Most families we work with in Miami-Dade have a version of the same story. A door was unlocked for ninety seconds. A gate latch was left open after a delivery. A child who had never opened the sliding door before opened it. The story usually ends with the child found quickly and everyone shaken, and it is almost always followed by the same question: why does she keep doing this, and how do we make her stop?
That question is worth reframing, because the framing changes what you do next. Leaving a supervised space — clinicians call it elopement, families usually call it wandering or bolting — is a behavior. Behaviors happen for reasons. A child who leaves is going toward something or away from something, and until you know which, environmental changes are guesswork and consequences are beside the point.
This article covers what a functional assessment actually identifies, the environmental changes that do the most work in a Miami-Dade home, why adaptive swim instruction matters here specifically, and how to write a one-page safety plan you can hand to a first responder. Nothing here is a guarantee. Layered prevention reduces risk; it does not eliminate it, and no therapy of any kind should be described as drowning prevention.
How common elopement is, and what the research actually says
The most-cited prevalence figure comes from a 2012 study in Pediatrics by Anderson and colleagues, who surveyed families of 1,218 children with autism spectrum disorder. Forty-nine percent reported their child had attempted to elope at least once after age four. Twenty-six percent had been missing long enough to cause concern. Among the children who went missing, 24 percent were reported to have been in danger of drowning and 65 percent in danger of traffic injury. The same study found elopement risk rose with autism severity, increasing roughly nine percent for every ten-point increase in Social Responsiveness Scale score.
You will also encounter a widely repeated claim that autistic children are "160 times more likely" to drown. We do not use that number, because we could not source it to the study it is usually attributed to. What the primary literature does report is different and still serious: Guan and Li, writing in the American Journal of Public Health in 2017, examined national death records from 1999 to 2014 and found 1,367 deaths among individuals with an autism diagnosis, of which 381 — 27.9 percent — were injury deaths. Drowning accounted for 74 of them, with a proportionate mortality ratio of 39.89. A proportionate mortality ratio is not a risk multiplier; it describes how over-represented drowning is among these deaths compared with the general population. It is a real and large signal, and it deserves to be quoted accurately.
A second 2017 paper by the same authors, in Injury Epidemiology, looked at the circumstances of 23 fatal drownings involving autistic children under 15, assembled from news reports between 2000 and 2017. Wandering was the reported activity preceding 73.9 percent of them. Just over half occurred in ponds. The average distance between the child's home and the drowning site was about 291 meters. Roughly three-quarters happened between noon and 7 p.m. That study is small and drawn from media accounts rather than a registry, so it should be read as descriptive rather than definitive — but the pattern it describes is the one Miami-Dade parents should recognize: close to home, in the afternoon, in standing water.
The practical takeaway is not that you should be more frightened. It is that the water nearest your door is the water that matters most, and that the hours after school and before dinner are when supervision most often lapses.
Why Miami-Dade changes the arithmetic
Most of the national guidance on wandering was written for places with a swimming season. Miami-Dade does not have one. Water here is a year-round, ambient feature of the built environment, and the density of it is unusual even by Florida standards. The county is threaded with drainage canals, many of them unfenced and running directly behind residential blocks. Retention ponds sit inside apartment complexes and shopping centers. Backyard and community pools are common enough that a child who leaves a home in Kendall, Doral, or Hialeah may pass several before anyone notices.
That density interacts badly with the finding that most incidents happen close to home. A 291-meter radius in a suburb with one distant lake is a manageable search area. The same radius in much of Miami-Dade can contain a canal bank, two community pools, and a retention pond. The question "where would my child go if he left right now" has more answers here than it does in most of the country.
It also means the seasonal framing many families use — being vigilant in summer, relaxing in winter — does not fit. Neither does the assumption that a child who cannot reach a pool at home is safe. We ask families to walk the actual route: leave your front door, and for ten minutes, walk in the direction your child tends to head. Write down every body of water you pass, including the ones behind fences that a determined child could get around. That list is the real risk map, and it is usually longer than parents expect.
What a functional assessment identifies
A functional behavior assessment is the process a BCBA uses to work out what a behavior accomplishes for the person doing it. For elopement, the answers cluster into a few recognizable patterns, and each one points toward a different intervention.
Some children leave to get to something. A child who consistently heads for the same neighbor's pool, the same canal bank, or the same playground is not escaping — she is traveling toward a preferred destination, and water is often the destination itself, because it is sensorily compelling in a way few other things are. Some children leave to get away from something: noise, a transition they did not expect, a demand, a crowded room. Some leave because leaving itself produces an immediate and reliable reaction from adults, which for a child with limited communication can be a very efficient way to get attention. And some leave because the movement is the point — the running, the door, the latch mechanism.
These distinctions matter because they change the plan. A child eloping toward water needs a different response than a child eloping away from noise. The first needs supervised, scheduled access to water so that leaving is not the only route to it, plus physical barriers on the path. The second needs the antecedent addressed — the transition warned, the noise reduced, the demand adjusted — because barriers alone will not resolve a child's need to escape something aversive. A plan built on the wrong function tends to fail in a specific way: the behavior stops at one door and reappears at another.
A functional assessment is also where communication comes in. If a child has no efficient way to request access to the pool, or to signal that a room has become intolerable, elopement may be the most reliable communication available. Building a functional alternative — a picture exchange, a sign, a device request, a word — is slower than installing a lock, and it is the part that generalizes. Our article on what happens during an ABA session describes how that teaching looks in practice.
Changing the home environment
Environmental change is the part of the plan that works immediately and requires nothing of the child. It should be done first, in parallel with assessment, not after it.
On doors: the standard recommendation is a secondary lock placed out of the child's reach, above the primary lock, on every exterior door. Add an audible alert — a chime, a door sensor, an inexpensive contact alarm — so that a door opening produces a sound in the room where the adults are, not just at the door. Sliding doors and windows need the same treatment; families frequently secure the front door thoroughly and leave a lanai slider on a single latch. If your home has a pool, a compliant four-sided isolation fence with a self-closing, self-latching gate separates the pool from the house rather than only from the street, which is the configuration that matters for a child leaving from inside.
On the perimeter: check gate latches monthly rather than assuming they hold, and check them after any service visit. Lawn crews, deliveries, and pool technicians open gates. A latch that a child cannot work at four may be workable at six. If a canal or retention pond sits behind your property, the fence between them is the single most important structure on the lot, and it should be inspected like one.
On neighbors: this is the intervention families skip and later wish they had not. Tell the immediate neighbors, and specifically tell the ones with pools, that your child may enter without warning and does not reliably respond to their name. Ask them to call you first and to check their pool first. In a county where the average incident happens within a few hundred meters of home, the people who live inside that radius are functionally part of your safety system.
Adaptive swim instruction
Swim instruction is not drowning prevention and should never be sold as such. A child who can swim can still drown, and no lesson removes the need for barriers and supervision. What swim competence does is change the outcome of a specific, common scenario: a child reaches water unsupervised and enters it. Skills relevant to that scenario — rolling to a back float, orienting to the nearest edge, getting to and holding the wall, exiting without a ladder — are the ones worth prioritizing over stroke technique.
Adaptive instruction differs from a standard group lesson in ways that matter for autistic children. The instructor works one-to-one. Instructions are short and often paired with visual models or gestures rather than delivered verbally at length. The lesson has a predictable structure and a visible ending, so that the child knows what is coming. Sensory factors are treated as real variables rather than complaints: water temperature, echo in an indoor pool, the smell of chlorine, goggles, the texture of a swim shirt. Progress is measured in repeatable skills rather than distance.
When families ask us how to evaluate a program, we suggest asking the instructor three things. How do you teach a child who does not follow verbal instructions? What do you do when a child becomes dysregulated mid-lesson — end, continue, or shorten? And how do you measure progress? An instructor who answers all three concretely is usually a good fit; one who answers only in terms of strokes and levels is usually not.
For children already receiving behavior-analytic services, the two can be coordinated. Water tolerance, following instructions in a loud and echoing environment, and staying with an adult in an open space are all skills that can be targeted on land first, where teaching is easier and safer, and then brought to the pool. That coordination is ordinary clinical practice — it is not a claim that therapy prevents drowning.
Identification, tracking, and what to hand a first responder
Identification serves one purpose: shortening the interval between a stranger finding your child and that stranger reaching you. Options range from an engraved bracelet or shoe tag to an iron-on label inside a collar to a temporary tattoo for outings. What matters is that it is worn reliably and that it says the minimum useful thing — a first name, "autistic, may not respond to name, non-speaking" if applicable, and a phone number. Ours is (786) 671-1601 if you want a template for how little text it takes.
Tracking devices are worth understanding before buying. Consumer Bluetooth trackers only report a location when another phone in the network passes near them; in a sparsely trafficked area, or beside a canal, that may be no help at all. GPS devices report continuously but need charging and a subscription. Some Florida sheriff's offices operate radio-frequency locator programs for residents with autism or dementia, which work independently of phone networks — it is worth calling your local office and asking what they run, because these programs are inconsistently publicized.
The single highest-value item is a one-page document prepared before it is needed. Put on it: a current photograph taken within the last three months, height and weight, what the child was likely wearing, whether the child is speaking or non-speaking, whether the child responds to their name, and how the child responds to strangers, sirens, and shouting. Then add the part most families omit — a list of the specific water features within a ten-minute walk, in the order a searcher should check them, and any place the child is known to be drawn to.
That last section is the one that saves time. Search protocols for missing children with autism generally prioritize water first, and a parent who can hand over an ordered list of nearby water rather than describing it under stress has materially improved the search. Keep the sheet somewhere retrievable in seconds — on the refrigerator and in your phone — and update the photograph seasonally.
Writing the plan down, and who else needs a copy
A safety plan that lives in a parent's head is not a plan, because the moment it is needed is the moment recall is worst. Write it down. One page. Who checks which exit, who calls 911, who walks which route, who stays home in case the child returns, and where the first-responder sheet is kept.
Then distribute it. The people who need a copy are the ones who supervise your child when you do not: the other parent, grandparents, a babysitter, the after-school program, the bus driver if there is one, and the school. If your child receives ABA services, the clinical team should have it too, because elopement targets belong in the treatment plan and the plan should reflect what the family is actually doing at home. Our article on the role of parent training in ABA therapy covers why that transfer between clinic and home is usually the deciding factor.
Review the plan when anything structural changes — a move, a new pool, a new school, a new skill like opening a deadbolt. Children acquire the ability to defeat barriers, and a plan written for a four-year-old is often obsolete for the same child at six. Summer is a reasonable default review point in most of the country; in Miami-Dade, where there is no off-season, tie it to something else — the start of the school year, or a birthday.
Finally, be realistic with yourself about supervision. The Anderson study found that 62 percent of parents of children who eloped said concerns about wandering had prevented them from enjoying or attending activities outside the home. That cost is real, and a written plan with layered barriers is partly a way of buying some of that life back. The goal is not permanent vigilance, which no one can sustain. It is a set of systems that keep working during the ninety seconds when vigilance fails. Our guide to summer without structure deals with the related problem of keeping routines intact when school is out.
