Minutes after an autistic 3-year-old lay her head on a Squishmallow stuffed toy and closed her eyes, Abigayle Hosking was ordered to take the makeshift pillow away and rouse the sleepy girl. The child was scheduled for four more hours of autism therapy at the Centria Healthcare clinic in Michigan where Hosking worked until last year.

Hosking said she felt terrible. “The younger kids, you know, I think they need a break,” she said.

Medicaid plans and private insurers will often pay for up to about 40 hours of autism therapy a week, and shut-eye doesn’t count. At Centria and some other large chains backed by private-equity investors, former employees said in interviews that they felt pressure to maximize therapy hours, claims the companies dispute.

The Wall Street Journal analyzed the billings of five of the largest private equity-backed providers of autism therapy to Medicaid patients, including Centria. Together, the group’s average weekly therapy hours for those patients were 29% higher than competitors’ in 2024. All but one of those companies delivered more hours than local rivals.

“It’s simple business math,” said Philip Painter, a physician who briefly served as chief medical officer at Centria. “You want to maximize the number of hours the child is doing.”Medicaid programs financed by state and federal taxpayers paid $10 billion for autism therapy last year, up from around $2 billion in 2021, driven in part by increasing numbers of treatment hours per patient.

In autism therapy, treatment hours are supposed to be tailored to each patient’s needs, but there is debate about how much is warranted.

Many autism-therapy providers and an industry group cite studies of young children showing high-hour therapy can help those with severe challenges. Government officials and insurers have scrutinized long hours, at times pointing to other research, disputed by some providers, that raised questions about the correlation between long therapy hours and improved outcomes. New federal guidelines for Medicaid say there is no correlation.

Private-equity firms have poured hundreds of millions of dollars into building national chains of clinics to serve children diagnosed with autism. The Journal’s autism-treatment analysis looked at Centria, majority-owned by THL Partners; Action Behavior Centers, backed by Charlesbank Capital Partners; Hopebridge, acquired by Arsenal Capital Partners; Cerberus Capital Management-backed Lighthouse Autism Center; and Learn Behavioral, majority-owned by Gryphon Investors, which billed fewer hours a week, on average, than its competitors.

All five companies said they individualize treatment, including weekly hours, based on patient needs, not financial considerations.

The four companies with higher-than-average hours said the Journal’s analysis of Medicaid claims didn’t account for differences in the severity of patients’ autism, and that their nap policies aim to accommodate children’s sleep needs. Centria, Action and Lighthouse said their patients had more severe autism than average, justifying more therapy.

Centria said waking a 3-year-old as described by Hosking wouldn’t have complied with its policies, and such an incident wasn’t documented at the time. Action disputed the assertion in the Medicaid guidelines that treatment intensity doesn’t correlate with outcomes.

Former employees of the four higher-hour companies said in interviews they felt pressure to recommend long therapy schedules for children, and that centers sometimes turned away children who needed low weekly hours.

The former employees said the companies applied the same approach to both Medicaid and privately insured patients.

In Arizona, Centria billed 94% more hours a week per patient, on average, than competitors in 2024, and in Colorado, Action billed 81% more than local rivals, the Medicaid data show.

LaDeana Thompkins’s son was scheduled for about 40 hours a week at an Action facility near Phoenix when she told the clinic in July that his hours would need to be cut by about half because he was starting kindergarten, according to emails reviewed by the Journal.

In a meeting the next day, an Action staffer told Thompkins and her husband that their son wasn’t ready for kindergarten, and proposed that the boy, then 5, could delay going to school and continue full-time therapy, according to Thompkins and her husband, as well as documents from the time.

The couple decided they didn’t want to hold off on kindergarten, where their son could get help from an aide and other support at their local school. The Action staffer told the couple that the clinic could accommodate the boy only if he was attending close to its recommended therapy hours, and that if he left and wanted to return, he might land on a waiting list, the Thompkinses said. “It felt like we just fell off a cliff,” LaDeana Thompkins said.

Action’s chief clinical officer, Charna Mintz, said in an interview that if a family chooses school, “even if it is not what is clinically recommended,” the company would support the family in the transition. In a letter, lawyers for Action told the Journal the company had “information that would thoroughly debunk the narrative this patient is being used to support,” but couldn’t share it due to a healthcare privacy law.

The big therapy chains focus on applied behavior analysis, or ABA, commonly given in clinics or at home to children as young as 2. Sessions can include practicing chores such as folding laundry, brushing teeth and working on communication skills. In ABA, common activities such as snacking can become therapeutic tasks, as workers help children learn how to request the food they want. Major medical groups recommend the therapy as an option for autism treatment.

rapy is delivered mostly by behavior technicians, who work one-on-one with patients. More highly trained behavior analysts, typically with master’s degrees, oversee the technicians.

Before starting therapy, each child is supposed to be assessed by a behavior analyst who develops an individualized treatment plan. The analyst’s recommendations, along with the family’s preferences, are supposed to determine weekly hours, and an insurer generally must authorize services.

“The hours should be given and directed based on the needs of the child and then the family,” said Whitney Ence, a psychologist who oversees an autism program at UCSF Health.

The four higher-hour companies said their behavior analysts make their own judgments about what treatment to recommend.

Former employees of Lighthouse said the company typically sought to enroll kids in full-time or half-time slots, with a preference for full-time. “There are so many families and kids I would have to turn away because they didn’t fit inside of our model,” said Macie Wallace, a clinical director at a Lighthouse center in Illinois until earlier this year.

Lighthouse said its therapy centers primarily serve patients with acute autism, that it has no minimum required level of therapy, and that it refers about 10% of prospective patients—generally ones with lesser needs—elsewhere. It said its focus on therapy at centers rather than in homes could help explain its higher hours, though the data show its hours to be higher than is typical of similar providers.

Centria said three-quarters of its patients, by its own measure, fell into higher tiers of autism severity. Hopebridge and Action noted they focus on young children, who typically receive more intensive treatment. Action said its focus on acute cases makes the Journal’s comparison “not apples-to-apples.”The Journal’s analysis showed that even when treating the same children as their rivals, the four higher-hour chains routinely delivered more hours of therapy.

Lighthouse said the difference might reflect patients’ changing needs, while Centria and Hopebridge said comparisons of billed hours don’t account for quality or appropriateness of care.

Action said its model enabled it to deliver the hours recommended in patients’ treatment plans, while competing providers might not have resources to do so. Action also said it didn’t unilaterally set treatment hours, which insurers must approve.

UnitedHealthcare, the nation’s largest insurer, removed Action from its Medicaid network in Arizona earlier this year. A UnitedHealth spokesman told the Journal that Action cost the company three times as much per patient annually as smaller competitors in that market. Action said the comparison didn’t account for the severity of its patients’ conditions, which UnitedHealth confirmed.

The Journal’s analysis relied on detailed Medicaid billing records that reporters accessed under a research agreement with the federal government. The data don’t include private insurance claims, and don’t include identifying details of patients, such as names.

Autism therapy can be lucrative, partly because labor costs for technicians are low. They often are paid around $20 an hour, while Medicaid pays the companies an average of about $60 an hour, according to the Medicaid data. Private insurers can pay closer to $90 an hour, according to a data analysis by health-tech company Turquoise.

vestor presentations last year, Centria projected a 2025 profit margin of more than 20%, before taxes and other items, on $382 million in revenue, and Action estimated a 24% profit margin by a similar measure on $1.5 billion in revenue. The letter from Action’s lawyers said its 2025 operating margin was 8.5%.

The slide deck for potential Centria investors noted 31 average weekly hours of therapy per patient at its unit for kids 18 months to 6 years old. Lighthouse investor slides that circulated in 2021, just before Cerberus invested, flagged weekly billable therapy at “30+ Hours/Client,” and noted “lifetime revenue per client” of $285,000.

The Action slides highlighted growth in “FTE patients”—an abbreviation typically used to mean “full-time equivalent.”Action said the term wasn’t used as a label for any patients. Ben Seib, a former chief financial officer at Hopebridge, said that company tracked young patients in terms of FTEs.

“Generally you’re looking at, like, just how many kids do we have, how many billed hours are we getting per week, and is it growing,” he said. The company generally sought to schedule children for either full-time or half-time therapy, which made it easier to coordinate treatment and staff scheduling, he said. Behavior analysts, he said, could recommend varying hours.

Former Hopebridge employees said clinics kept whiteboards tracking weekly metrics. A photo of one, reviewed by the Journal, including a “billed hours” row with target and actual figures. Hopebridge said it tracks such measures “to improve access to treatment.”After Hopebridge acquired Autism in Motion Clinics, a smaller autism-therapy company, in late 2021, executives told employees about new standard therapy schedules: either half-days (8:30 to 1 or 1 to 5:30), or full days (9 to 4:30), according to slides reviewed by the Journal. The slides said Hopebridge focused on children up to 9 years old.

Former employees said parents were supposed to be told if their kids didn’t match the new model. “That’s not a conversation you can really have with a family—‘You don’t fit into our box anymore,’” said Lindsay Watkins, a behavior analyst who helped run an AIM clinic in Apex, N.C. She left the company before she had any such discussions.

Hopebridge said its scheduling is meant to ensure children receive quality care, and decisions about whether patients should be treated at a center are based on clinical needs. The company said the claims data analyzed by the Journal don’t reflect patients’ needs or outcomes. Managing children who want to nap is a challenge across the industry.

A Centria handbook for technicians, reviewed by the Journal, noted that “a client who falls asleep during therapy is allowed 1/2 hour of rest without stopping the session,” but “after 30 minutes elapses, client must wake up, or technician will conclude therapy for the day.”Former employees said that often meant a child would be sent home. Centria said children could nap longer than 30 minutes when approved by the clinicians overseeing their care.

Autism therapy is often intended to help children transition into school and other settings, and experts recommend that hours be reduced for kids who are achieving treatment goals.

The Journal analysis found that children getting autism therapy from Lighthouse, Hopebridge, Action and Centria, on average, were treated for a similar number of hours a week after months and in some cases years of treatment.

In Colorado, the 224 Medicaid patients who began treatment with Action in 2022 and 2023 and received at least a year of therapy quickly ramped up to about 31 hours a week, on average, of service provided by behavior technicians, the data show. A year later, they were still averaging more than 30 hours of therapy most weeks. The average over the second year remained at about that level.

Action said the Journal’s analysis was flawed, and it provided four different analyses of its own that showed reductions of up to 17% over a year, and more over two years. One of the company’s approaches didn’t count weeks with 25 or fewer therapy hours during the first three months of treatment to account for the ramp-up, but did count those short-hour weeks in later periods.

The Journal shared its own methodology with several experts in autism therapy and health-services research, who said it was sound. David Mandell, a University of Pennsylvania epidemiologist and health-services researcher who studies autism treatment, who also reviewed Action’s description of its methodology, said the approach of counting only high-hour weeks early in treatment “is not consistent with any best practice or understanding of how we measure treatment dose.”

Callie Baird, an analyst who worked at Action from 2023 until April 2025, mostly in Illinois, said she recommended cutting hours for one of her patients so he could transition to kindergarten. He was doing well on his goals, she said, and had gotten bored with the therapy. She felt he needed more social peers. She said her supervisor resisted. “It would be like, ‘No, clearly he’s not ready,’” she said.

Eventually, the child’s mother decided she wanted his hours reduced and sent him to school, where he did well. “I was constantly having to push against the system there,” said Baird, who now runs her own therapy company.

Action said it works to reduce hours to prepare children for school and that every patient has a titration plan. The company said Baird was closely supervised because she had limited experience as a behavior analyst when she joined Action. Baird responded that at the time she had more than six years of experience in the field.

Centria said 42% of its requests to insurers to continue treatment included a reduction in hours. Hopebridge and Lighthouse also said reducing weekly hours was part of their approach.

Some of the higher-hour providers point to guidelines by industry group Council of Autism Service Providers, which say patients with significant needs can benefit from 30 to 40 hours a week of ABA treatment.

That guideline is largely based on studies beginning in the 1980s, including one that found 19 patients who started full-time therapy by age 4 outperformed a comparison group getting about 10 weekly hours when measured in later assessments.

Eric Larsson, an autism-therapy expert who worked on some of that research, said the economics of the treatment provides an incentive to keep hours high. “If it costs you money to do 10 hours a week, but you profit at 40 hours a week, even if you’re ethical, how do you balance those incentives?” said Larsson, the executive director of autism-therapy provider Lovaas Institute Midwest and a CASP board member. He said he wasn’t speaking for CASP.

Recent guidelines from the federal Centers for Medicare and Medicaid Services, which oversees Medicaid, said the number of hours should vary depending on patients’ needs. CMS said scheduling young children for 40 hours a week of ABA is “not a best practice,” and children need time for normal activities like napping and eating.

That’s how many providers already operate. During the period examined in the Journal’s data analysis, 11 Michigan Medicaid patients getting autism therapy switched from Centria to Game Changer Therapy Services, a local provider. Their average weekly therapy hours dropped 25%.

“If a client says, I want my kid to go to karate or do a basketball league, 100%, families shouldn’t have to choose between therapy and creating meaningful experiences for their kids,” said owner Scott Staszak.

Thompkins, the Arizona mother who sent her son to kindergarten, said he has switched to a new therapy provider, Cortica, where he is getting about 25 hours a week of ABA after school. His therapy is covered by the insurance she gets through her employer, a software company.

Thompkins said she believes her son benefited from the therapy at Action, which records shared by the family show he started in April after one of the company’s analysts found he needed significant support. She and her husband had always aimed for their son to enroll in kindergarten, after he earlier attended preschool.

An Action discharge report reviewed by the Journal said he was departing to begin school against its clinical recommendation.

Action said a decision by the Thompkinses not to permit the company to share the boy’s medical details called into question the reliability of the family’s account. The company said that by way of example a 5-year-old with low scores on a standardized assessment and limited communication skills, who received treatment for just a short time, “would typically require high-intensity services.”The company also pointed to a study co-written by Mintz that said 71% of Action patients it characterized as having “graduated” from therapy were placed primarily in general education at school, while those it said didn’t graduate were more likely to receive special-education services. The majority of patients in the study chose school against clinical advice.

Thompkins said her family’s decisions are “guided by what we believe is in our son’s best interest.”

Thompkins said kindergarten is going well for her son, now 6. The boy loves story time and learning numbers, and he hasn’t had any meltdowns at school. “He wakes up and he says ‘New day!’” she said. “And he gets ready and he waits for his bus and he looks like he’s just so happy.”The Journal analyzed Medicaid claims for autism-therapy services that states reported to the U.S. Centers for Medicare and Medicaid Services from 2019 through 2024 to determine patients’ weekly hours of therapy.

The analysis used a definition of autism-therapy services cited in the federal Centers for Medicare & Medicaid Services’s Applied Behavior Analysis Toolkit that reflects medical billing codes that were adopted by most states. The analysis excluded duplicate claims, five states that reported insufficient data to accurately calculate weekly therapy hours, and a small share of patient weeks with implausibly high hours.

The analysis identified the largest private-equity companies in terms of autism-therapy services billed to Medicaid, using the claims data, records from the National Plan and Provider Enumeration System and publicly available deals announcements.

For each large private-equity-backed provider included in the analysis, the Journal calculated the number of hours of therapy services each patient received in each week for 2024. To determine how those companies’ treatment patterns compared with competitors, the Journal identified a comparison group of other providers in the states serviced by the private-equity-backed companies.

It then calculated average-weekly hours for each of those comparison groups, and weighted that average to account for the fact that some of the private-equity-backed companies provided more of their services in some states than others. That weighting method resulted in smaller reported overall differences between the large providers and the comparison groups than would result from using raw averages or weighting by beneficiaries served.

The Journal analysis also examined the weekly therapy hours of patients before and after switching to or from large private-equity-backed providers to determine whether the large providers delivered different amounts of treatment hours. That part of the analysis included only patients who received at least 10 weeks of service both before and after any such transition, and covered 2019 to 2024.

To examine the degree to which companies reduced service intensity over time, the Journal examined patients who began services with each company in 2022 or 2023 and missed no more than six weeks in their first full year of therapy. The analysis included only therapy services provided by behavior technicians. Graphics in this article are based on this methodology.