# Mechanism Diligence narration and animation standard

Working guide for original educational stories. It applies the existing ten-perspective opening prompts to additional medicines. It is a writing and production standard, not a claim that a draft has passed clinical review.

## What the research informs

- **JoVE:** use a clear introduction, a sequence of observable steps, representative evidence/results, and a conclusion. Give each chapter a visible label; synchronize narration with the action; leave graphics legible long enough to follow. JoVE-produced videos use human voice-over and editorial production. Our browser voices do not meet that production standard, and the interactive stories should not be presented as JoVE-quality video. Source: [JoVE author video guidelines](https://www.jove.com/authors/publication/author-submitted-video) and [JoVE-produced video workflow](https://www.jove.com/authors/publication).
- **Educational YouTube examples:** George Zaidan's [TED-Ed pain-reliever explainer](https://www.youtube.com/watch?v=9mcuIc5O-DE) and the Institute of Human Anatomy [ibuprofen explainer](https://www.youtube.com/watch?v=Zeem6k2bDFk) were selected earlier for approachable narrative and visible anatomy. Selection was editorial, not a verified like-count ranking or clinical endorsement. The saved review notes that full content review was incomplete for some other candidates. We learn from concise framing and concrete visual transitions; we do not copy their narration, art or footage.
- **Cell Press graphical abstracts:** make the main biological message understandable quickly, with one dominant relationship and restrained, legible labels. The graphical abstract is an entry point, not a substitute for evidence. Source: [Cell Press graphical abstract guide](https://crosstalk.cell.com/hubfs/Files/GA_guide.pdf).
- **Nature, Science, Cell and PNAS research:** use relevant peer-reviewed work to identify the actual experimental model, perturbation, comparator, exposure and observed result. A paper's figure or graphical abstract is not enough to generalize a mechanism to people. Follow the cited methods, supplemental data and limits; redraw original teaching diagrams and retain the evidence link. Do not reuse journal artwork without permission.

## Shared prompt

> Draft a narrated, evidence-linked mechanism story for **[generic medicine; relevant brand names]** for **[one of the ten perspectives]**. Use only the supplied evidence ledger. Begin with that perspective's distinct real question; do not imply that you know why a visitor takes or was prescribed the medicine. Explain the medicine's intended benefit and the population/evidence supporting it before discussing potential harms. Move in a causal sequence from biological context to target/pathway, drug action, exposure or timing when established, human-relevant outcome, and important limits. Make one claim per spoken beat. State whether evidence is biochemical, cellular, animal, clinical or label-based. Keep preclinical results separate from established human effects. Explain uncertainty and plausible alternative causes of symptoms. Never infer individual benefit, treatment choice or dose. Use short, speakable sentences with varied rhythm, pronounceable abbreviations and no dramatic hook. For every beat return: (1) narration and subtitle text, (2) one original visual action, (3) evidence/source IDs, (4) model and uncertainty labels, (5) a proportionate safety boundary, and (6) review questions for a clinician, pharmacologist and language reviewer. If evidence does not support a claim, say so or omit it.

## Benefit-first, proportionate safety prompt

> Write for calm understanding, not fear or persuasion. Lead with the medicine's supported purpose and what benefit was observed, for whom, and in what kind of evidence. Do not promise that an individual will benefit. Describe common effects as common only when reliable sources say so; distinguish them from uncommon or rare serious harms, and include a frequency or relevant risk context only when the source supports it. Do not turn a rare event into a prediction, use alarming language, or present a long undifferentiated list in the spoken story. Keep material contraindications, important interactions, urgent warning signs, and complete label information available in a clearly labeled safety section. A calm tone must never hide, soften into ambiguity, or omit a risk that could change a patient's action. Explain that effects are possibilities, not predictions; tell people to discuss concerns with their clinician or pharmacist and not to start, stop, or change treatment from this story. Do not call a medicine "safe" in general: suitability depends on indication, product, dose, duration, health conditions, other medicines, and the person. Frame the care team as a partner, not an opponent.

## Perspective prompt cues

Use one distinct opening and choose emphasis, order and vocabulary for the selected perspective. Keep the underlying scientific claims the same.

| Perspective | Opening and emphasis |
|---|---|
| Patient | What effect is intended, what might change, and what should not be assumed? Speak directly and calmly. Do not attribute symptoms to the medicine from timing alone. |
| Caregiver | What questions can be discussed with the person's care team? Do not make the caregiver diagnose or monitor beyond their role. |
| Pharmacist | Product, formulation, absorption, interactions, contraindications and label variation. |
| Clinician | Target-to-physiology chain, relevant clinical outcomes, patient context and evidence limits. Avoid treatment recommendations without full context. |
| Nurse | What should be confirmed in the treatment plan and what change should be escalated or reassessed? Do not create orders. |
| Student | Explain the causal chain and distinguish mechanism, pharmacodynamics and clinical outcome. |
| Scientist | Connect target engagement to pathway and phenotype; specify model, concentration, perturbation and unresolved alternatives. |
| Chemist | Connect structure, formulation and exposure only where evidence supports it; distinguish a schematic from a measured binding pose. |
| Investor / diligence analyst | Separate established evidence, proposed mechanism, clinical value and remaining failure points. Never imply commercial success or investment advice. |
| Other | Ask what part of the medicine's story the visitor wants to follow; do not infer identity or purpose. |

## Animation and narration rules

1. One visible process per beat: highlight the actor, show the direction of change, then reveal the consequence. Use a persistent orientation cue when moving between body, cell and molecule.
2. Chapter labels and narration must use the same terms. Keep the active subtitle near the animation, preserve a scrollable transcript, and let people jump to any chapter or sentence.
3. Timing is an explanation aid, not biological time. Label schematic movement, simplified scale and omitted steps. Never imply tissue-selective delivery, exact target occupancy or measured kinetics unless supplied evidence actually measures it.
4. Use a neutral, well-paced voice; keep level and pronunciation consistent. A browser voice is only a fallback. For production narration, use recorded or high-quality commissioned voice, edit noise/reverb, and review pronunciation and synchronization with a human listener.
5. Provide mute, captions and a visible transcript. A caption-only mode is not proof that audio quality is good.

## Evidence and review checks

- Keep a claim-by-claim ledger with source URL, exact passage or figure/table, species/model, dose or concentration, route, comparator, endpoint, date checked and reviewer.
- Prefer current official prescribing information and government/patient resources for product scope and safety; use primary studies and rigorous reviews for mechanisms. A high-impact journal name is not a substitute for checking the methods or whether the paper's claim matches the narration.
- Distinguish: established drug action; supported but context-dependent pathway; hypothesis; illustrative diagram; unavailable evidence. Show which is which.
- Include confounding and disease progression when a symptom worsens. Temporal order alone does not establish drug causation; underlying disease, concomitant medicines and other causes may explain it.
- Before public release, have a clinician/pharmacologist check claims and safety, an independent language reviewer check translations, and representative users test clarity. Until then label narration and translations as drafts.

## Existing-story update rule

When revising an existing medicine, audit each audience and language against the benefit-first prompt. Put the expected benefit before the caution; replace comparisons or sweeping phrases such as "not safe" with specific, evidence-backed context; say when a serious event is rare only when a source supports that frequency; and distinguish common effects from serious red flags. Preserve clinically important warnings and the full safety information. Do not remove a warning just to make the story feel reassuring. Invite a patient to ask the care team a question rather than challenging or changing the plan on their own.

## Current implementation note

Atorvastatin, levothyroxine and metformin are selected from the 2024 U.S. prescription-fill list (not a global popularity measure). The first stories are concise mechanism drafts in English, Spanish, Hindi, Kannada and Marathi. Kannada and Marathi translations need human review; all four medicines need independent clinical review. The added narrated pages provide perspective-specific openings and shared mechanism sections. They are not ten separately validated clinical scripts per drug. The patient openings and ibuprofen patient safety wording were revised to state benefits first and place risks in context; label warnings and urgent signs remain available. This is a tone and organization pass, not clinical approval or a claim that any medicine is risk-free.