AMR Kansas City Field Guide 2026

Medical Director: Dr. Ameet Deshmukh, MD Authorized under: Kansas Board of Emergency Medical Services Effective: 2026

Quiz available at https://amr-kc-protocols.github.io/amr-kc-protocols/quiz.html

Installing on Your Phone

1. Open your browser and go to: https://amr-kc-protocols.github.io/amr-kc-protocols
2. On iPhone: Tap the Share button → "Add to Home Screen"
3. On Android: Tap the menu → "Add to Home Screen" or "Install App"
4. The app will install and work fully offline

Contents

• 28 IFT Clinical Protocols (Section II)
• 34 Section VIII 911 Mutual Aid Protocols
• 22-drug Formulary with full dosing by route and age
• 11 Procedure Cards
• Reference: Vital targets, Pediatric dosing calculator, LTV 1200 alarms, GCS, Cincinnati Stroke Scale, Broselow, Rule of Nines, APGAR

Simulators

Two interactive trainers, both single self-contained pages, both reachable from the home screen.

• vent-ltv1200.html — the LTV 1200 transport ventilator.
• lifepak-15.html — the LIFEPAK 15 as the AHA's Rhythm Recognition and Electrical Therapy station: a rhythm on the monitor, name it from four, then deliver the right therapy on the unit's own keys. Sixteen rhythms covering the ACLS set, four answers — defibrillate, cardiovert, pace, or recognise that none of them is the answer — and a reason line on every one.

Why it is a station and not a console

It began as the whole unit with a walkthrough, seven coached cases and three multi-phase assessments, and it was too much for the screen it is worked on. The drill is the thing that teaches, and everything around it was in the way.

So the page is one loop now. What decides the answer is the rhythm and the patient: monomorphic VT is in the bank three times — pulseless, unstable with a pulse, and stable with a pulse — and takes a different answer each time. A station that mapped rhythm to therapy one-to-one would be teaching a lookup table rather than the decision an ACLS provider makes.

The two shock answers are told apart the way the real unit tells them apart: whether SYNC was armed when the shock landed. That is the thing crews get wrong, so it is asked of the device rather than accepted as a word.

none is never "do nothing" — it is "nothing electrical", and the reason line says what the patient does need instead. A learner who reads "none" and stops has taken the wrong lesson from a right answer.

The screen it is worked on

A phone held sideways, and nothing scrolls. A drill whose answer is below the fold is a drill nobody finishes.

That decided the layout. The unit is laid out fluidly rather than drawn at 1083×704 and scaled — the scaled chassis was exact and could not be made to fit, putting every control at 21–26pt against a 44pt floor on a phone. And it is down to the nine keys electrical therapy is delivered with: ON, SYNC, ENERGY SELECT, CHARGE, SHOCK, PACER, RATE, CURRENT, PAUSE.

The record keys, the monitoring and display areas, LEAD, SIZE and the speed dial are gone. ANALYZE is gone for a different reason than the rest: it is a working control that would answer the question being asked. A learner deciding whether a rhythm is shockable could press it and be told.

Held upright the page says to turn the phone rather than stacking into a scroll, and the unit is inert behind that notice — covering is only paint, and without it the whole chassis stays in the tab order, SHOCK included.

It is also capped at the size of the thing it is a picture of. A LIFEPAK 15 is 15.8 × 12.5 inches, so past that the chassis stops growing and centres instead: left to fill the screen it rendered at roughly twice life size on a 2160-wide Toughbook, which reads as a mockup rather than a unit. CSS inches are nominal — 96px to the inch, before the panel's own DPI and the OS scaling factor — so this is about life size, not calibrated to it. Phones are unaffected, being well under 15.8in wide. The keys are still larger in proportion than the real unit's, because the station shows nine of them where the device has far more; that is the cost of dropping the keys the drill does not use.

The unit is sized against the visual viewport, not the layout viewport. Sideways in mobile Safari with the toolbars showing, the layout viewport stays full height while the visible area is 80–100 px shorter, so a chassis pinned to bottom:0 puts CHARGE and SHOCK behind the browser chrome — the two keys the drill ends on. It tracks visualViewport instead and gives up density in tiers as the visible height falls, down to 285 px.

How the tracings are drawn

The rhythm on the screen is the question, so it has to carry the finding the item is asking about. Most of them are built rather than replayed.

The original simulator drew each rhythm as one keyframed beat stretched to fill the R-R interval. That is fine at 75/min and wrong everywhere else: a QRS is about 0.09 s wide whatever the rate, so stretching made it broad at 45/min and a spike at 200 — the exact feature that separates a wide-complex tachycardia from a narrow one, scaled away by the rate. It also cannot show a rhythm whose parts run on different clocks.

So the station's rhythms are summed from parts measured in seconds — a P wave, a normal QRS-with-T, a ventricular complex — laid on the trace where the rhythm puts them, at durations that do not change with the heart rate. What changes with the rate is the gap between complexes.

Measuring them turned up a second problem: the normal complex gave the QRS nearly half its length, which put a normal beat at about 140 ms — already wide by the under-120 rule the learner is being taught to apply — and left a fast VT barely broader than it. The QRS now takes the first quarter and the ST-T the rest, so a narrow complex measures about 75 ms against 165–245 ms for the ventricular ones, and the comparison the wide-complex items turn on is actually there to be made. That gets each rhythm its diagnosis rather than its outline:

• Atrial flutter — the flutter waves are their own generator at 300/min and run straight through the QRS, so there is no isoelectric baseline anywhere. The conduction ratio falls out of the ventricular rate: 150 is 2:1, 100 is 3:1, without either being written down.
• Atrial fibrillation — no two R-R intervals alike and no pattern to them, from a closed-form sum of incommensurate terms, with the same QRS in every one of them and fibrillatory waves in place of a flat line.
• Mobitz II — the P waves march at a fixed P-P whether or not the ventricle answers, and the PR on the conducted beats never varies. That constant PR is what separates it from Wenckebach.
• Third-degree block — the P waves are on their own clock at 76/min and the wide escape is on the patient's, fully dissociated. A single repeating keyframe cannot show that, and the dissociation is the diagnosis.
• Ventricular tachycardia, the paced rhythm and the third-degree escape share one wide complex with a T pointing away from the R, because they are the same complex arriving for three different reasons. The paced rhythm used to be a normal sinus beat with a spike in front of it — and it is a distractor on the wide-complex tachycardia items, so it has to be wrong for the right reason.
• Junctional escape — narrow complexes at 40–60 with no P in front of them, which is the whole read and what tells it from the third-degree block sitting next to it in the options.
• Ventricular fibrillation is not a template at all: a few incommensurate sinusoids under a slow envelope, so it never repeats cycle to cycle.

The lead transform is applied to the parts rather than to fixed fractions of the cardiac cycle. Applied by phase it cut notches into whatever happened to be at that fraction of the R-R — at 32/min, a visible step in the baseline a third of the way between escape beats, nowhere near a complex.

The boundary

The unit is ported from the CES simulator, where it is half of a two-window system: a facilitator drives the patient from a control panel and the monitor never writes a field of patient state. There is no facilitator here, so the station plays one — in a facilitator's place, rather than by letting the keys reach into the patient. The crew own the device; the patient answers to the station, on the event a press generated. test/lifepak.test.mjs drives the whole therapy set through the unit with the station's writer stubbed out and asserts that not one field of the patient moved.

What the clinical content is checked against

Two sources, answering different questions. The device's behaviour and its numbers come from Stryker/Physio-Control's documentation for the LIFEPAK 15 — the data sheet and the Setup Options guide. The clinical calls come from the 2025 AHA Guidelines for CPR and ECC, Part 9. The suite asserts the published figures directly, so drift fails rather than sits there.

Matching the device: the manual-mode energy ladder, the 200-300-360 J adult sequence that is its factory default, pacing at 40–170 PPM from a 60 PPM default and 0–200 mA, the 60-second disarm.

Matching the guidelines: atropine before pacing; pacing an asystolic arrest taught as ineffective, which is what the 2025 evidence review concluded; and SYNC ruled out for VF and pulseless VT.

On doses, the station teaches the standing orders these learners are actually held to, not a reading of the guidelines. protocols-2026.pdf in this repo names them in four places — V-15 (Synchronized Cardioversion), the LIFEPAK 15 device operation page, the II-2/II-3/II-4 tachycardia protocols, and the tachycardia reference — and they agree:

Therapy	AMR Kansas City 2026
Manual defibrillation, VF / pulseless VT	200 J biphasic
Synchronized cardioversion, SVT	50–100 J
Synchronized cardioversion, atrial fibrillation	200 J biphasic (cites AHA 2025)
Synchronized cardioversion, VT with a pulse	100 J

Every item whose answer is a shock names the joules for that rhythm, and the suite asserts it.

This corrects two earlier mistakes of mine, in opposite directions. I first stripped the fixed figures out on the grounds that the 2025 AHA no longer sets them. That was wrong twice over: the AHA does set the atrial fibrillation one — 200 J or more on a biphasic unit, a reversal of the older low-and-climb approach, because a 200 J first shock converts more often and is less likely to induce VF — and the service sets all of them regardless of what the AHA does. The protocol's own citation of AHA 2025 for the A-fib figure is what confirmed it; ahajournals.org and cpr.heart.org are both unreachable from the environment this was written in, so the guideline could not be read directly. The protocol is the better authority here anyway: it is what the students are examined and employed against.

LIFEPAK is a trademark of Stryker. This is a training simulation and is not a medical device. Nothing here has been through clinical review by the Medical Director; it is checked against published sources, which is not the same thing.

Backend

Academy completions and Ask the Educator submissions can be mirrored to Supabase so a CE record survives a lost or wiped phone. The app stays offline-first: localStorage remains the source of truth for what a learner sees, and writes that cannot go out are queued and retried on reconnect.

All five academies are covered — the four in the repo root plus the Ventilator Training Academy. Caregiver Signature Form records go to the same place; the PDF stays the deliverable and signature images are never uploaded.

Learners can optionally link an email, which turns a self-typed name into a verified identity and lets their progress follow them to another device.

educator-dashboard.html shows completions (with CSV export), the Ask the Educator inbox, and filed caregiver forms to approved educator accounts. It is read-only: nothing there can edit or delete a record. Access is an allowlist in the database, not a check in the page.

Setup lives in supabase/README.md. Until a project is configured the sync layer is an inert no-op and every page behaves as it did before.

Confidential — For Authorized AMR EMS Personnel Only