Designing Emergency Information for People Under Stress

# Designing Emergency Information for People Under Stress

A person finds someone unconscious on the floor. Their hands are shaking. Their heart rate is elevated. They may not have medical training. They may not read fluently in the language the emergency app uses. They pull out their phone and open whatever guidance is available.

What they encounter is often a menu. A list of conditions. A paragraph explaining the difference between cardiac arrest and fainting. A diagram with unlabeled icons. A video that requires buffering on a weak connection. A "terms of use" screen.

This is not a failure of the user's intelligence. It is a failure of design.

Emergency information is almost universally authored for a user who is calm, literate, attentive, and connected. The person standing over an unconscious body is none of those things. They are experiencing acute stress — and stress changes what the brain can do.

This article is about that mismatch. It is not medical guidance. It does not tell readers what to do in an emergency. It examines what research and authoritative communication frameworks tell us about how emergency information systems should be built for people whose cognitive capacity is temporarily, predictably, and severely reduced.

The design assumption that fails first

Most emergency information — pamphlets, apps, warning systems, dispatcher scripts, wall posters — is designed around a implicit user profile:

• Can read at approximately a high-school level
• Has time to scan options and choose the correct path
• Can hold several pieces of information in working memory simultaneously
• Can switch attention between the information source and the emergency scene
• Has a stable internet or phone connection
• Shares the cultural and linguistic conventions of the information designer

This profile describes a preparedness workshop attendee. It does not describe a bystander in crisis.

The gap is not anecdotal. The U.S. National Institute of Standards and Technology, in its guidance on emergency communication for building occupants, notes that verbal comprehension can drop an average of four grade levels under high levels of stimulation, stress, and distraction — all of which are present during emergencies. The CDC's Crisis and Emergency Risk Communication (CERC) program states plainly that when people are scared or anxious, they have difficulty taking in and remembering large amounts of information.

If a preparedness brochure is written at a tenth-grade reading level, a stressed reader may effectively be trying to comprehend sixth-grade material — except they are also managing fear, noise, time pressure, and a person who may be dying. The arithmetic is brutal: the information gets harder to process exactly when it becomes most consequential.

Hypothesis (not yet field-validated at scale): Emergency information systems that test comprehension only with calm, literate users in usability labs will systematically overestimate real-world performance. Stress-validated testing should be a design requirement, not an advanced practice.

What stress does to cognition — and why it matters for design

Stress is not merely an emotional state. It is a physiological condition that reallocates cognitive resources toward immediate threat response and away from complex reasoning.

A 2025 narrative review of cognitive neuroscience in crisis management describes how acute stress activates limbic structures — particularly the amygdala — producing hypervigilance and emotional reactivity that can override the prefrontal cortex processes responsible for analytical reasoning, working memory, and deliberate decision-making. The brain, in effect, trades breadth for focus. That trade has direct design implications.

Tunnel vision and narrowed attention

Under high stress, people experience perceptual narrowing — often called tunnel vision. They focus intensely on the central element of a situation and miss peripheral information. Research on law enforcement and civilian eyewitnesses has found that individuals may look directly at peripheral hazards — including mock improvised explosive devices placed in their field of view — and fail to register them mentally. In one study, all participants had visually fixated on a peripheral IED, but only about 1% reported it.

For emergency information design, this means a user may stare at an injured person and fail to notice critical on-screen instructions, environmental hazards, or secondary actions (such as sending another bystander to call for help). Information cannot assume the user is monitoring the screen continuously. It must be redundant, persistent, and recoverable.

Auditory exclusion

Stress also affects hearing. The phenomenon known as auditory exclusion — temporary reduction or loss of hearing under acute stress — has been documented in police officers during gunfire and in experimental settings where elevated heart rates impaired auditory processing. Research cited by situational awareness specialists suggests hearing ability may begin to decline when heart rate exceeds approximately 175 beats per minute.

The design implication is stark: voice-only phone guidance, often considered the gold standard for dispatcher-assisted emergencies, may fail precisely when the user is visually focused on a patient and cognitively overloaded. Conversely, a user reading text instructions may miss auditory cues from the environment. Multimodal redundancy — visual plus audio plus haptic feedback — is not an accessibility enhancement. It is a stress-design requirement.

Cognitive load and working memory

Cognitive load theory, applied to emergency contexts, distinguishes between the intrinsic difficulty of a task (performing chest compressions), the extraneous load imposed by poor information design (navigating a confusing interface), and the germane load of actually learning and executing the task. Under stress, extraneous load becomes catastrophic. Every menu, every jargon term, every unnecessary choice consumes working memory that the user needs for the emergency itself.

Research on interaction modes for emergency mobile applications proposes a useful framework: when cognitive capacity is fully consumed by the emergency, the interface should require one-click or automatic interaction; when partial capacity remains, guided step-by-step interaction with simple choices; only when the user is safe and calm should full interactive modes be appropriate. Most emergency apps are designed for the third state. Most users are in the first or second.

A 2021 study of U.S. county emergency managers found that perceived information overload from digital sources was associated with higher stress and a shift toward quicker, less reflective thinking. If trained emergency professionals degrade under information overload, untrained civilians facing a personal crisis almost certainly degrade further.

Why most emergency information fails under stress

Synthesizing across CDC CERC guidance, WHO emergency communication principles, NIST building communication research, and peer-reviewed studies on disaster literacy and inclusive alerting, several recurring failure modes emerge.

Too much information, too soon

Emergency messages often begin with background context: what cardiac arrest is, how it differs from a heart attack, when to call for help versus when to start compressions. CERC guidance recommends the opposite: initial messages should be simple, immediately relevant, and free of background. Repetition of the main action helps compensate for shortened memory retention under anxiety.

Many apps and pamphlets violate this by front-loading explanation before action. Under stress, explanation without immediate action steps is often lost entirely.

Text-heavy design in low-literacy contexts

The disaster literacy model proposed in _Disaster Medicine and Public Health Preparedness_ defines disaster literacy as the ability to read, understand, and use information to make informed decisions during disasters. The model's central observation is that the literacy demands of preparedness materials frequently exceed the literacy skills of the most vulnerable populations — including adults with physical, mental, and educational disabilities.

A 2024 synthesis on inclusive emergency alerts found that many populations may need voice delivery because they lack the literacy skills to interpret text-based messaging. In multilingual contexts — including much of India — this is not a minority concern. It is the default condition.

Designing text-heavy emergency information without audio, pictorial support, and plain-language alternatives is designing for exclusion.

Abstract visuals without labels or context

Visual-first design is often advocated as the solution to literacy barriers. The evidence is more nuanced. A cross-cultural study of disaster pictorials found that symbols designed as "universal" by Dutch designers were comprehended significantly better by Dutch participants than Chinese participants — even when the relative difficulty ranking of symbols was similar across groups. Context photographs improved comprehension for both groups.

Abstract, minimalist, unlabeled icons — the kind common in modern app design — perform particularly poorly under stress. Industry analysis of emergency and pandemic apps documented users skipping key steps or exiting systems early because they could not determine what unlabeled icons meant, especially in low-light or high-anxiety conditions.

The design principle is redundancy: icon plus label plus contextual image where possible. Never icon alone.

Navigation that assumes calm decision-making

Emergency apps frequently present scenario menus: "What happened?" followed by a list of twelve conditions. Under stress, this is a high-stakes multiple-choice exam administered to someone whose working memory is compromised.

NIST guidance recommends message templates pre-drafted for specific scenarios to reduce error under stress. The equivalent in interactive design is scenario detection or progressive narrowing ("Is the person breathing?" → yes/no branch) rather than open-ended category selection. Research on dispatcher-assisted CPR demonstrates that structured, sequential prompting outperforms unstructured information delivery — and that video-assisted guidance with simultaneous visual demonstration can improve performance over voice-only guidance, though the benefit depends on operational design rather than video presence alone.

Positive framing neglected — or over-applied

CERC recommends positive action framing when feasible ("drink bottled water") rather than negative framing ("do not drink the water"), because negatives can be confused or forgotten under heightened emotion. This is a valuable default for design.

But the guidance also acknowledges that some hazards require negative instructions. Design systems need both modes, with positive framing preferred for action steps and negatives reserved for critical prohibitions — clearly highlighted, repeated, and paired with the acceptable alternative action.

Connectivity assumed

This article focuses on cognitive design, but stress and connectivity failure often co-occur. During disasters, cell tower failures affect a significant fraction of infrastructure. An emergency information system that requires live loading of instructions fails at the intersection of two predictable vulnerabilities. Offline-capable, pre-cached content is a cognitive design issue as much as an infrastructure one — loading delays add extraneous cognitive load while the user waits.

Design principles for stress-accessible emergency information

The following principles synthesize authoritative guidance and research evidence. They represent design direction, not validated prescriptions for every context. Where evidence is strong, it is noted. Where it is extrapolated from adjacent domains, that is stated.

1. Design for the worst cognitive moment, not the best

Assume the user has four grade levels less reading comprehension than your content targets. Assume working memory holds one instruction, not three. Assume attention is split between the screen and a person in danger. Assume hearing and peripheral vision are degraded.

If the design works under these assumptions, it will work better for everyone — including calm, trained users. The reverse is not true.

2. One action per screen; zero navigation during acute guidance

Each screen or voice prompt should contain a single action step. "Check if the person is breathing." Not: "Assess responsiveness, check breathing, note skin color, and determine if CPR is needed."

Branching should be binary where possible: yes/no, conscious/unconscious, breathing/not breathing. Multi-option menus should be deferred until the user's cognitive state permits them — or eliminated entirely in favor of automated scenario detection.

3. Plain language as a safety requirement, not an editorial preference

WHO emergency communication guidance specifies plain language, familiar languages, and visual formats. The CDC CERC program requires avoidance of jargon and technical terms in initial messages. Plain language is not dumbing down. It is precision under constraint.

Write at a sixth-grade reading level or lower for acute guidance. Use short sentences. Use active voice. Name the action before explaining why. Test with representative users — including low-literacy and non-native speakers — not only with colleagues.

4. Multimodal redundancy by default

Combine text, audio, icons with labels, and haptic feedback. Allow the user to switch modes in real time as situational demands change — a requirement identified in disability-inclusive emergency app research, where participants noted that stress impairs the ability to rely on any single communication channel.

Voice prompts should be repeatable on demand. Visual instructions should remain visible while audio plays. Do not force a choice between looking at the patient and reading the screen.

5. Layer information; do not dump it

A layered information architecture — basic action layer first, extended explanation available on demand — reduces cognitive burden while accommodating users who want more detail. Research on public health communication diagrams found that even college-educated participants preferred basic layers over extended information during initial engagement, though both were valued.

The basic layer is the emergency product. The extended layer is the reference product. Conflating them is a design error.

6. Test with stress, not just with usability

Standard usability testing is insufficient. Design evaluation should include time pressure, background noise, simulated physical exertion, and divided-attention tasks — conditions validated in high-risk decision-making research as influential on performance. Few emergency consumer products currently meet this bar. This is a major evidence gap and a major design opportunity.

7. Cultural and linguistic validation of all visuals

Every pictogram, icon, and color-coded element should be tested with users from the populations who will depend on it — not assumed universal because it follows ISO or Western conventions. Context photographs improve comprehension across cultural groups. Labels improve icon comprehension under stress. Color coding alone is unreliable across educational and cultural backgrounds.

8. Persistent "call for help" as the universal fallback

When in doubt, the system should make calling professional emergency services (112 in India, 911 in the U.S., or local equivalents) the always-visible, zero-navigation primary action. Research on the civilian window (see related article on first responders before professionals arrive) suggests that effective help-seeking may be as consequential as specific bystander interventions in many scenarios. No design should make calling for help harder than finding scenario-specific guidance.

Practical implications for emergency guidance platforms

For teams building emergency information platforms — whether public sector alert systems, NGO preparedness tools, or civilian guidance apps — several actionable implications follow from the evidence.

Pre-cache everything that matters. Acute guidance content should be available offline. Loading states during crisis are unacceptable extraneous load. Scenario content, pictograms, audio prompts, and plain-language scripts should ship with the application or be downloadable in advance.

Build a stress-state interaction model. Map user journeys for three cognitive states: crisis (one-action, voice-first, auto-detect where possible), elevated concern (guided binary branching), and preparedness (full content, richer explanation). Most products collapse these into a single interface.

Instrument comprehension, not just clicks. Analytics should measure where users abandon guidance, which steps generate repeat requests for audio replay, and which branches are most frequently mis-selected. These are proxies for cognitive failure points.

Treat accessibility as the core user profile. Disability-inclusive emergency app requirements research demonstrates that features designed for people with intellectual, hearing, and visual disabilities — pictograms, plain language, multimodal interaction, adjustable presentation — improve usability for stressed users generally. Designing for the least prepared person is designing for everyone under stress.

Do not substitute AI chat for structured guidance without evidence. There is no authoritative evidence that generative AI conversational interfaces improve bystander performance under stress. There is substantial reason for caution: hallucinated instructions, unpredictable response length, and the added cognitive load of reading and evaluating unstructured text are precisely the wrong characteristics for acute crisis design. Structured, validated, sequentially presented content remains the evidence-backed approach.

Integrate with professional response rather than replacing it. Dispatcher-assisted CPR research shows that remote professional guidance combined with well-designed visual support improves layperson performance. Platform design should facilitate connection to professional dispatch — sharing location, scenario type, and actions already taken — rather than positioning the app as a standalone substitute for emergency services.

What we do not yet know

Honesty requires naming the gaps.

There is limited field evidence testing emergency information comprehension with validated stress induction in diverse lay populations — particularly in LMIC contexts with high linguistic diversity and variable literacy. Most visual communication research uses calm participants. Most CPR guidance research uses simulation with young, educated volunteers.

There is no consensus on the optimal number of information units per message during acute crisis. CDC suggests keeping initial messages to a few key points; UX practitioners advocate one action per screen. These have not been rigorously compared for bystander medical scenarios in real emergencies.

There is a tension in the literature between minimal messaging (recommended under information overload) and richer messaging (which some studies find improves behavior change during preparedness campaigns). The likely resolution is phase-dependent design — minimal during acute action, richer before and after — but this hypothesis needs empirical validation.

There is almost no India-specific research on emergency information comprehension under stress across the country's linguistic and literacy landscape. Design principles from U.S. and European contexts may transfer incompletely.

And there is a fundamental open question: can information design alone close the gap between bystander willingness and bystander capability — or does it require pairing with training, legal clarity, and community infrastructure? The evidence from bystander CPR research suggests information tools improve performance when they offload specific procedural steps, but they cannot replace recognition training or the confidence to act. Design is necessary. It may not be sufficient.

The question beneath the question

Emergency systems invest heavily in professional training, equipment, and dispatch infrastructure. They invest comparatively little in the cognitive design of the information that untrained civilians receive at the point of need.

This is partly because stress-degraded cognition is invisible in planning meetings. Designers and policymakers review materials while calm, literate, and focused. The content seems clear. The icons seem intuitive. The reading level seems acceptable. The test is performed without elevated heart rate, without a person bleeding on the floor, without the auditory chaos of a crowd.

The question this article raises is not whether civilians should receive emergency guidance. They already do — from apps, posters, dispatcher scripts, social media, and each other. The question is whether that guidance is designed for the brain state they actually have when they need it.

If it is not, the failure will look like user error. It will look like panic, ignorance, or indifference. It will rarely be recognized as what it actually is: a predictable mismatch between information design and human cognition under stress.

Designing for the least prepared person means designing for the most stressed person. The research suggests that is not a niche accessibility concern. It is the defining constraint of emergency information systems — and one that most current systems have not yet taken seriously.