Crash Course in Behavior Analytic Toilet Training: Translating Research to Practice

Why Toilet Training Earns a Spot on the Treatment Plan

Autistic people are more likely than their peers to deal with incontinence (no bladder or bowel control). That can last into the adult years. The research base also shows a second point. Many autistic children do not pick up toileting the way other kids do at home. The loose methods parents use with siblings often fall flat. These children need systematic teaching (planned, step by step instruction). Not all of them do. But enough do that teams should plan for it.

The social case is strong. Accidents can keep a child out of inclusive settings. Parents also report that it leaves their child cut off from peers. Age raises the stakes. The older a person gets without this skill, the more open they are to abuse. Families carry a financial load too, from diapers, pull-ups, and wipes. That load grows when a child outgrows the usual sizes. Yet parents often report that they put off toilet training. It feels like a huge thing to take on.

Before you start, check one prerequisite. The child should tolerate sitting on the toilet for about five minutes. No void is needed, and a reinforcer (something the child will work for) is fine. What you do not want is challenging behavior (crying, hitting, or refusing) paired with the toilet.

If they can't sit on the toilet without challenging behavior, we don't really want to target toilet training quite yet.

From the talk — Dr. Jessica Osos

What Johnson et al. (2021) Found About Dose and Timeline

Johnson et al. (2021) ran a systematic review (a structured look at many studies). They pulled in 55 articles and 219 participants with intellectual and developmental disabilities. Their focus was intensity, not just which steps got used. They looked at dose, how often, and how long.

Some numbers are worth keeping in your head. Rapid training averaged about 6.7 sits per session. Sessions averaged about 7.3 hours. The mean inter-sit interval (time between toilet sits) was about 56 minutes. To get an effect, studies averaged about 373 hours of treatment. That worked out to about 245 trips to the bathroom.

Translate that for a team. In a seven hour school day, 373 hours is about 45 to 46 days of active work. That is a big ask for a family to sign on to. The review also flagged a real risk. Slow progress leads people to quit too soon.

Two features showed up in the shortest treatments. One was a progressive sit schedule that starts dense and stretches out. The other was running the plan at least five days a week. The authors wrote that the ideal is every day and across settings. No study compared settings head to head, so read that as guidance and not proof.

And this is why setting realistic resource and timeline expectations upfront matters for parents too.

From the talk — Dr. Jessica Osos

The Parts That Make Up Most Toilet Training Packages

The same review counted which parts show up most. Diaper removal led with 41 studies. Punishment came next with 33. Fluid increase (more drinks early in training) showed up in 31. Differential reinforcement for staying dry (a reward for dry pants) showed up in 26. Systematic instruction, which usually means a sit schedule, showed up in 22. A urine alarm (a sensor that detects wet) and request training (teaching the child to ask) each showed up in 16.

Underwear is the heavy hitter. Two ideas may explain why. First, it is the one part you never fade out. Second, wet underwear is hard to miss, and wet pants feel unpleasant. That gives the accident a small natural cost. Dr. Osos advises against underwear worn under a pull-up during active training. For desensitization (getting used to a new item) she has no problem with it.

The urine alarm is simpler than it looks. A small pack holds the alarm, and a cord runs to a sensor. The sensor pinches the outside of the underwear fabric. It never touches skin. When it senses wet, the pack beeps and buzzes. The pack clips to a collar or a pocket. It does two jobs. It tells the adult that an accident just happened. The sound can also stop the void, so the child can finish on the toilet. Some children need to get used to wearing it first.

Studies have shown that just simply changing the type of undergarment can speed up the toilet training process, even when everything else is remaining the same.

From the talk — Dr. Jessica Osos

Dense Sit Schedules Versus Progressive Sit Schedules

Two sit schedules dominate the research. A dense sit schedule (short sits, frequent trips) means three minutes on the toilet and 30 minutes off. Once the child hits mastery criteria (the rule that says done), teams probe a 90 minute schedule. If the child passes, the next step is fading the reward. If not, teams add small chunks of time and climb toward 90.

A progressive sit schedule (breaks grow over time) front loads toilet time on day one. Then the break between prompted sits grows by set rules. LeBlanc et al. (2005) made this popular with 12 levels. Level one can be 10 minutes on and five minutes off. So the first day means a lot of time in the bathroom. Books, songs, and toys on the potty are fair game. Just keep the strongest reward for voids in the toilet.

The design question is really about motivation. A dense schedule can empty the bladder before the child ever feels full. Then there is no reason to ask. A progressive schedule builds that full bladder feeling across the day. That is what sets up a self-initiation (the child asks on their own). One more rule: no audible timer. Use a vibrating interval timer or a watch. A beeping timer can teach the child to ask when the beep sounds.

Um, cause the end goal is, is that the child should be asking to go under the control of a full bladder.

From the talk — Dr. Jessica Osos

The Punishment Era: Azrin and Foxx Through LeBlanc et al. (2005)

Azrin and Foxx ran the first toilet training study for people with developmental disabilities. Their gear caught voids right away, both in the pants and in the bowl. The bowl signal was paired with reward, so that sound became a good thing. Off the toilet, the person sat in a chair. That taught the difference between chair and toilet. They also raised fluids, rewarded good voids, and taught dressing steps.

They also used heavy punishment. An accident brought a reprimand and restitutional overcorrection (clean and restore the whole area). It also brought an hour with no food, no drinks, and no social contact. Nine adults with intellectual disabilities got trained. Few teams would run those consequences today. But the basic parts they built are still in use.

LeBlanc et al. (2005) used a non-concurrent multiple baseline design (staggered starts across children). The package held a progressive sit schedule, rewards for voids and self-initiations, and extra fluids on day one. It also used request training, a urine alarm, and positive practice (repeat trips in wet clothes). Day one ran about seven hours in clinic. Staff modeled for two hours, then caregivers took over. Days two and three ran at home. In baseline (the plan before treatment), all three children had already failed a low intensity package. Accidents dropped and stayed low at follow-up, and social validity ratings (how acceptable people found it) stayed high.

So why move away from it? Dr. Osos was trained on this package and ran it for years. Then one child got very upset during positive practice. He pushed past her planned ignoring and told her he was all done. A 2024 article by Donnelly and Karsten looked at barriers to continence. The authors warn that punishment can raise challenging behavior tied to toileting, so they caution against starting there.

And that day I was literally like, there has to be a better protocol out there.

From the talk — Dr. Jessica Osos

Reinforcement Only: Greer et al. (2016) and Perez et al. (2020)

Greer et al. (2016) asked a cleaner question. Which parts actually carry the load? Teachers in an early childhood center ran the study with 20 children. The median age was about 26 months, and most were typically developing. Only one child was autistic. Six children got the full three part package: underwear, a dense sit schedule, and the dry reward. The rest got parts added one at a time.

The full package beat every single part. Alone, the dense sit schedule and the dry reward did little. Underwear was the strongest single part. It worked for two of four children on its own, and for four of six when added later. The authors also flagged the dense schedule. Staff had to suppress self-initiations to keep the schedule running. That is the exact problem a progressive schedule solves.

Perez et al. (2020) took that package to 13 autistic children. They used a preference assessment (a ranking of favorite items), request training, and a dry reward before each scheduled sit. All children wore underwear and followed the dense schedule with a 90 minute probe. Eleven children were trained, and 12 of the 13 who finished stayed dry through the day. Several were under four years old, which pushes back on older advice to wait.

There is a catch worth knowing. The authors noted faulty stimulus control (the wrong cue controls asking). Children got prompted to ask every 30 minutes whether they needed to go or not. Across the graphs, self-initiations stayed low or absent. Dr. Osos reads that as a sit schedule problem, not a request training problem.

We don't need to use punishment from the onset of targeting toilet training in order to see a treatment effect.

From the talk — Dr. Jessica Osos

A Standardized to Individualized Package for Autistic Preschoolers

Dr. Osos took the parts she trusted from LeBlanc et al. and Perez et al. and built one package. She calls it a standardized to individualized toilet training package. The clinic study ran with five autistic preschoolers, all four years old. The site was a university based preschool with behavior analytic services. Children attended four hours a day, five days a week. Each one worked one to one with an undergraduate or master's level instructor. Outside of bathroom steps, their day stayed normal.

The standardized phase had five parts. A multiple stimulus without replacement preference assessment (pick one, remove it, repeat) ran at the start. It ran again two hours in. The top ranked item was saved for voids in the toilet only. Items two and three paid for dry pants every 30 minutes. Children wore underwear and got request training. The progressive sit schedule carried the rest. After an accident, staff prompted a sit and changed the child with little attention. A void finished in the toilet still counted as a success.

Now the part that makes behavior analysts twitch. Sit levels moved on time, not on data. Session one climbed one level per hour. Sessions two and three held one level each. From session four on, each level lasted two days. Mastery meant 80% or more successful voids at level 13 across five sessions in a row.

Three of the five children mastered the standardized package alone. Two stalled out. For them, the team restarted the schedule at level one. Progress then depended on hitting 80% across two sessions before moving up. That fixed it for one child. The last child needed two added parts: a flat 10 minute sit and a real push on fluids. Both targeted one hunch, that he was too tense to release on the toilet. The average was 30.4 sessions to mastery. In Perez et al., mastery on the dense schedule took 139.1 sessions.

So that is like one key difference from my study to past studies is that we faded to this level 13.

From the talk — Dr. Jessica Osos

Data Sheets, Treatment Integrity, Caregivers, and Fading

The data sheet is simple on purpose. Each row is one void or one dry check. Columns cover time, sit level, wet or dry check, urination, bowel movement, and notes. Codes are short. A is an accident. S is a success. I is a self-initiation. A hash mark means the child sat and nothing happened. Train staff to log one more thing. How long did the child sit before the void? A void at minute two says the toilet is pulling that response. A void at minute nine does not. A child can also get off the toilet once the void is done.

Thirty-six different instructors ran this protocol in the clinic study. Treatment integrity (staff run it as written) and IOA (two observers agree on data) both held up. Oversight covered about a third of sessions, for roughly an hour each. That is a useful signal for supervisors. Brief training plus clear written steps can carry a package like this.

The next study moved the package into homes. Four caregiver and child dyads (one caregiver, one child) finished. The children were two to three years old. One caregiver was a grandparent who watched the child during work hours. Training ran about 30 to 45 minutes before baseline and before each new phase. Caregivers also got written protocols and instructions for each phase. For readiness, Dr. Osos uses a 2019 list of 12 signs built for typically developing children. She asks for six of the 12, not all of them, so the skill is not gatekept.

Assent (the child agrees to take part) got built in. Some children said potty later when prompted. Caregivers accepted that and re-prompted five minutes later. Three re-prompts was the most any child needed. Fading needs the same care. In the clinic study, all parts stayed in place for the first week except dry checks. By the four week check, only underwear and praise were left. Four children held at or near 100%. One did not. He fell to 50% at three weeks and 0% at six, and challenging behavior rose. The reward had been holding the whole chain together, not just the void.

No protocol, however, evidence-based you think it is overrides family context and like the child's ascent.

From the talk — Dr. Jessica Osos