A veteran behavior analyst walks through what quality autism therapy actually demands, and the points where rapid, investor-backed growth quietly erodes it.
Key Takeaways
- Quality is a sequence, not a setting. Good ABA runs on individualized protocols, hands-on training, steady observation, and frequent data review, each step depending on the one before it. Skip a step and the therapy a child receives drifts without anyone deciding it should.
- Thin staffing is where it goes wrong. When a provider runs too few behavior analysts across too many cases, the training and monitoring that hold quality together are the first things to give. That, more than any single bad actor, is what produces cookie-cutter care.
- The workforce is green. The typical board certified analyst has held the credential about three years, and open jobs outnumber working analysts by roughly three to two. New graduates land in supervisory roles they have not yet grown into.
- More hours is not more therapy. Intensity is meant to be set child by child, not defaulted to 40 hours for anyone with a diagnosis. To Green, the remedy is unglamorous: smaller caseloads, real supervision, and progress measured one client at a time.
Before a new therapist is allowed near a child, the good programs make them practice on a pretend one. A colleague plays the client. The trainee runs the protocol, the written, step-by-step plan built for a particular child, while a behavior analyst watches and corrects, again and again, until the procedure is right. Only then does the trainee work with the real child. It is slow, unbillable, and, in Dr. Gina Green’s telling, exactly the sort of thing that goes missing when a company is growing too fast to pay for it.
Green has watched the field from close range for four decades, more than ten years of it at the New England Center for Children, six on the Behavior Analyst Certification Board, and a term as president of the California Association for Behavior Analysis, where she now consults on public policy. (She helped write the field’s current practice guidelines, a role disclosed here because it shapes her sense of what counts as adequate care.) The word everyone reaches for to describe what has curdled as investor money poured into autism therapy is “cookie-cutter.” Green treats it less as an insult than as a diagnosis, one with specific and locatable failures.
She is careful with terms, an occupational reflex. “We use terms like treatment fidelity or procedural integrity to refer to the accuracy with which interventionists implement treatment protocols,” she said in an interview with Acuity. “What we’re talking about more here is clinical quality: ensuring the quality of the services is acceptable.” The two are bound together, because quality starts with the protocol and then lives or dies on whether the person delivering it sticks to the plan.
The work behind a good session
A protocol is only as good as its design: built for the specific child, drawn from research, and clear enough that someone else can follow it, down to how each behavior is counted and recorded. A vague or generic plan hobbles everything that comes after. Yet even a strong plan is inert on paper. “You can’t just hand somebody a written set of instructions and say, now go do this with this client,” Green said. “The protocol should be individualized to each client, and the behavior analyst has to train each person who’s going to be implementing it.” That is the role-play, and it reaches the whole delivery team, from registered behavior technicians and assistant analysts to the parents expected to carry procedures into the kitchen and the car.
Then comes the part that never ends. People stop following instructions, not out of bad faith but out of fatigue and habit. “You get fatigue. People drift,” Green said. “They start not following the rules, kind of making up their own, changing the procedures on their own. So the behavior analyst should be regularly watching people implement the protocols and recording data on whether they’re actually performing each of the steps the way they were intended.” Observe, record, retrain, adjust. The loop is what keeps a program aimed at the child in front of it, and it is the single most time-consuming thing a behavior analyst does.
What growth does to the math
Investor-backed expansion presses on that loop from above. “When investors purchase programs or agencies, they want to maximize return on investment,” Green said. “One way they do that is by cutting costs, and one way to cut costs is to have as few people at the professional behavior analyst level as possible and lots of people at the technician level.” Fewer analysts, more technicians, more billable hours. On a spreadsheet it reads as efficiency. On a caseload it reads as an analyst who cannot keep up, a ratio buyers and clinicians now weigh as a diligence risk.
“You end up with the professional behavior analysts having caseloads that are too large, so many that they can’t really effectively develop individualized protocols, train people to implement them, make sure they’re implementing them correctly, and review the data on a regular basis,” Green said. The independent research she points to attaches numbers to the complaint. An analysis published by the Center for Economic and Policy Research found supervisory ratios the field pegs at roughly one analyst for every 10 to 15 clients stretched, at some acquired agencies, to one for 25 or even 40. Workforce planners generally treat eight clients per analyst as reasonable.
The staffing squeeze compounds a demographic one. The typical board certified analyst has held the credential only about three years. In 2025, employers posted more than 132,000 jobs for a workforce of roughly 84,000 active analysts, only about 30,000 of them in direct clinical care, while first-time exam pass rates slid from 66 percent in 2020 to 51 percent. Into that supply gap step new graduates, handed the hardest work in the field. “If you’re right out of a master’s program, you’ve just attained your board certification, that’s it in terms of your practical experience,” Green said. “If you’re now responsible for conceptualizing cases and developing protocols for a lot of clients, some of them may not be the kinds of clients you’ve had experience with. That would be a tall order for quite an experienced behavior analyst, much less one fairly new to the field.”
Above the analyst, often, sits someone who has never run a session. “They may put their own people in clinical and administrative decision-making roles, and they’re not behavior analysts,” Green said of some corporate owners. “They may have no experience with people with autism. So you get this tension where well-meaning, well-trained behavior analysts answer to people who don’t have that background and may not have realistic expectations about what the behavior analyst does.” A young clinician reporting to an executive counting billable hours is not well placed to say no, a dynamic visible across the largest multi-state chains now reshaping the field.
The forty-hour question
If cookie-cutter care has a signature, Green says, it is the reflexive prescription of maximum hours. Weekly hours stand in for intensity, and intensity is supposed to be a clinical call. Many children, especially young ones, do benefit from intensive intervention, and for an older client who is self-injuring or bolting into traffic it can be a matter of safety. For others, ten focused hours accomplish more than forty scattered ones. “It is not at all the case that every person with autism should be getting a high number of hours. That has to be individualized,” she said. “One of the misapplications with some of these big corporate entities is that they just prescribe a high number of hours for every person. That’s not what our professional standards require, and it’s not necessarily supported by research for everybody.”
Put a four-year-old on a 40-hour plan without first assessing what the child needs, then, and the number is likelier a business decision than a clinical one. “It all feeds into that emphasis on billable hours and profit margins instead of evidence-based practice,” Green said. That is the pattern now drawing payer and Medicaid scrutiny, though she notes those reviews tend to pore over claims data rather than the clinical variables that decide whether a child actually improves.
Families are in no position to referee any of this. Most parents cannot rank one provider against another, and many have little choice to begin with, boxed in by which agencies their plan contracts with and what it will authorize. “Parents may not be well informed,” Green said. “One of the things a lot of us have tried to do over the years is educate parents and payers about what good, genuine ABA looks like and what it requires.” Her own fix is unglamorous and structural: the better agencies build systems, the standing policies that keep caseloads reasonable, keep technicians trained, keep someone reading the data every week. Voluntary accreditation exists to check for precisely that. As the field’s billing and coding fights grind on, none of it substitutes for the thing hardest to scale and easiest to cut: a well-trained analyst with enough time to watch one child, then the next, and catch the moment something stops working.
