learning objective — recognisable monitor behaviour, the salient physiological relationships and instructor control — not maximum engineering resemblance to a particular patient, monitor brand or whole clinical room. The simulation-education literature does not make fidelity a simple ladder: added realism helps when the objective needs it, while irrelevant complexity can consume attention and bury the very cue being taught. So NAP's monitor-only scope, instructor-authored physiology and clamp-don't-evolve behaviour are deliberate instructional choices — simplify to teach, never to mislead. This rationale explains NAP's design target; it does not, on its own, validate clinical accuracy or unlock any module's fidelity claim.
Hamstra et al., 2014 — Reconsidering fidelity in simulation-based training
The distinction between physical resemblance and functional task alignment — matching realism to the applied learning context, not to a specific device.
Norman, Dore and Grierson, 2012 — The minimal relationship between simulation fidelity and transfer of learning
Across the reviewed direct comparisons, higher-fidelity simulation was rarely superior to lower-fidelity — there is no simple "more realism, more learning" rule.
Dieckmann, Gaba and Rall, 2007 — Deepening the theoretical foundations of patient simulation as social practice
Goal-oriented matching of realism to the desired outcome, and the recognition that more physical realism does not uniformly improve learning.
Cook et al., 2013 — Comparative effectiveness of instructional design features in simulation-based education
The weight of instructional design — feedback, interactivity, practice and difficulty — over technology or resemblance alone.
Issenberg et al., 2005 — Features and uses of high-fidelity medical simulations that lead to effective learning
The conditions under which simulation supports learning: feedback, repetitive practice, curriculum integration, appropriate difficulty and explicit outcomes.
Edge et al., 2021 — Cognitive load in simulation-based learning experiences
Why extraneous load matters — unnecessary fidelity, complexity and distraction can crowd out the very cue being taught.
Society for Simulation in Healthcare — Simulation Dictionary (3rd ed.) and Accreditation Standards
The functional-versus-physical fidelity vocabulary, and the standard that modality and realism should meet the learning objectives — higher fidelity is not required for its own sake.
INACSL Standards Committee, 2021 — Healthcare Simulation Standards of Best Practice: Simulation Design
Purposeful simulation design that starts from identified objectives and optimises achievement of the expected outcomes.
Sweller, van Merriënboer and Paas, 2019 — Cognitive architecture and instructional design: 20 years later
Why guidance beats unassisted discovery for novices — the worked-example and guidance-fading effects: novices learn more from studying a guided solution than from searching for it, and guidance is faded only as expertise grows. The learning-science rationale for NAP being instructor-driven, not self-directed.
INACSL Standards Committee, 2021 — Healthcare Simulation Standards of Best Practice: Facilitation
Facilitation as a distinct, standards-recognised practice: the facilitator "provides the structure and process to guide participants… to achieve desired outcomes" and delivers cues toward the objectives. Grounds NAP's instructor-driven model — NAP is the instrument a facilitator uses, not a facilitator itself.
DiGregorio et al., 2025 — The impact of simulation facilitation on learning outcomes: a systematic review
Recent synthesis that trained-facilitator-led simulation is associated with better learner outcomes — facilitation earns its place empirically, not just by standard. Cited with the review's own caveat that the small, heterogeneous evidence base limits pooling and generalisation.
How NAP uses evidence
Evidence review, implementation verification and external face-validation are separate gates. Matching a cited range or passing a code test does not demonstrate clinical realism. NAP uses evidence in four bounded ways.
- 01
Concept selection
Guidelines identify the rhythms, states, device interactions and observable relationships worth teaching.
- 02
Directional grounding
Primary literature informs whether simulated variables should rise, fall, equalise, vary with respiration or change morphology together.
- 03
Starting references
Published bands and conventions provide candidate defaults and soft plausibility fingerprints — a green test is not clinical validation.
- 04
Explicit limits
Sources also show what a monitor-only simulator cannot represent: neurological findings, laboratory criteria, clinical context or proprietary algorithms.
The evidence, by module
Grouped as NAP's modules are built. Each group opens to its full source list; every link is the real, verified reference behind that design area.
Base monitor
Vital signs, rhythms, respiration, arrest and perfusion — the recognisable direction and relationships used in teaching scenarios, not universal “typical patient” values.
Validation status
Guideline-grounded · descriptiveWhat this covers
Blood pressure
When the two blood pressures are moved together, they follow a curve rather than a fixed ratio: as the systolic rises the gap between the two widens, the way stiffer arteries behave with age, and the curve differs for a child. An instructor can always unlink them and set any pair — necessary, not merely permitted, because two patients with the same systolic can genuinely need very different diastolics.
Pulse rate, and the HR–PR difference
The monitor also shows a photoplethysmograph-derived pulse rate beside SpO₂, counted from the simulated pleth pulses NAP detects — so in rhythms such as rapid atrial fibrillation the pulse rate can read below the ECG heart rate, and in pulseless electrical activity it blanks entirely, producing an HR–PR difference resembling a pulse deficit. NAP renders this difference; it does not diagnose or label it, and makes no claim to a validated pulse-deficit magnitude.
Whether a beat makes a pulse
Whether a beat produces a pulse is judged relative to the heart's own expected cycle rather than against a fixed interval: a beat is weak because it arrived early, not because the patient is fast. So a sustained fast rhythm keeps a smaller but real pulse — in keeping with reports of adults paced at 175/min sustaining an arterial trace, and of infants remaining asymptomatic in supraventricular tachycardia at rates near 300 — while a markedly premature beat, as in bigeminy or the short cycles of atrial fibrillation, may eject too little to register.
One circulation judgement feeds every signal
Whether a beat perfuses is decided once, per beat, and every pulse-derived signal reads that same judgement — the oxygen saturation, the plethysmograph, the pulse rate, the blood pressure, and whether the patient breathes spontaneously at all. That is why the picture moves together rather than a value at a time: a pacer that gains capture restores every one of those numbers at once, and a rhythm that loses circulation blanks them together, as at the bedside — with the deliberate exception that an agonal breathing pattern still gasps, because agonal respiration is brainstem-driven and is one of the things an arrest is recognised by. NAP renders that coherence; it does not evolve the patient, and the instructor still sets what the patient is doing.
During chest compressions
During chest compressions the pulse tone is silenced and the pulse rate blanks: a pulse oximeter cannot separate a returning spontaneous pulse from the compression signal, and resuscitation guidance recognises return of circulation by other means — a rise in end-tidal CO₂, an arterial trace, or a pulse check at the rhythm check — not by the oximeter.
What the thresholds rest on
The shapes of these relationships are drawn from published work; the specific thresholds are documented teaching values, not validated measurements.
Grounded in 20 sources
US National Center for Health Statistics, 2022 — NHAMCS emergency-department visit estimates
Prevalence rationale for including common chest-pain, dyspnoea and stroke presentations in the scenario coverage map.
Franklin et al., 1997 — Hemodynamic patterns of age-related changes in blood pressure (Framingham Heart Study)
That systolic and diastolic do not move in proportion: with age the diastolic falls while the gap between the two widens. This is the shape behind linking the two pressures along a curve.
SHEP Cooperative Research Group, 1991 — Prevention of stroke in older persons with isolated systolic hypertension
A reference pair at the top of the range (mean entry pressure 170/77), used as a landmark so a high systolic does not drag the diastolic up with it.
Beckett et al. for the HYVET Study Group, 2008 — Treatment of hypertension in patients 80 years of age or older
A second, independent landmark (173/91) in a group not selected on diastolic — so the wide gap belongs to the age group, not to the entry criteria.
American Heart Association, 2025 — Adult advanced life support
Shockable versus non-shockable arrest behaviour, CPR/capnography cues, reversible-cause coverage and ROSC directionality.
Resuscitation Council UK, 2025 — Special circumstances guidelines
Tension-pneumothorax and other reversible-cause recognition context.
World Allergy Organization, 2020 — Anaphylaxis guidance
Circulatory and respiratory direction of the anaphylaxis teaching state.
SCCM / ESICM, 2021 — Surviving Sepsis Campaign guidelines
Low-perfusion / MAP emphasis in the sepsis scenario.
Spahn et al., 2023 — European guideline on major bleeding and coagulopathy after trauma (6th ed.)
Haemorrhagic-shock and major-trauma perfusion direction, and the distinction between a resuscitation target and a presenting value.
NICE, 2016 (updated) — Major trauma: assessment and initial management (NG39)
Urgent recognition and haemorrhage-control context for tension pneumothorax and major trauma.
ADA / EASD / JBDS / AACE / DTS, 2024 — Hyperglycaemic crises in adults: consensus report
DKA case definition; prevented an obsolete glucose threshold being presented as current.
SRLF / SFMU, 2024 — Oxygen therapy in acute hypoxaemic respiratory failure
The hypoxaemic-respiratory-failure archetype, while making clear a monitor without blood gas cannot represent the whole syndrome.
BTS / SIGN, 2019 — British guideline on the management of asthma
Recognisable severe-asthma vital-sign direction and deterioration cues.
Global Initiative for Asthma, 2024 — Global strategy for asthma management
Severe exacerbation and exhaustion context used alongside BTS/SIGN.
World Health Organization, 2025 — Opioid overdose fact sheet
Respiratory depression and loss of consciousness as the central monitor lesson in opioid overdose.
World Health Organization, 2011 — IMAI district clinician manual
Respiratory-arrest escalation and supportive-monitoring context for poisoning scenarios.
AAPCC consensus panel, 2007 — Tricyclic antidepressant poisoning out-of-hospital guideline
Importance of QRS morphology and haemodynamic compromise in the TCA teaching case.
ACC / AHA, 2022 — Diagnosis and management of aortic disease
Aortic-dissection instability context and its acknowledged inability to show bilateral pulse/BP or neurological findings.
AHA / ASA, 2026 — Acute ischaemic stroke guideline
Stroke as a deliberate negative-control scenario: important, but not diagnosable from monitor values.
American Heart Association, 2020 — FAST recognition
The non-monitor recognition boundary documented for the stroke case.
12-lead ECG and rhythm morphology
Educationally recognisable vectors, lead relationships and morphology features.
Validation status
Guideline-grounded · descriptiveWhat this covers
What it does not do
NAP does not calculate or interpret a diagnostic ECG, and its display is not claimed to meet diagnostic acquisition standards.
Setting a target QTc
An instructor can set a target QTcF — QT corrected for heart rate by the Fridericia formula — to teach the long-QT picture; the interval readout then MEASURES that back off the drawn trace rather than repeating the number that was typed in. The dial’s 450 / 480 / 500 targets follow the categorical outlier boundaries the ICH E14 guidance suggests for QTc analysis; they are authoring conveniences, not diagnostic categories.
Two corrections, printed together
The readouts print both QTcF (Fridericia) and QTcB (Bazett), each labelled, with QTcF as the primary value the ≥500 marker is evaluated against. Fridericia is used for marking and authoring because NAP’s rendered QT/RR relation is close to a cube-root law, so QTcF stays nearly rate-flat over the drawn trace, while Bazett — the correction many bedside monitors print by default — visibly over-corrects as the heart rate rises. Showing both on the same beat lets that divergence be demonstrated live: at a healthy infant’s default rate the drawn trace reads QTcB above 480 while QTcF stays in the low 420s. The E14 Q&A (R3) describes Bazett as an inferior correction in adults and Fridericia as likely to be appropriate in most situations; GE’s 12SL analysis program computes Bazett, Fridericia and Framingham, with Bazett the default reported value. None of this is a claim of diagnostic or screening validity — the two numbers are measurements of a simulated trace, labelled by method.
Which lead the intervals come from
The monitor measures the lead it is sweeping, and the 12-lead measures across all twelve — from the earliest onset in any lead to the latest offset in any lead, which is the standard for 12-lead measurement. Each readout says which it is. The two can legitimately differ for the same patient: a single lead reads shorter when its own onset or offset happens to sit on the baseline, so a bundle-branch block that measures about 177 ms on the monitor spans about 190 ms across twelve leads. For the same reason the monitor may decline to report a QT it genuinely cannot see, where the 12-lead can.
The ≥500 ms marker
At or above a QTcF of 500 ms the value carries a ≥500 marker — evaluated against the Fridericia value only, never the Bazett secondary. The 500 ms line follows the highest of the ICH E14 categorical outlier boundaries (450 / 480 / 500 ms), applied identically across simulated age groups: no guideline body publishes an age-banded QTc threshold above the neonatal period, and NAP does not invent one. The marker is a description of a measurement, neither an age-specific reference range nor a prediction that the rhythm will deteriorate — selecting torsades remains the instructor’s deliberate act.
The unauthored baseline
⚠ The unauthored baseline is a simulation output built from each age band’s resting QT rather than a fixed corrected target. The displayed QTcB therefore varies with heart rate (rising as the rate rises), while the displayed QTcF stays nearly flat; both baseline values are simulation outputs, not clinical normal ranges.
Grounded in 16 sources
Kligfield et al. for AHA/ACCF/HRS, 2007 — ECG standardization, Part I
Standard lead nomenclature, acquisition concepts, the waveform-generation-vs-interpretation distinction, and global measurement across simultaneously recorded leads as the basis for 12-lead interval durations.
ICH E14, 2005 (+ Q&A R3, 2015) — Clinical evaluation of QT/QTc prolongation
The categorical outlier boundaries (QTc >450 / >480 / >500 ms, §3.2.2) behind the QT dial’s targets and the ≥500 marker, and the Q&A R3 (Q1.5) assessment that Bazett is an inferior correction in adults with Fridericia likely to be appropriate in most situations — grounding for which correction NAP marks against, not a claim that NAP performs E14 analyses.
GE Healthcare, 2019 — Marquette 12SL ECG Analysis Program Physician’s Guide (2056246-002C)
Real-device convention for printing a method-labelled QTc: 12SL computes Bazett, Fridericia and Framingham corrections, with Bazett the default reported value unless denoted — the precedent for NAP printing both corrections labelled by method.
Surawicz et al. for AHA/ACCF/HRS, 2009 — Part III: intraventricular conduction
Bundle-branch-block morphology and conduction terminology.
Rautaharju et al. for AHA/ACCF/HRS, 2009 — Part IV: ST, T, U, QT
ST/T morphology conventions for ischaemia, pericarditis and electrolyte overlays.
Wagner et al. for AHA/ACCF/HRS, 2009 — Part VI: acute ischaemia / infarction
Contiguous-lead and reciprocal-change relationships for the STEMI teaching overlays.
Thygesen et al., 2018 — Fourth Universal Definition of Myocardial Infarction
Localisation / contiguous-lead context, and the warning that NAP’s ST amplitude is a teaching default, not a universal diagnostic threshold.
ACC / AHA, 2025 — Management of acute coronary syndromes
ACS / STEMI scenario scope and contemporary clinical context.
Pava et al., 2010 — R-wave peak time at DII
The delayed lead-II R-wave peak that makes simulated VT recognisable on a single-lead monitor.
Vereckei et al., 2008 — aVR wide-QRS tachycardia algorithm
Initial-versus-terminal activation features used in NAP’s wide-complex morphology.
Brugada et al., 1991 — Regular wide-QRS tachycardia
The prolonged RS-interval feature represented in the VT teaching morphology.
Sokolow and Lyon, 1949 — Ventricular complex in LVH
The voltage-pattern teaching overlay for LVH; no diagnostic sensitivity/specificity is asserted.
Miner, Grigg and Hart, 2023 — Wellens syndrome
The characteristic precordial T-wave teaching pattern and its clinical-context caveat.
KDIGO Controversies Conference, 2020 — Potassium management in kidney disease
The association between hyperkalaemia severity and ECG change, without treating one overlay as the full progression.
Lindner et al. for KDIGO, 2020 — Acute hyperkalaemia in the emergency department
Emergency-recognition context and the documented gap between NAP’s peaked-T overlay and a full graded QRS/sine-wave sequence.
European Society of Cardiology, 2025 — Myocarditis and pericarditis
Diffuse ST/T and pericardial-disease teaching context, including tamponade cross-reference.
Defibrillation, cardioversion, pacing and NIBP
Device actions are simplified for instruction.
Validation status
Guideline-grounded · descriptiveWhat this covers
What the guidelines do and do not establish
Guideline pathways inform state transitions and recognisable indications — they do not make NAP a treatment recommender or establish real-device timing, energy, capture or blood-pressure accuracy.
Mean arterial pressure
The mean arterial pressure appears beside the non-invasive systolic/diastolic just as a real monitor shows it, computed with the standard bedside formula (MAP = diastolic + one-third of the pulse pressure); the cuff readings themselves are instructor-controlled, not a validated oscillometric measurement.
Synchronised cardioversion
A synchronised cardioversion is recognised as its own action rather than a defibrillation, and is judged against its own lower adequacy threshold.
Adenosine
Adenosine appears as an ask-instructor prompt, not a modelled drug: the instructor chooses and records the outcome — a conversion, or a transient block that unmasks an underlying atrial rhythm without stopping it — because in NAP the instructor is the physiology and the simulator never evolves the patient on its own.
AED mode
AED mode is a representative AED training model, not a copy of any device: the learner attaches pads, presses analyse, and — when advised — delivers a fixed shock through a two-minute CPR cycle with spoken prompts.
It does not analyse the ECG
Crucially, NAP does not analyse an ECG signal; the advice is derived from the rhythm the instructor has set, modelling what a representative AED would decide — so it cannot mis-detect, and it makes no claim to reproduce any manufacturer’s detection algorithm, timings or brand. It advises a shock for ventricular fibrillation and rapid ventricular tachycardia and withholds one otherwise; the rate above which it treats monomorphic ventricular tachycardia as shockable is a representative teaching threshold, not a validated cutoff — the 1997 American Heart Association task force deliberately set no universal figure and commercial devices genuinely disagree at that boundary.
Adult only
AED mode is adult-only and carries no paediatric-AED or PALS claim.
AED prompts during CPR
It can also interact with CPR and coach compression quality — but here too it only reflects what the instructor has set, never anything it measures: when the instructor is running compressions the AED waits and says “stop compressions” before analysing, and it voices “push harder” or “good compressions” according to the Good/Poor quality the instructor authored, with a metronome at the authored compression rate whenever compressions are running. This is not real-time CPR feedback and nothing is sensed.
Two honest boundaries
Two honest boundaries: refusing to analyse during compressions models a conventional AED (modern devices analyse through compressions), and the quality coaching models a CPR-feedback-capable device rather than every AED — both are representative teaching behaviours, adjustable per room (auto-analyse and coaching are instructor options), and neither names a manufacturer.
Grounded in 16 sources
Kerber et al. for the American Heart Association Task Force on Automatic External Defibrillation, 1997 — AED arrhythmia-analysis algorithm performance
That the AHA sets high sensitivity for shockable and high specificity for non-shockable rhythms as the AED targets, and deliberately specifies no universal ventricular-tachycardia rate cutoff — grounding NAP’s AED as a representative model (not a validated detection algorithm) and its VT rate gate as a teaching threshold.
Nishiyama et al., 2015 — Diagnostic accuracy of commercially available automated external defibrillators
That commercial AEDs disagree markedly at the VT boundary (shock-advice for VT above 180/min ranged 36–96% across four devices) — the basis for NAP presenting its rate gate as representative, not a validated cutoff matching any one device.
de Graaf et al., 2021 — Analyzing the heart rhythm during chest compressions (cprINSIGHT AED algorithm)
That modern AEDs can analyse the rhythm through ongoing compressions — the honest basis for NAP stating that its “wait, stop compressions to analyse” behaviour models a conventional AED, a teaching choice rather than a universal truth.
Christenson et al., 2009 — Chest compression fraction determines survival in out-of-hospital VF; Cheskes et al., 2011 — Perishock pause and survival
That minimising hands-off time (higher compression fraction, shorter peri-shock pause) is tied to survival — the teaching point behind modelling the AED’s hands-off cadence at all.
American Heart Association, 2025 — Electrical cardioversion algorithm
Synchronised-versus-unsynchronised teaching behaviour and the unstable-tachyarrhythmia transition model.
American Heart Association, 2020 — Adult tachycardia with a pulse
Instability / cardioversion logic for SVT, AF with rapid ventricular response and VT with a pulse.
American Heart Association, 2025 — ACLS tachycardia (with a pulse) algorithm
The per-rhythm synchronised-cardioversion energy ranges behind treating cardioversion as a separate, lower adequacy threshold than defibrillation, and the shift toward device-recommended energies.
Soar et al. for the European Resuscitation Council, 2021 — Adult advanced life support guideline
European synchronised-cardioversion energies, cross-checked against the American ranges so the threshold is not drawn from a single guideline tradition.
Glover et al., 2008 — Biphasic energy selection for transthoracic cardioversion of atrial fibrillation (BEST AF)
That first-shock success is materially higher at higher biphasic energies — the basis for NAP stating openly that a single low adequacy floor is charitable to atrial fibrillation rather than pretending one threshold fits every rhythm.
Panchal et al. for the American Heart Association, 2020 — Adult basic and advanced life support
Adenosine’s place in regular narrow-complex tachycardia and the teaching caveat against using it in irregular wide-complex tachycardia. In NAP the drug is a prompt the instructor records and authors, never a modelled effect, and no dose is surfaced in the interface. The prompt now spans paediatric rooms too — SVT is the classic paediatric tachyarrhythmia — with the authored outcomes and their rates age-scaled to the class (and the flutter-reveal outcome shown only in the older bands, never in the infant, where it would be a false note). This stays a descriptive, instructor-authored teaching depiction, not a PALS or validated-paediatric claim, and remains under NAP’s open external paediatric review.
Page et al. for the ACC/AHA/HRS, 2015 — Guideline for the management of adult patients with supraventricular tachycardia
The differential the authored outcomes depict: adenosine either terminates an AV-node-dependent re-entry, or by transient block transiently slows a non-nodal atrial mechanism — unmasking it without terminating it.
DiMarco et al., 1983 — Adenosine: electrophysiologic effects and therapeutic use for terminating paroxysmal supraventricular tachycardia
That the transient block preserves atrial activity rather than producing a flat line — why NAP draws the reveal as an ongoing atrial rhythm at a slowed ventricular response, and why the effect is brief.
American Heart Association, 2025 — Adult bradycardia with a pulse
The compromised-bradycardia and transcutaneous-pacing teaching path.
American Heart Association, 2020 — Post-cardiac-arrest care
The direction of post-ROSC perfusion and capnography changes (not one universal post-ROSC vital set).
ACC / AHA / ACCP / HRS, 2023 — Diagnosis and management of atrial fibrillation
Irregular AF rhythm behaviour and urgent-cardioversion context when instability is represented.
Muntner et al. for the American Heart Association, 2019 — Measurement of blood pressure in humans
Cuff-selection, positioning and oscillometric conventions that frame NIBP simulation and paediatric cuff checks — it does not validate NAP’s generated readings.
Alarms
NAP’s alarm behaviour is a teaching abstraction whose priorities follow the priority framework of IEC 60601-1-8 — potential harm crossed with onset, so a condition that can kill within seconds ranks high regardless of any numeric.
Validation status
Guideline-grounded · descriptiveWhat this covers
Crisis alarms
Crisis alarms therefore behave as a real patient monitor ranks them: life-threatening rhythms — ventricular tachycardia, ventricular fibrillation and asystole — and apnoea raise a high-priority alarm by recognising the condition, not only when a number crosses a limit, and a lethal alarm can’t be switched off with a per-vital toggle.
Apnoea waits
When a patient stops breathing but is still circulating, the apnoea alarm waits a short onset window (20 s) before it annunciates, because a monitor does not call apnoea on a single missed breath; the respiratory-rate number still drops to zero immediately, and an arrest — which has no breathing to wait for — alarms at once.
Other alarms annunciate at once
Apart from apnoea, an alarm sounds the moment a value crosses its limit — a real monitor usually waits a beat first, to avoid firing on a noisy sample or a momentary blip. NAP has no such transients to filter: the instructor is the physiology, so a value only changes because someone changed it, and the monitor never moves a number on its own. A pause between your action and the alarm would read as the simulator being slow rather than as anything about alarms, so the delay is left out deliberately.
Simultaneous alarms take turns
When more than one alarm of the same priority is active, the banner cycles through them rather than showing only the first — so a desaturation cannot sit hidden behind a heart-rate alarm on a screen being watched from across the room.
The alarm sound is NAP’s own
The alarm tones are a house sound, modelled on the pitches a real monitor documents for its standard-conformant setting, in NAP’s own shorter cadence. That is the position real devices occupy too — the major bedside monitors ship their own alarm sounds by default and offer the standard set as an option — so NAP does not claim to reproduce any particular device’s alarm, or the standard’s full burst.
What NAP is not
NAP is not an IEC-conformant device and does not replicate real arrhythmia-detection performance: it does not substantiate exact factory limits, delays, sounds, colours or offsets.
Grounded in 4 sources
International Electrotechnical Commission, 2020 — IEC 60601-1-8:2006+A1:2012+A2:2020
The distinction among alarm-condition priorities and general alarm-signal conventions — priority as potential harm crossed with onset, so a condition that can kill within seconds ranks high regardless of any numeric.
Patient-monitor manufacturer — instructions for use and configuration guide
The red / yellow / technical three-channel split, the red arrhythmia list, apnoea as a high-priority alarm after a configurable onset delay (default 20 s), and the guard that keeps lethal alarms from being switched off. Vendor interface concepts, not device equivalence or NAP’s exact numbers.
Patient-monitor manufacturer — arrhythmia-algorithm clinical resource
That a “lethal” arrhythmia class (asystole, VF/VT) is a conventional non-optional floor — the concept behind lethal codes bypassing the per-vital off toggle. Interface concept, not algorithm equivalence.
Suba et al., 2024 — Performance comparison of 6 in-hospital patient monitoring systems in the detection and alarm of ventricular cardiac arrhythmias
That treating asystole/VF/VT as the high-severity class is universal across vendors rather than a single-vendor quirk. Establishes convention, not NAP validation.
Respiration & respiratory rate
How NAP decides a respiratory rate is measurable and which source it trusts.
Validation status
Guideline-grounded · descriptiveWhat this covers
Setting a pattern sets a rate
Choosing a breathing pattern also sets a starting rate for it, which the instructor can still adjust, and a gentle note appears if the rate chosen reads oddly for that pattern — the rate set is the rate the monitor shows, with no hidden multiplier.
The patterns it renders
The patterns it renders — normal, Cheyne–Stokes, Kussmaul, Biot/ataxic, agonal and apnoea — are recognisable rate-and-depth teaching cues whose defining cadence is grounded in standard clinical descriptions; NAP does not reproduce any individual patient’s respiratory mechanics, and each pattern carries a physiologically-plausible starting rate — an adjustable clinical-judgment teaching value grounded in the pattern definitions, not a measured or validated reference standard.
Where the rate comes from
On a real monitor RR is derived from impedance pneumography (ECG electrodes), the pleth (SpO₂ probe), or capnography; NAP names the best available source (capnography > pleth > impedance) and blanks the rate when none is attached.
The rate is counted over a window
Like a monitor counting breaths, NAP reports the respiratory rate averaged over a short rolling window rather than instant by instant. So an irregular pattern — agonal gasping, or the pause in Cheyne–Stokes — reads as a low rate between breaths rather than dropping to zero, while apnoea still reads a true zero. At very slow gasping rates a breath can fall outside the window altogether, and the rate then reads zero until the next one: the monitor reports what it counted, which is also what a real one would do. Steady patterns are not averaged, so an ordinary rate change still shows immediately. The window length is a teaching convention chosen for clarity, not a figure taken from a published standard or copied from a particular device.
What the sources establish
The sources below inform this source model, the pattern definitions and the reliability ordering — not a validation of NAP’s rate values, which remain instructor-set.
Grounded in 5 sources
Porter and Graham for StatPearls, 2025 — Abnormal Respirations
The eupnoeic 12–20/min baseline and the defining features of the abnormal patterns — Kussmaul as deep and fast, Biot/ataxic as clusters separated by irregular apnoeas, agonal as slow gasping — informing the recognisable cadence of each pattern (not a per-patient mechanics claim).
Rudrappa, Modi and Bollu for StatPearls, 2023 — Cheyne–Stokes Respirations
The smooth crescendo–decrescendo of tidal volume separated by central apnoeas and the characteristically long cycle — grounding the waxing/waning Cheyne–Stokes cue and the decision that its apnoeic trough must not false-fire the apnoea alarm.
Charlton et al., 2018 — Breathing-rate estimation from the ECG and photoplethysmogram (IEEE Rev Biomed Eng)
Pleth-derived (photoplethysmographic) respiratory rate as an established, reviewed measurement class — not a niche add-on.
Eisenkraft et al., 2023 — Clinical validation of a wearable respiratory-rate device (Chron Respir Dis)
Modern agreement of pleth-derived RR against capnograph and ventilator references.
Touw et al., 2018 — Accuracy of remote continuous respiratory-rate monitoring (Can J Anaesth)
Capnography as the reference method and impedance pneumography as the widest limits of agreement — the source-reliability ordering.
Advanced haemodynamics
Reference bands, derived relationships, directional preset fingerprints, waveform conventions and respiratory interaction.
Validation status
Beta · external face-validation openWhat this covers
A number only when the line is in
A number appears only when the line that measures it is actually in place — central venous pressure needs a central line; pulmonary pressures, thermodilution cardiac output and mixed-venous saturation need a pulmonary-artery catheter; and the wedge shows only while the balloon is inflated, because it is a brief manoeuvre rather than a standing reading. A saturation taken from a central line is labelled as such and never presented as a mixed-venous one.
Cardiac output, by method
Cardiac output can also be estimated from the arterial line alone — an uncalibrated pulse-wave-analysis estimate, labelled as such and kept distinct from the pulmonary-artery-catheter thermodilution figure; it is presented as a trend rather than a calibrated measurement, and NAP shows one patient’s cardiac output under one method label at a time rather than inventing a difference between the two methods.
Coronary perfusion pressure during CPR
The same rule governs the coronary-perfusion-pressure readout shown during CPR: it is a pressure difference between two sites, so it appears only when an arterial line and a central line are both in place, and it leaves the screen when compressions stop. It marks the level historically associated with return of circulation as a teaching landmark, without asserting anything about a particular patient’s outcome.
A derived number is withheld rather than invented
Every derived index is computed from the values on screen at the moment it is read, and it is shown only while the quantities it depends on make it meaningful. A resistance is a pressure difference divided by a flow, so it is withheld when there is no forward pressure gradient — a central venous pressure at or above the mean arterial pressure, or a wedge pressure at or above the mean pulmonary artery pressure — rather than displayed as a negative number no monitor could produce. The same rule blanks the flow-derived indices when there is no perfusing rhythm. NAP would rather show nothing than something a real device could never show.
Waveform timing comes from the same cardiac clock
The venous trace is drawn from the same beat timeline the ECG is drawn from, rather than from a second rhythm of its own. In rhythms where the atria and ventricles are dissociated, the a-wave is keyed to that independent atrial timeline — the same one that places the P waves on the ECG — so an atrial contraction falling within ventricular systole shows as an accentuated a-wave on that beat: the recognisable cannon a-wave teaching pattern, aligned with the rhythm above it. The pronounced narrowing seen under the tamponade picture — where it falls on inspiration — and under positive-pressure ventilation, where the swing reverses with the breath, is likewise applied as a share of each affected channel’s own pulse pressure rather than as a fixed number of millimetres; the ordinary respiratory variation is a bounded shift of the whole trace, and is not re-modelled as a narrowing. The timing and the proportions are structural; the rendered magnitudes are teaching values, with no claim to validated waveform fidelity.
What this establishes
These establish starting concepts — not clinical validation of NAP’s numbers or waveforms. This module remains governed by an open external face-validation checklist.
Grounded in 24 sources
Marino, 2014 — Marino’s The ICU Book (4th ed.)
General adult haemodynamic reference bands used as soft fingerprints rather than clamps.
Haemodynamic-monitoring manufacturer — normal adult parameters pocket card
Starting adult ranges for invasive pressures, flow, resistance and mixed-venous saturation.
Vanderbilt University Medical Center CVICU, 2020 — Hemodynamic parameters
Calculation definitions and normal-value cross-checks for derived haemodynamic channels.
Haemodynamic-monitoring manufacturer — normal parameters reference
An additional cross-check for the normal adult haemodynamic reference set.
van Diepen et al., 2024 — Mixed cardiogenic-vasodilatory shock
The mixed-shock directional fingerprint linking flow, filling pressure and vascular resistance.
Kapur et al., 2022 — Stages of cardiogenic shock severity
Cardiogenic-shock stage variables that constrain the scenario fingerprint rather than dictate one presentation.
SCAI — Mechanical circulatory support in cardiogenic shock: QI tips
The haemodynamic-variable selection used in the cardiogenic-shock teaching case.
Paradis et al., 1990 — Coronary perfusion pressure and return of spontaneous circulation
The coronary-perfusion-pressure landmark marked on NAP’s CPR readout. A small 1990 cohort under very different resuscitation practice, used as a legible teaching landmark rather than a validated cut-off.
Lim et al., 2020 — Cardiogenic shock: oxygen delivery and utilization
Aetiology-dependent profiles and the relationship between cardiac output, filling pressure and tissue oxygen delivery.
Gopal, Sokup Ivanov and Meer, 2026 — Cardiac tamponade
Diastolic pressure equalisation and tamponade waveform / preset direction.
Van Dam, Hashmi and Fitzgerald, 2024 — Pulsus paradoxus
The clinical definition and respiratory-interaction context for NAP’s pulsus-paradoxus model.
Curtiss et al., 1988 — Pulsus paradoxus and cardiac tamponade severity
The magnitude / direction of respiration-linked systolic pressure variation in tamponade.
Klopfenstein et al., 1985 — Mechanism of pulsus paradoxus in tamponade
The mechanistic relation between respiration, ventricular interaction and arterial pressure variation.
Piazza and Goldhaber, 2006 — Acute pulmonary embolism, Part I
The PA-pressure, wedge-pressure and pulmonary-resistance direction of the massive-PE preset.
European Society of Cardiology, 2019 — Acute pulmonary embolism
The high-risk-PE / shock context and risk-marker selection used in scenario review.
AHA / ACC, 2026 — Evaluation and management of acute pulmonary embolism
The current clinical framing of high-risk PE.
Bootsma et al., 2022 — The contemporary pulmonary artery catheter, Part 1
Invasive-pressure waveform anatomy, zeroing / levelling concepts, and fast-flush and wedge teaching behaviours.
Quaal, 2001 — Accuracy of pulmonary artery catheter measurements
Caution around waveform acquisition and interpretation — it does not validate NAP’s invasive-pressure rendering.
Bootsma et al., 2021 — The contemporary pulmonary artery catheter, Part 2
The measurement set a pulmonary-artery catheter actually provides — pulmonary-artery pressure, occlusion (wedge) pressure, thermodilution cardiac output and mixed-venous saturation — i.e. the numbers that would be false notes without one.
Monnet and Teboul, 2015 — Minimally invasive monitoring
The decline of routine pulmonary-artery-catheter use in favour of less-invasive cardiac-output methods — the context for treating thermodilution cardiac output as catheter-gated, and why a cardiac-output estimate from an arterial line is worth teaching.
Saugel et al., 2021 — Cardiac output estimation using pulse wave analysis
The starting concept behind the arterial-line cardiac-output readout: an uncalibrated pulse-wave-analysis estimate inferred from the arterial pressure waveform, labelled as an estimate and kept distinct from thermodilution — a trend, not a calibrated measurement. It grounds the teaching frame, not any accuracy claim about NAP’s number.
Lequeux et al., 2010 — Continuous mixed venous and central venous oxygen saturation in cardiac surgery
That a central-venous saturation reads on average higher than the true mixed-venous value — the offset behind labelling the central-line-only reading separately.
Varpula et al., 2006 — Mixed venous oxygen saturation cannot be estimated by central venous oxygen saturation in septic shock
A cross-check of that offset’s magnitude and of the non-interchangeability of the two saturations.
Lorentzen et al., 2008 — Central venous oxygen saturation cannot replace mixed venous saturation in cardiac surgery
That the relationship between the two, while offset on average, is too variable to be interchangeable — so the central-line value is never relabelled as mixed-venous.
Processed EEG / depth of anaesthesia
NAP presents a generic 0–100 processed-EEG index plus simulated EEG-derived features, including a density spectral array (DSA) — a frequency-over-time colour strip of the same spectral estimate the edge frequency is read from.
Validation status
Beta · external face-validation openWhat this covers
The density spectral array
The DSA is a generic depth-and-suppression view: it shows how the spectrum shifts with anaesthetic depth and how burst suppression appears, and it deliberately does not model anaesthetic-agent-specific spectrogram signatures.
What it does not reproduce
It does not reproduce any proprietary algorithm, values, performance or device behaviour.
Validation status
This module’s transitions remain subject to a separate external face-validation gate.
Grounded in 12 sources
Depth-of-anaesthesia monitor manufacturer — product information
The display vocabulary — processed index, signal quality, EMG and suppression ratio — used to define the module’s visible channels.
Depth-of-anaesthesia monitor manufacturer — operator-documentation index
The official route to operator documentation; used only for interface concepts, not algorithm equivalence.
Nimmo et al., 2019 — Safe practice of total intravenous anaesthesia
Clinical context for processed-EEG monitoring during TIVA, and the decision not to portray one index as a complete anaesthetic state.
Gao et al., 2018 — Processed EEG and anaesthetic awareness: meta-analysis
The cautious treatment of outcome claims and the separation of a teaching display from clinical efficacy.
Zhang et al., 2011 — Index-guided TIVA and awareness in high-risk patients
The awareness-monitoring context, without being used as validation of NAP’s index.
Evered et al., 2021 — Anaesthetic depth and delirium: BALANCED substudy
The deliberately cautious relationship between processed-EEG targets and postoperative outcomes.
Purdon et al., 2015 — Clinical EEG for anesthesiologists, Part I
Recognisable awake, propofol, general-anaesthesia and burst-suppression EEG morphology.
Akeju et al., 2014 — EEG signatures of propofol and sevoflurane
Agent-linked spectral / morphology distinctions used in the teaching presets.
Sleigh et al., 2021 — Raw and processed EEG monitoring
The relationship — and possible disagreement — between raw EEG, derived values, artefact and clinical state.
Schuller et al., 2015 — Index decline after neuromuscular blockade in awake volunteers
The EMG / neuromuscular-blockade confound represented in the simulator.
Purdon et al., 2015 — Age-related changes in EEG during propofol anaesthesia
Age as a modifier, and the decision not to treat one morphology / index relation as universal.
Shanker et al., 2021 — Burst suppression: causes and mechanisms
The qualitative burst-suppression state and suppression-ratio teaching channel.
Paediatrics
An Adult↔Paediatric patient class with four age bands (Infant, Small child, Bigger child, Adolescent).
Validation status
Beta · clinician face-checkedWhat this covers
Validation standing
A paediatric clinician face-checked the numbers as plausible teaching values — not a validated paediatric-fidelity claim, and not a PALS/APLS/ERC-compliance claim.
What the sources informed
These sources informed the age bands, growth-derived body size, vital ranges, alarm defaults, weight-based defibrillation, BP-cuff measurement, fluid references and haemodynamic BSA/CI scaling; each paediatric number is a teaching value grounded in them (⚠ clinical judgment), not a reference standard.
Age-adjusted ECG morphology
The ECG waveform is now age-adjusted: NAP’s existing one-vector, one-beat ECG model is parameterised by band so the limb-lead axis, the precordial R/S progression and the juvenile T-wave pattern move with age — the infant’s right-sided emphasis giving way to a progressive leftward shift, the chest-lead R/S maturing toward the adult direction, and the juvenile T-wave inversion resolving left-to-right (V3, then V2, then V1) across the bands. Each band draws one representative deterministic trace inside the published normal-limits envelopes below; those envelopes describe a population, not one universal child, so this is a plausible teaching ECG for the age — not a validated, diagnostic or PALS ECG.
Out of scope
Neonatal morphology (birth–28 days) is deliberately out of scope: NAP’s youngest band is an infant at six months. Earlier releases drew the adult ECG pattern at paediatric rates; this replaces that.
Still outstanding
A stronger clinically-reviewed claim still awaits an external paediatric-ECG specialist’s review of the rendered age-band plates.
Grounded in 16 sources
Royal Children’s Hospital Melbourne — Acceptable ranges for physiological variables
The shipped V1 age-banded reference vital ranges and the need for age-specific alarms.
Royal Children’s Hospital Melbourne — Basic paediatric ECG interpretation
An accessible age/rate cross-check for HR, axis, PR/QRS duration, voltages, R/S ratios and juvenile T waves — grounding the age-adjusted ECG morphology across the four bands.
Rijnbeek et al., 2001 — New normal limits for the paediatric ECG
Age- and sex-specific digital normal limits for intervals, axes and voltages — the calibration envelopes for the age-banded ECG, not validation of NAP’s selected targets.
Saarel et al. / Pediatric Heart Network, 2018 — Electrocardiograms in healthy North American children in the digital age
Contemporary age/sex percentile envelopes for HR, P/PR/QRS/QTc, axes and precordial amplitudes, and the caution that sex, race and acquisition method affect normal measurements.
Dickinson, 2005 — The normal ECG in childhood and adolescence
The developmental pattern the age-adjusted ECG evokes — ventricular dominance, R/S progression, the PR–rate relation and the juvenile T-wave sequence.
World Health Organization — Weight-for-age child growth standards
The proposed age-to-weight reference inputs for body-size and weight-based simulation.
World Health Organization — Length/height-for-age child growth standards
The proposed age-to-height reference inputs used with weight for body-surface-area calculation.
World Health Organization — Growth reference data for 5–19 years
Extended the growth-derived body-size basis into school-age and adolescent bands.
Mosteller, 1987 — Simplified calculation of body-surface area
The proposed height-and-weight BSA derivation used for indexed haemodynamic quantities.
Resuscitation Council UK, 2025 — Paediatric life support
V1 age-related physiology, paediatric arrest pathways, weight-based defibrillation and circulation teaching behaviour.
American Heart Association, 2025 — Paediatric cardiac-arrest algorithm
Cross-checked the shipped weight-based defibrillation and arrest-state transitions.
Lewis et al., 2016 — Acute management of refractory and unstable paediatric supraventricular tachycardia (J Pediatr)
That in acute paediatric SVT, adenosine and first-line vagal/medical management terminate most episodes and synchronised cardioversion is reserved for unstable or adenosine-refractory SVT — grounding SVT cardioversion as an instructor-decided post-shock outcome rather than a routine step, and the infant band as the harder case (first-dose adenosine response lower in infants).
Flynn et al. for the AAP, 2017 — High blood pressure in children and adolescents
Paediatric cuff / measurement considerations and the warning that age-only BP limits are incomplete.
NICE, 2015 (updated) — Intravenous fluid therapy in children (NG29)
The proposed weight-based fluid interaction and its age / weight constraints.
Sproul and Simpson, 1964 — Stroke volume in normal children
Historical primary data for the direction of paediatric stroke-volume and cardiac-output scaling; modern external review remains required.
Lewis et al., 2016 — Acute management of refractory and unstable paediatric supraventricular tachycardia
That in children most SVT episodes settle with vagal or medical management, so synchronised cardioversion is reserved for the unstable or refractory case and is rarely performed — which is why NAP treats it as an outcome the instructor decides rather than a routine step, and why the infant band is the hardest case.
Blood gases
An instructor-authored sample whose acid–base quantities are derived from published equations rather than typed in.
Validation status
Beta · external face-validation openWhat this covers
What is authored, and what NAP works out
A gas starts from one of a catalogue of clinical pictures, which carries the whole measured set — tensions, bicarbonate, electrolytes, lactate, glucose and haemoglobin. From there the instructor moves the acid–base axis directly — pH, carbon-dioxide tension, bicarbonate and base excess — and can also set the potassium and the glucose. The remaining values (sodium, chloride, lactate, haemoglobin, oxygen tension) come from the chosen picture and are not separately adjustable today. What is never authored at all is the pH, the base excess, the anion gap and the calculated oxygen saturation: those are recomputed from the measured set every time the result is drawn, so a NAP gas cannot be internally inconsistent in the way a hand-typed one can. Bicarbonate and base excess share a single control under two labels, which means the pair cannot be moved into disagreement with each other.
The equations
The pH comes from Henderson–Hasselbalch in its clinical form; the standard base excess from the Van Slyke equation; the anion gap from the convention that excludes potassium; and the calculated saturation from Severinghaus’ standard oxyhaemoglobin dissociation curve. Each is cited below. The base-excess, saturation and anion-gap expressions were each checked against the published source rather than against a secondary summary; the pH equation is Henderson–Hasselbalch in the standard clinical form given in the acid–base review cited below.
Standard base excess, and why haemoglobin is not an input
The base excess NAP shows is the STANDARD one: the Van Slyke equation evaluated at the haemoglobin of the extracellular fluid rather than of blood. That is what makes the number describe the extracellular space and stay steady when the haematocrit moves, and it is why the haemoglobin the instructor authors deliberately does not feed it. The other quantity — actual base excess, computed at the patient’s own haemoglobin — is a different measure, and NAP reports the standard one.
Where the model stops
NAP applies no temperature correction and no Bohr pH shift, using the standard curve at pH 7.40 and 37 °C uncorrected — a stated omission rather than an approximation, because a correction with a remembered coefficient would be worse than none. There is no co-oximetry, so no methaemoglobin or carboxyhaemoglobin, and the saturation shown is calculated from the oxygen tension rather than measured. There is no alveolar–arterial gradient and no inspired-oxygen context.
Arterial and venous
A gas can be taken arterially or venously. The venous sample is the same authored patient read differently: pH and carbon-dioxide tension carry the published venous offsets, and the oxygen tension is a typical peripheral value rather than a conversion — the meta-analysis below finds arterial and venous carbon-dioxide tensions are NOT comparable, with a bias interval that crosses zero, so a real venous value can sit either side of its arterial pair. NAP therefore presents a venous sample as its own result and never as something convertible into an arterial one; the card says so on its face, and shows no calculated saturation at all, because none of it is derived from the arterial pair. Because both readings describe one patient, the optional highlight marks the same abnormalities either way — switching how the blood was drawn does not change what is wrong with the person.
A result slip, not a second patient
A released gas is frozen at the moment it is taken, like a printed slip: it does not evolve, and it is not coupled to the monitor in either direction. Editing the gas never moves the patient’s displayed saturation, and the patient never rewrites the gas. Where the two disagree, NAP preserves the disagreement rather than quietly reconciling it — the same rule that governs every other NAP module, that the instructor is the physiology. Reading that disagreement is the instructor’s job; NAP does not currently annotate it.
Reference bands, and the highlight
An optional teaching highlight marks each value that falls outside its reference band. With one exception these are ordinary adult teaching intervals rather than sourced values, and they are listed below as still-open provenance; the anion-gap band is the one carrying a citation. Real intervals are analyser-dependent and vary between laboratories, so no single set would be right everywhere. The comparison is made against the value as PRINTED, in whatever units the room is set to — not the value as stored — so the card cannot show a number sitting inside its band with an “outside” arrow beside it, or outside it with none.
Units describe the reader
Tensions can be read in kilopascals or millimetres of mercury, glucose in millimoles per litre or milligrams per decilitre, and haemoglobin in grams per litre or per decilitre. These are properties of the room reading the result, not of the sample, so they are settings a room carries and can change with a result already on screen — not values frozen into the gas when it was taken.
Grounded in 6 sources
Siggaard-Andersen, 1977 — The van Slyke equation (Scand J Clin Lab Invest Suppl 146:15–20)
The standard-base-excess form NAP ships. The paper states the equation with haemoglobin as a free variable; solved for base excess at the extracellular-fluid haemoglobin of 3.1 mmol/L it reproduces both of NAP’s constants exactly, which is what makes the quantity STANDARD rather than actual base excess.
Morgan, Clark & Endre, 2000 — Accuracy of base excess: an in vitro evaluation of the Van Slyke equation (Crit Care Med 28:2932–6)
That base excess computed by the Van Slyke equation tracks metabolic acid–base status in vitro and is, in the authors’ words, “little affected by large simultaneous alterations in P(CO2), or by very low hemoglobin concentrations similar to that used to calculate standard base excess” — the direct support for evaluating it at a fixed extracellular haemoglobin.
Severinghaus, 1979 — Simple, accurate equations for human blood O₂ dissociation computations (J Appl Physiol 46:599–602)
The oxygen saturation NAP calculates from the oxygen tension. NAP ships the published expression unmodified; the paper reports it fitting the standard human dissociation curve to within ±0.0055 fractional saturation, and it is that STANDARD curve — uncorrected for temperature and pH — which NAP uses.
Kraut & Madias, 2007 — Serum anion gap: its uses and limitations in clinical medicine (Clin J Am Soc Nephrol 2:162–74)
The potassium-excluded anion gap NAP derives, defined there as chloride and bicarbonate subtracted from sodium; and the caution NAP repeats rather than papers over, that the normal value “can vary widely, reflecting both differences in the methods that are used to measure its constituents and substantial interindividual variability”.
Berend, de Vries & Gans, 2014 — Physiological approach to assessment of acid–base disturbances (N Engl J Med 371:1434–45)
Henderson–Hasselbalch in its clinical form and the expected-compensation relationships used to BUILD the preset pictures, so each one hangs together as a recognisable clinical story rather than a set of independently plausible numbers.
Byrne et al., 2014 — Peripheral venous and arterial blood gas analysis in adults: are they comparable? (Respirology 19:168–75)
How a venous sample is presented. The meta-analysis puts arterial pH about 0.03 above peripheral venous, but concludes venous and arterial carbon-dioxide tensions are NOT comparable — a bias whose prediction interval crosses zero. NAP uses the offsets to GENERATE a plausible venous sample from one authored patient, never as a conversion, and says so on the card: the unreliability is the teaching point.
Where provenance is still open
Some exact NAP-authored values are not yet strongly enough sourced to support a fidelity claim, and are marked as citation-to-verify in the canonical bibliography: exact adult scenario / recovery values and timing arcs, exact alarm thresholds and offsets, exact processed-EEG transition constants, the exact paediatric defaults, which remain authored teaching values rather than citable reference standards, and — for blood gases — both the individual preset values and all but one of the reference bands the optional highlight judges against, which are ordinary adult teaching intervals rather than sourced ones. Deliberately compressed timings — such as a shortened NIBP cycle — are product conventions, labelled as such rather than retrofitted with a clinical citation.
The clearest example is how fast EtCO₂ follows a change in perfusion. NAP moves it several times faster than the published time course, so that a class watching a resuscitation sees the consequence of better or worse compressions while it is still the thing being discussed. That is a teaching decision about tempo, taken deliberately and recorded as one. NAP does not claim that EtCO₂ responds this quickly in patients, and no resuscitation guideline specifies a response time — neither AHA 2020 Part 3 nor ERC 2021 ALS contains one. The direction and the mechanism are grounded; the speed is chosen for the lesson.
Evidence informs; face-validation confirms.
A module may only make a fidelity claim after independent clinicians review its rendered behaviour, transitions and teaching interpretation against a documented checklist. Until then, everything here is provenance for the design — not proof of clinical realism. This page renders NAP's single canonical bibliography, kept current as modules are designed and links are re-checked.