Teaching Autism Safety Skills Without Turning to Surveillance

Two approaches shape autism safety planning: containment, with locks and trackers, versus communication, including through AAC devices, which asks what prompted the leaving and how the person would ask for help. The 2012 Pediatrics study of 1,218 children with autism found that 49% of parents reported an elopement attempt after age four, and 26% said the child had been missing long enough to cause concern.

Those numbers make safety planning a core SLP responsibility. A safety plan built only on surveillance assumes the person cannot communicate their way out of danger. The stronger plan treats communication as the safeguard.

Why Safety Planning Belongs in the SLP's Scope

Safety planning for an autistic client means preparing for what happens when someone leaves a safe space unexpectedly, encounters a stranger during a crisis, or needs help from someone who does not know them. Too often in autism speech therapy this gets treated as a behavior problem to manage rather than a communication gap to close, and that framing puts the wrong professionals in the driver's seat.

Elopement Is a Communication-Access Problem

When a child wanders, the danger is not the walking away itself. It is what happens next: a police officer, a neighbor, or a stranger at a busy intersection has no way to understand who this person is, what they need, or how to help them communicate. That gap is exactly where speech-language pathologists work every day. Reframing elopement risk through a communication lens, rather than a purely behavioral one, puts SLPs at the center of prevention and response planning, not on the sidelines.

AAC Training Is Safety-Critical Work

Teaching someone to use an AAC device, and coaching the people around them to actually understand and respond to it, is not an enrichment activity to fit in after "real" goals are met. When a nonspeaking or minimally speaking person can indicate their name, a caregiver's number, or "I need help" to an unfamiliar adult, that skill can be the difference between a quick reunion and a prolonged, dangerous search.

This framing draws directly from Nikita Babkin's September 2026 article "Safety Without Surveillance" in Autism Spectrum News, which argues that a safety plan should also function as a communication plan. That thesis matters for clinical practice, and it rests on a foundation ASHA has long emphasized: AAC is not a privilege earned through demonstrated ability, it is a right to communicate that underlies every safety strategy that follows.

The Data Behind Wandering and Elopement Risk

Elopement is not a rare edge case, and the numbers make that clear. A 2012 study published in Pediatrics surveyed the families of 1,218 children with autism spectrum disorder, and what parents reported reframes wandering as a predictable clinical risk rather than an occasional crisis. For SLPs, these figures are the rationale for treating communication access as a safety intervention.

AAC Access Should Never Require Proving Cognitive Ability

No autistic person should have to pass a cognitive test to earn the right to communicate. Yet that is exactly what happens in clinics and school evaluations across the country when trials of AAC devices are delayed until a child "demonstrates readiness" through matching tasks, symbolic play checklists, or IQ cutoffs. ASHA defines AAC simply as tools and strategies that supplement or replace spoken communication when speech is unreliable or absent. Nothing in that definition mentions a minimum cognitive threshold, because none exists in the clinical literature.

Where Gatekeeping Hides in Funding Rules

The barriers show up less in formal "cognitive eligibility" rules and more in the fine print of who pays. Medicaid covers speech-generating devices in every state, and coverage for children under 21 is especially strong through EPSDT protections,² but prior authorization is standard practice nationwide. Louisiana's Medicaid policy requires proof that a beneficiary can physically and mentally operate the device³, a capacity screen that can deny access to the very individuals with the most complex needs. Pennsylvania requires six months of prior therapy before funding a device. Colorado and Medicare will not cover tablet-based AAC apps at all, only dedicated speech-generating devices. Private insurers typically demand a formal evaluation, documented medical necessity, and prior authorization, and because Medicaid is the payer of last resort, families with both coverages often need a private denial in hand before Medicaid will step in.

What SLPs Should Do Instead

None of this justifies waiting to introduce AAC. Presume competence and begin trials immediately, using low-tech options while high-tech funding paperwork moves through the system. Document functional communication need, not scored cognitive prerequisites, when building a case for coverage. Treat every delay as a clinical and safety cost, not just an administrative one.

The Safety Stakes Are Real

A person without reliable access to AAC for nonverbal autism cannot tell a caregiver they are in pain, cannot alert someone to danger, and cannot signal that they are lost or scared. If a child wanders and encounters a stranger, a police officer, or a crowded street, their ability to communicate distress may depend entirely on whether their SLP gatekept access months or years earlier. Safety planning starts with communication access, not after it.

Questions to Ask Yourself

Building a Communication Plan That Doubles as a Safety Plan

The CDC's current wandering guidance for families of children with disabilities lists seven core actions: keep the person's information current, secure the home, keep identification on the person, inform neighbors and school staff, alert first responders in advance, teach safety skills, and have an emergency plan ready to activate.¹ For SLPs, every one of those actions has a communication layer, and none of them work if the plan does not describe how the person actually communicates.

Core Components to Document

A plan that functions as both a communication profile and a safety document should specify:

• Signaling basics: how the person indicates yes, no, pain, fear, and "stop" , across their primary AAC system, gestures, facial expressions, and behavior that may look like refusal but signals distress.
• Backup methods: what to do if the device is lost, dead, broken, or taken away. A laminated card with core vocabulary, a printed communication board, or a wearable with key phrases keeps communication possible when the tech fails.
• Contact and context: who to call, where the person is likely to go (water, roads, rail corridors are CDC-flagged priorities)³, sensory triggers, and interests that can help redirect or locate them.
• Responder-facing information: a short profile responders can read in seconds, covering communication abilities, behaviors that might be misread, and how to approach.

Make It Travel, and Revisit It

HHS's National Autism Missing Person Alert Initiative and CDC guidance both assume responders can quickly access current identifying details.² That only happens if the plan is physically with the person: a device lock-screen image with name and one emergency contact, a wearable ID, a laminated wallet card, and a shared digital copy with caregivers and school. Schedule an SLP emergency preparedness review at least annually, and any time the AAC system, living situation, school placement, or medical picture changes.

Low-Tech Backup Is Non-Negotiable

AAC devices fail. Batteries die, screens crack, bags get left behind, and first responders may not know how to use a speech-generating device. Every plan should include a paper or wearable backup with core requests, a photo, the person's name, a caregiver number, and a brief line on how they communicate. The backup is not a downgrade, it is the version of the plan most likely to be in someone's hand during the emergency itself.

Safety Goals Across the Lifespan: Children, Teens, and Adults

Safety risks shift as an autistic person moves from childhood into adulthood, and so should the communication goals attached to them. The comparison below groups typical priorities by life stage, but stage is a starting point, not a rule: goals should track a person's current skills, communication profile, and daily environments rather than their birthday. A 17-year-old may still need elopement-focused work, and an 8-year-old may be ready for self-advocacy scripting.

Teaching Safety Skills Without Creating Compliance Dependence

The field is shifting away from safety drills that reward doing exactly what an adult says in the moment, and toward teaching autistic learners to recognize and respond to natural cues on their own. Yet there is no single peer-reviewed comparison that cleanly pits prompt-fading safety instruction against compliance-based training and shows one produces better real-world autonomy. The closest evidence comes from smaller studies, including a 2026 randomized comparison of 27 autistic children that found equivalent gains, mastery, and generalization across constant time delay, most-to-least prompting, and flexible prompt fading.¹

Build Your Own Evidence Base

Rather than waiting for a definitive trial, use a simple search process. In PubMed, Google Scholar, and association databases, combine terms such as "prompt fading autism safety skills generalization" and "compliance-based safety training autonomy." Cross-check these sources:

• BLS.gov for injury and safety statistics.
• School district websites for special education safety curricula and transition plans.
• ABAI, APA Division 33, ASHA, and the Autism Society for practice guidelines, position statements, and conference proceedings.

Look for systematic reviews and cited-by trails. A smaller peer-reviewed report found most-to-least prompting produced fewer errors during training for two participants and recommended most-to-least with time delay when learning history is unknown or errors increase problem behavior.² Response-prompting strategies have the strongest evidence for time delay, simultaneous prompting, and most-to-least prompting across autistic learners and learners with moderate to severe disabilities.

Questions That Reveal the Approach

Ask providers these questions, not as a script, but to compare programs:

• Prompt fading: How do you fade prompts across settings and people? What happens when a prompt is not available?
• Autonomy goals: Which self-advocacy and initiation targets are in the plan? Does the learner ever say no, ask for help, or flag danger unprompted?
• Generalization: How is mastery measured with unfamiliar partners and in new environments? What criteria show the skill is independent, not just cued?
• Data: How do you distinguish prompt dependence from learned safety responding?

Red Flags for Compliance Dependence

• The learner responds only when an adult gives a specific verbal or gestural prompt.
• The program measures compliance more than spontaneous help-seeking, refusal, or environmental scanning.
• No prompt-fading schedule exists, or prompts stay embedded in the skill.
• Safety instruction stops at following adult directions and does not teach recognizing unsafe situations, seeking a safe adult, or refusing unsafe requests.

Working With Families, Schools, and First Responders Without Over-Sharing

The most practical safety collaboration today is built on minimal disclosure, not open-ended sharing of a student's full history. First responder training is shifting toward communication and sensory support, and the SLP can help families decide what truly needs to travel with a child in an emergency.

Start With a One-Page Communication Passport

A useful passport fits on one side of one page. It includes the person's name, primary communication method, a few key phrases or symbols, sensory triggers, calming strategies, and one emergency contact. It does not include full diagnostic history, cognitive labels, or behavioral intervention records. Families and schools can use the same passport so a responder does not have to sort through school paperwork.

Training Programs Are Already Expecting This Level of Detail

National and local trainings now emphasize communication supports rather than diagnosis as the main identifier. The Autism Society's Safety on the Spectrum First Responder Program, IBCCES First Responder 2 CE Team Training, and Johns Hopkins All Children's Hospital's Law Enforcement Officer Training all include autism recognition, de-escalation, and communication strategies. On the state level, the Delaware Network for Excellence in Autism has trained 225 responders from 55 agencies since November 2022¹. The Autism Society of Central Virginia trained 177 police, EMS, and fire personnel in 2024 using autism-friendly strategies, scenarios, and sensory tools². Responders increasingly ask for "how does this person communicate" rather than "what is this person's diagnosis."

Coach Families on What to Disclose and What to Withhold

SLPs can role-play the first minute of an emergency with caregivers. Ask: Would this piece of information change how a responder approaches the person in the first five minutes? If the answer is no, leave it out. For example, the AAC devices a child uses and "she may not stop when asked" matter, while a full developmental report does not.

Keep School Coordination Focused on Access, Not Escalation

In IEP or safety meetings, push for language that centers communication access. Instead of "increase supervision after elopement risk," write "call the SLP if the person cannot access their AAC system; do not restrict movement as a response to communication breakdown." This frames safety as support, not surveillance.

Nonspeaking and Minimally Speaking Individuals: What Safety Looks Like

Absence of speech is not absence of communication, and safety planning that treats it that way puts people at risk. ASHA describes AAC as the tools and strategies that supplement or compensate for difficulties with spoken communication, and for a nonspeaking autistic person in an emergency, those AAC devices are the difference between being understood and being misread as noncompliant.

Map every modality the person actually uses

A safety profile should document the full range, not just the device. That includes AAC (high-tech and low-tech), gestures, facial expression, body movement and posture shifts, pictures, typing, and any individualized signal the person has developed with familiar partners. Write down what distress looks like for this specific person: pacing, a particular vocalization, hand placement, going still. A responder who does not know that stillness signals panic will read it as cooperation.

Build a low-tech backup before you need one

Devices get dropped, left behind, and run out of battery at the worst possible moment. Every plan should include a redundant layer that survives those failures. A laminated card set on a lanyard or clipped inside a jacket works well: one card for "I am hurt," one for "I need help," one for "I am lost," one with the person's name and an emergency contact. Practice pulling them out in calm conditions so the motor routine is familiar under stress.

Tell responders how to approach and how long to wait

This is the piece most plans skip. Spell out the approach: reduce noise if possible, stand at an angle rather than face-on, use short single-step sentences, offer a written or pictured option alongside speech. Then name a specific wait time, often 20 to 30 seconds after each question, and state plainly that silence does not mean the person did not hear or understand.

Presume competence throughout. Assuming someone cannot report pain, danger, or an unsafe person because they do not speak is the assumption that gets ignored in exactly the situations where reporting matters most.

Questions to Ask Yourself

Legal, Ethical, and Autonomy Considerations for SLPs

ASHA's Code of Ethics and the SLP scope of practice obligate speech-language pathologists to honor client self-determination as a core principle of service delivery, which makes safety planning a clinical and ethical responsibility, not a compliance afterthought. For autistic clients, that obligation collides with three tensions: guardian consent versus personal assent, protection versus autonomy, and compliance-based teaching versus trauma-informed practice.

Consent Is the Floor, Assent Is the Practice

A guardian's signature on a consent form authorizes treatment, but it does not erase the client's right to say no in the moment. Assent means continuously checking in: Does the client understand what will happen? Do they agree to participate right now? For nonspeaking individuals, assent may show up as body language, vocalizations, turning away, or selecting a "stop" symbol on an AAC device. Treat withdrawal of assent as information, not noncompliance to override.

Autonomy and Trauma-Informed Safety Planning

Forced or surprise safety drills can create stress and teach a child that their body is not their own. Trauma-informed practice means building safety skills through predictable routines, client-chosen roles, and rehearsal that respects sensory and communication needs, a core principle of Neurodiversity-Affirming Speech Therapy. An autistic person who has experienced restraint or coercion may interpret "safety training" as another demand to comply. Instead, frame goals around self-protection, choice, and asking for help, not around obedience to adult commands.

Documenting Assent Next to Consent

Clinical records should capture the consent conversation and the ongoing assent process. Note what the client agreed to, how they communicated agreement or refusal, and any adaptations made. This documentation supports ethical practice and helps future providers understand what safety planning looked like from the client's perspective. When guardianship limits decision-making authority, the SLP still documents efforts to involve the individual in decisions about their own safety and communication supports.

GPS, Technology, and Environmental Safeguards: Used, Not Relied On

Families often ask SLPs whether a tracker will solve an elopement risk. The honest answer is that devices buy time; they do not communicate for the person, and they fail at predictable moments. The Centers for Disease Control and Prevention frames personal tracking technology as one element of layered safety measures alongside supervision and prompt emergency response, not a stand-alone solution. Use the comparison below to help families weigh options, then revisit every choice on a set schedule as skills and independence grow.

Questions to Ask Yourself

Related Articles