Lesson L3-1 · authored fragments (docs/43)
Primary review content — the AI-avatar narration, in order.
Before any medication goes to a resident, you have to be sure you've got the right person. This one matters, because the right pills going to the wrong resident can really hurt someone. Here's the trap: after a few shifts, you'll know your residents by face, and that starts to feel like enough. It isn't. Knowing a face is not a check, and you never want to guess on this one -- so make it an active check, every single time. Start with the MAR. Open it to that resident's page, and match the name on the MAR to the name on each medication label. Then confirm the person standing in front of you. Now, how you ask matters. Don't ask, "Are you Mary?" -- a resident may nod along, or say yes, to almost anything, so a yes like that tells you nothing. Instead, ask them to say their name. Something like, "Can you tell me your name?" -- and let them tell you. If your resident is nonverbal, or can't give you a reliable answer, check their photo. So it comes down to three things that all have to match: the name on the MAR, the name on each label, and the resident in front of you. All three, every time. Never assume.
13 of 13 critical claims are verified_by: null (Gate-2 SME sign-off pending → preview only, production-blocked under the Rule-3.6 waiver). · runtime budget: 4:30
# Concept mirror — MA.L3-1.C (Verify the Client — the Right Person) > **MIRROR / NOTES ONLY.** The authoritative concept entry is **docs/42 §E, `MA.L3-1.C`** > (Instructional Domain Model). This file is the local traceability ledger + decision queue for the > derived-asset bundle; if it ever disagrees with docs/42, **docs/42 wins** (docs/43 §22 authority flow). | Field | Value | |---|---| | `concept_id` | `MA.L3-1.C` | | `competency_id` | `C-3.1` (⇄ `C-2.2` Right Person) | | `lesson_ids` | `[L3-1, L2-2]` + all route lessons (cross-cutting — identity is verified on every route) | | `clinical_risk` | **Critical** → 100% KC mastery, sim force-fail (person-gate), SME-protected KC key | | `medication_error_prevented` | Wrong-person administration — the right dose reaching the wrong client | | `sme_status` | `in_review` (docs/42) | | `apd_traceability_status` | sourced — one open checkpoint anchor (`TODO(source-L31C-1)`) + one SME-expansion flag + one `[Waiver360-added]` mnemonic | ## Rule-3.6 author-ahead waiver (recorded 2026-07-05) The concept is `in_review`, not `approved`. Per docs/42 §I.1 / docs/43 §4, this bundle is authored **ahead of concept approval under a Rule-3.6 waiver** as part of the Oral Route reference build. Consequence: **every fragment in this bundle is `sme_status: "draft"`, preview-flag only, and production-publish-blocked** until (a) the concept reaches `approved` and (b) a Gate-2 SME sign-off is recorded in `content_revisions.sme_signoff_*`. ## Source ledger (projects into `payload_json.sources`) - `curriculum_source`: "APD BMA Module 1 (identity verification; the yes-nod trap) + Module 2 (name/photo match); Mod3 v2 Q9 checkpoint" - `form_reference`: "APD Form 65G-7.008 A (MAR) — front header / client name field" | # | Claim | Tag | Source | verified_by | |---|---|---|---|---| | C1 | Client identity must be **actively confirmed** before administering; a familiar face is not a check; never assume | `[APD]` | Module 1 s65; Module 2 s55 | null | | C2 | Confirm identity with an **open-ended** question (ask the client to state their name), **not** a leading yes/no question like "Are you Mary?" — the **yes-nod trap** | `[APD]` | Module 1 s65 | null | | C3 | If the client is **nonverbal** (or cannot give a reliable answer), verify with their **photo** | `[APD]` | Module 2 s55 — **"or unable to give a reliable answer" is an SME expansion** (see legacy block-04); flag | null | | C4 | **Triple match**: name on the MAR ↔ name on each medication label ↔ the client in front of you (photo) | `[APD]` | Module 2 s55; MAR front header/name | null | | C5 | The **correctly-phrased identity question** keyed for the KC is an open-ended request for the client to state their own name; leading questions and recognition-by-familiarity are not verification | `[APD]` | Module 1 s65 + checkpoint **Mod3 v2 Q9** — **`TODO(source-L31C-1)`** (Q9 verbatim keyed answer not in hand) | null | | C6 | Reinforcement rail label **"person before pills" / Right #1** | `[Waiver360-added]` | Mnemonic scaffolding of APD's Right-Person-first framing (Module 1); not APD verbatim | null | ## Patient-safety chain (owner directive — Critical concept) 1. **Medication error prevented.** A MAP hands a correctly-prepared dose — right drug, right dose — to the wrong resident, because identity was *assumed* instead of *confirmed*. The right pills reach the wrong person, and a medication that was safe for one resident can harm another. 2. **Why staff make this mistake — the real-world reason.** It is almost never "forgot." In a small group home the MAP sees the same handful of residents every shift and knows them on sight, so saying a name out loud feels redundant — even a little silly. Under med-pass time pressure the shortcut is the leading question — "Are you Mary?" — and a resident who wants to be agreeable (or who can't give a reliable answer) nods yes to almost anything. Familiarity plus a yes-nod *feels* like a check, but it confirms nothing — this is the **yes-nod trap**. 3. **How the lesson/transcript prevents it.** Block `l31-verify-client` replaces "I know who this is" with an active check: open the MAR to the resident's page and confirm identity with an **open-ended** request — ask the resident to state their own name — never a leading yes/no question. It names the yes-nod trap out loud, gives the **nonverbal fallback** (verify by photo when the resident can't give a reliable answer), and ties it together as the **triple match**: name on the MAR ↔ name on each medication label ↔ the resident in front of you. 4. **How the simulator reinforces it.** The MAR Simulator uses a **person-gate**: the MAP must click the resident's name and photo to confirm the match before the administration modal will close. Recognition alone will not advance the scenario — skipping the identity confirmation stops the med pass. 5. **How the KC verifies mastery.** The knowledge check asks the MAP to pick the **correctly-phrased identity question**; every distractor is a leading yes/no question. Choosing a leading phrasing is a miss — mastery is "confirm identity with an open-ended name request, not recognition-by-familiarity," verified against the exact trap that causes wrong-person errors. ## Decision queue → owner / SME (Michele + Nicole) 1. **`TODO(source-L31C-1)` — confirm checkpoint Mod3 v2 Q9 verbatim + keyed answer** for the identity-question KC. The bundle keys the correct answer as the *open-ended name request* (from Module 1 s65 + the legacy verify-client narration). Until SME confirms Q9, C5 stays `verified_by: null` and the KC key stays preview-only. **This is the item that gates the KC key to production.** 2. **Confirm the C3 expansion** — "or unable to give a reliable answer" was added to APD's "nonverbal → photo" rule by the SME (legacy block-04). Confirm it reads as an [APD-aligned] SME edit vs. strictly `[Waiver360-added]`, so the narration/KC-B key can be verified. 3. **Confirm C6 framing** — "person before pills" / Right #1 acceptable as `[Waiver360-added]` reinforcement of APD's Right-Person-first numbering? 4. **Person-gate sim mechanic** — the "click name + photo before the modal closes" gate is a new MAR Simulator interaction (see `cms_enhancement_recommendations`); not built by authoring. ## Bundle → lesson storyboard frame `V-L3-1-04` (triple-match graphic). Manifest/README owned by the later lesson-aggregation step (docs/43 §18).
You're about to give Mary her morning medications. She's verbal and can answer questions. Which is the correct way to verify you have the Right Person?
Why: An open-ended question makes the client state their own name — that's active confirmation. APD (Module 1 s65) warns that leading yes/no questions invite a nod from almost anyone, so they don't actually verify identity.
Error prevented: Wrong-person administration — the right dose reaching the wrong client because identity was assumed, not confirmed.
Misconception: "'Are you Mary?' is verification" and "familiarity replaces checking."
You're about to give Robert his morning medications. Robert is nonverbal and cannot tell you his name. How do you verify you have the Right Person?
Why: When a client can't state their name, the photo is the verification. Match the photo to the person and the name on the MAR — that's active confirmation without relying on an answer he can't give.
Error prevented: Wrong-person administration to a client who cannot self-identify.
Misconception: That the photo path is optional, or that a nod / staff say-so / room context substitutes for confirming identity when a client can't state their name.
Entry: Learner opens the administration modal for any dose in the Don Montana MAR Simulator. A 'person-gate' must be cleared before that modal can close and initials can be recorded.
Mastery: 100% — force-fail on the person-gate; no dose administration without an active identity confirmation.
Force-fail: FORCE-FAIL. The administration modal cannot close and no dose can be recorded until identity is actively confirmed. The sim does not silently fail: it surfaces the same compassionate-framing modal as the KC, with a direct link to the L3-1 verify-client re-anchor clip and a 'Try again' CTA.
Shot: Triple-match graphic: MAR name ↔ medication label name ↔ client photo, joined by checkmark arrows. Split-screen two-take of a leading vs. an open-ended question.
Camera: Hold on the triple-match graphic while the three names line up; split-screen for the question contrast; no camera move during the contrast so the wording reads.
Avatar: Voiceover.
Visual: Make identity an active three-way match and expose the leading-question trap (MA.L3-1.C visual: triple-match MAR↔label↔photo; animation: leading vs. open-ended, two takes). Critical concept.
Animation: Three checkmark arrows connect MAR↔label↔photo in sequence. Leading-question take dims and is rejected; open-ended take highlights. Photo path slides in for the nonverbal branch.
Infographic: Callouts: 'Name on MAR = name on label = person in front of you'. Two caption cards: dim/rejected '“Are you Mary?” → nod trap' vs. highlighted '“Can you tell me your name?”'. Nonverbal path: 'no reliable answer → check the photo'. [Waiver360-added] 'person before pills · Right #1' chip.
Triple-match graphic (MAR name ↔ label name ↔ client photo joined by checkmarks) plus a leading-vs-open-ended question two-take.
🖼 image prompt: Flat-vector triple-match diagram: a MAR name header, a medication label, and a synthetic client photo card, joined by three green checkmark arrows; caption 'name on MAR = name on label = person in front of you'; a small nonverbal-path note 'no reliable answer → check the photo'. Neutral clinical palette, synthetic client only, no real names, no logos. [APD] Module 1 s65 + Module 2 s55; 'person before pills · Right #1' chip is [Waiver360-added].
🎬 video prompt: Three checkmark arrows connect MAR↔label↔photo in sequence; a caption card '“Are you Mary?”' dims and is rejected while '“Can you tell me your name?”' highlights; the photo path slides in for the nonverbal branch; transparent background. [APD] open-ended-vs-leading identity check.
Target runtime: 4:30 · 7 chapters · not rendered (url: null)