Rad N Bad: 30–40 Hours of ABA — Dosage, Titration, and the Data Behind Service Intensity

Why 30 to 40 Hours Became the Default Recommendation

The hosts do not trash the model, and they give it credit first. The 30 to 40 hour recommendation did not appear from thin air. They point to studies showing that more hours can lead to more progress, especially in young learners. Repetition, structure, and intensity all matter. They matter most when you are building early skills like communication, behavior regulation, and adaptive functioning.

Sean compares it to physical therapy. You tear something, you go in for therapy, and you also do the exercises at home. The dose is intense on purpose. He says he would expect results from that much service, and he would demand them.

Mike adds the school comparison, since young kids already spend 30 plus hours a week in a classroom. Learning is a full-time job when you are little. So a high weekly dosage (how many hours per week) is not strange on its face.

Then comes the caveat that carries the rest of the episode. The model works for some kids, and only when the hours get used well. An authorization (the approved block of hours) can list 30 hours. That does not mean all 30 hours are quality service. The hosts spend the episode on that gap.

so the whole point of this episode is just because something is standard doesn't mean it's right for everyone

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

The Funding Model Pays for Hours, Not Outcomes

Here is the part the hosts say out loud. The payment system rewards quantity. Insurance pays on a fee for service model (paid per hour delivered), and it pays for hours rather than results. That shapes what gets recommended, even when everyone means well.

One host owns an ABA company and said so on air. He called the financial pressure real, and he said he is tired of pretending it does not color recommendations. He still believes most BCBAs are grinding ethically and writing honest hour requests. His point is about the system, not about individual bad actors.

His ask is simple. A field that calls itself science-based should be willing to audit its own incentives. If hours pay the bills, then hours deserve extra scrutiny from the people writing them.

The hosts also flip the logic in a way you can use tomorrow. Authorized hours are not a mandate. Mike's example is a child moving into school. The hours stay available for half days or new needs. But they do not get staffed just because they exist.

Right now, insurance reimburses based on hours, not outcomes.

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

How Assessment Checklists Manufacture Years of Service

Both hosts describe the same habit. You run a standardized assessment, then you write objective 1a, 1b, 1c, and keep going down the alphabet. Then you move to domain two, three, and four. Each new row justifies more hours across more authorizations. Mike says he has met families who have been in ABA for five, six, seven, or eight years.

The fix starts with how you treat the assessment. A tool like the VB-MAPP (a common skills assessment) is a reference point, not a curriculum and not a schedule. Life does not move in a straight line from milestone one to milestone two. Mike pulls from several sources at once. He uses VB-MAPP items, ESDM (a play-based developmental model) content, and EFL (a functional life-skills checklist) goals. He also asks a blunt question of the field: can you run more than one assessment at a time?

The hosts have a sharp image for what happens otherwise. Two clinicians sit next to each other and compare plans, and it feels like bingo. Do you have 1.6? Yep. Every custom plan reads the same, and goals get added to fill a schedule instead of to change a life. Their standard is the reverse order: program for the person, then pick the hours.

if we are truly focusing on what matters and programming for the individual not the assessment then we wouldn't have this problem

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

Titration Plans That Track Progress Instead of Percentages

Titration (stepping service hours down slowly) sits in most treatment plans already. The hosts say it is usually decoration. The classic version reads like this: when the client hits 75% of goals, reduce by two hours. They want to know where that number came from. They list the usual suspects of 40, 35, 30, 25, and 20. Mike also calls out the hour cliff, where hitting 15 hours means goodbye and graduation.

Their version starts much earlier. If you call the service intensive, plan the exit at the start. Use the hours hard while you have them. Name the point where response to treatment says you can come down, inside this authorization or by the next one. Fading is not kicking a family to the curb. It gets paired with generalization programming.

What replaces direct hours is usually teaching. If progress is real, titrate and shift toward caregiver training. The hosts make the credential argument plainly. A registered behavior technician (frontline therapy staff) can be trained with a high school diploma. So families can be taught those same skills too. The hosts also push hybrid models, with direct therapy plus school support plus caregiver coaching plus respite. Anyone who is a pillar in that child's life belongs on the team.

we need to actually use titration plans for what they're actually supposed to be used for and that's to reflect progress

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

Stop Using Fidelity as a Blanket Justification for High Hours

The hosts flag a move they hear all the time. We need all these hours because only our staff can run the program with fidelity (doing it exactly as written). That claim rests on two kinds of data. The first is procedural integrity (staff ran steps correctly). The second is interobserver agreement, or IOA (two observers scoring alike).

So they ask for it. Show the integrity data, show the IOA, and show that you know how to calculate it. One host guesses that most clinicians billing private insurance cannot produce either number. Without them, the claim that your intervention caused the progress is just a signature on a session note.

Their proposal goes past internal practice. Put integrity and IOA data in your reports and send them to funders. If you have nothing to hide, outside eyes are not a threat. They would also like someone at the funder who can actually read a graph. Peer reviews with a BCBA do exist, but the hosts say they are rare.

Show me that the data are actually valid, reliable, and accurate.

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

Trend, Stability, and the Three-Data-Point Myth

The data skills segment is where this episode earns its credit. Mike makes a point he likes to teach. Show one line graph to 51 people and you will get 51 interpretations. Simple line graphs and bar graphs are easy to read, and that is why we use them. Readable is not the same as objective.

So looking is not the whole job. Add numerical values and run a real analysis. Check two things before you touch the program. The first is trend (a steady direction over time). The second is stability (how tightly the points sit). Three to five points heading one way do not guarantee a trend. The percentages underneath may not be stable at all.

Mike's on-air quiz is the practical takeaway. Can you say what a stability envelope (a band around the data) is? Can you say how many points you need, and how to calculate the envelope? If not, that is the skill to go learn. The cost of skipping it is concrete. We fade prompts or raise criteria because the graph looked ready. Then a child melts down because they were not.

can you tell me what a stability envelope is and how to determine what or what the trend is

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

Incomplete Trial Recording Inflates Apparent Mastery

Mike's second data warning is about recording. He describes opening a goal that sits at 100%, with three points across the top of the graph. Then he looks closer. Those three were the trials the learner got right, and the 15 wrong ones never made it into the data.

He does not read that as cheating. He reads it as reinforcement (a reward that strengthens behavior) working on the adult. Staff see progress, celebrate the good trial, and forget the hard ones. The fix is accountability rather than blame: the same definitions, the same teaching procedures, and the same training for everyone who records data.

This loops straight back to integrity and IOA. If my definition and your definition differ, our numbers differ too. When the numbers differ, the data are less reliable than we believe. That barely increasing trend may not exist. Mastery can be an artifact of what got written down.

They offer one more measurement idea, this time aimed at the hours themselves. Take partial interval recording (scoring short time blocks) on staff and clients during sessions. Then you can say how much of those 30 to 40 hours is actual therapy.

sometimes when our people see that they've made progress they're like oh i got the good trial and they forget to record the not good trials

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA

A Checklist for Data-Based Dosage and Titration Decisions

Pull the episode together and you get a checklist you can run on your own caseload. Name the goals that require this dose, in writing. Show the procedural integrity and IOA data behind the claim. Confirm trend and stability before any programming or hour change. Write the titration trigger in progress terms rather than percentages. Name the caregiver-training step that replaces direct hours. Program generalization from the start, not after a box gets checked.

One more item comes from a story. The only peer review one host ever triggered was for underutilization, which means using fewer hours than approved. The family had real reasons, including transportation and other children at home. A family can qualify for 40 hours and only commit to 25. The funder said the reasons were not good enough. So document the reasons for your utilization as carefully as you document the need. The system can punish you in both directions.

The closing frame is the one worth keeping. There are 168 hours in a week, so even full-time clinic attendance is under 25% of a child's life. The hosts point back to Baer, Wolf, and Risley in 1968 and all seven dimensions (seven marks of real ABA). Their bar for the 30 to 40 hour model is that it be earned by data, then sustained by quality practice.

the gold standard is the individualization of how many hours that kid needs to make clinical progress

From the talk — Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA