The Wrong ABA Providers May Be Winning – and Payers Are Helping. Quality Metrics Can Change That

The Same Rate for Everyone

A CPT (billing) code does not know who billed it. When an ABA organization submits a claim for units of adaptive behavior treatment, the payer’s system sees a code, a diagnosis, a date, and a rendering provider number. It does not see whether the technician who delivered those units was well trained or not. It does not see whether a BCBA watched the session and coached the technician afterward. It does not see whether the treatment plan was followed with fidelity or improvised.

It does not see the organization behind the technician: whether it invested in a culture where staff stay, show up rested, and pay attention, or whether it churns through people who count the minutes while totally checked out. It sees the code and the bill. Every provider in a network is paid the same fee schedule for the same code, regardless of what went into the hour, reviewed on the same cycle, and trusted to the same degree.

That was never ideal, but not as noticeable when ABA was small. That is no longer the case. Medicaid and CHIP payments for ABA rose to $10.1 billion in 2025, according to a CMS toolkit for states released in August 2026. Medicaid is the largest payer, but it is not the only one. Since Tennessee became the last state to act in 2019, every state has required its state-regulated commercial plans to cover autism treatment, including ABA, and those plans use the same codes, the same authorizations, and the same flat-fee schedules.

The response to that growth has been cost containment, and almost none of it has anything to do with quality. One state has capped a child’s lifetime ABA at 4,000 hours, set weekly caps by diagnosis, cut its fee schedule 6 percent with more cuts to follow, and frozen new agency enrollments for an initial six months. Another has set weekly hour standard limits and requires a treatment plan review at least every 90 days. Commercial payers, in my experience, have taken the same road: more paperwork at reauthorization, shorter intervals between reviews, tighter hour limits. The changes are not clinically informed. These are blunt instruments applied equally to every provider. A lifetime cap does not distinguish a child who is progressing from one who is parked. A shorter reauthorization cycle asks the best provider and the worst for the same paperwork twice as often. A rate cut takes the same 6 percent from the organization that trains its staff well and one that practices RBT “baptism by fire”.

None of this makes a single hour of therapy better. It makes every hour cheaper and less accessible. ABA is not under-regulated; its providers are unsorted.

What Quality Costs Up Front

I ran an ABA and pediatric therapy company for more than 12 years. We lived quality as a value because we believed in it. It was in our DNA, and we invested in it. We built training roles whose only job was to prepare technicians before they ever sat across from a child. We supervised above the minimum the certification board required. We paid people to ensure high-fidelity treatment. Every one of those choices came out of the same reimbursement everyone else received for the same code.

Down the street, a competitor could skip all of it. It could hire technicians, bill the maximum authorized hours, offer little or no training, and keep supervision to a minimum. That competitor’s margin on each hour was higher than ours, and the difference did not go into the clinical program. It went into sales, marketing, and new locations. The payer treated us identically. In practice, it treated the lower-quality operator better because it left that operator more money to grow with.

This is the central distortion in how ABA is paid. High quality requires upfront investment: training, supervision, measurement, compliance staff, lower caseloads. Low quality requires only the minimum of each and produces the same claim. A flat rate does not stay neutral between the two. It subsidizes the one that invests less in the child.

The distortion is also expensive for payers, just not on the day the claim is paid. The bill for a provider that delivers 30 hours a week to one child for four years without measured progress is the cost of quality that was never bought. So is the reauthorization cycle itself, which exists in part because the payer does not know whom it is dealing with. A child supported by a well-trained, low-turnover, supervised team makes development gains faster and moves toward fewer hours and, eventually, discharge. A child who is not properly treated stays on the schedule. Quality is less expensive in the long term.

The Idea of Sorting Providers Is Not New

Elsewhere in health care, payers sort providers by quality and attach money to the sort. Medicare raises the payment benchmark five percentage points for a Medicare Advantage plan rated four stars or better, and KFF estimates those quality bonuses will cost at least $13.4 billion in 2026. CMS withholds 2 percent of hospitals’ Medicare inpatient payments and redistributes the pool according to mortality, infection, safety, patient-experience, and cost scores. One national insurer’s behavioral health program scores its therapists on standardized outcome data and gives the top group a 12-month 3 percent increase in their fee schedule, while top-scoring facilities get what the insurer calls a “streamlining” of clinical review.

Almost none of that infrastructure has reached ABA. There are designations, pilots, and one announced value-based contract, these are all great first steps. It is time to supercharge these efforts to separate the ABA companies who are dedicated to clinical quality, who put families first, from the companies that are focused on short-term profits at the expense of everything else.

How the Grades Could Work

Here is what I would consider. Every ABA organization gets a grade for example, from A through F, on a weighted composite of metrics.

Clinical outcomes: Standardized assessment gains at six and twelve months, goal mastery, and whether children are moving toward less intensive service rather than sitting at maximum hours indefinitely. Every serious provider already collects this data for its own treatment planning, on instruments every BCBA learned in graduate school. What is missing is not the data.

Integrity: Credential verification, supervision documentation, audit history, and claims accuracy.

Workforce investment: Training before a technician’s first billable session, supervision above the minimum, technician and BCBA retention.

Operational reliability: Authorization approval rate, documentation completeness, timeliness. Approval rate belongs in the composite but should never be the whole of it.

An A provider earns the highest reimbursement and the lightest paperwork. An F provider is on prepayment review or out of the network. Everyone in between knows exactly what it would take to move up.

The grades carry consequences in both directions. An A provider earns the highest reimbursement, the longest authorization periods, the lightest documentation on reauthorization, expedited credentialing when it adds sites or clinicians, and the presumption of good faith when a claim looks unusual. B providers get most of that. C providers get the status quo, which is what everyone gets today. D providers get lower rates, shorter authorizations, and concurrent review. An F provider is on prepayment review or out of the network. The payer has then spent its scrutiny where the grades say the problems are, rather than spreading it evenly across the network.

That ladder does two things a flat rate cannot. It returns the up-front investment in quality to the organizations that make it, and it removes the growth subsidy from the ones that do not. Then it lets the market finish the job. An A provider keeps doing what it does well and keeps the resources to do more of it: more referrals, faster credentialing when it opens a site, more room to hire and train. A D provider spends its thinner margin clearing hoops instead of on marketing. Over time, market share moves toward the organizations with the best outcomes; more children are served by them, and the payer ends up with the lowest cost that was ever actually available: the cost of care that works the first time.

Who Would Assign the Grades

There are three ways to get there, and they are not mutually exclusive.

The first is an independent body that grades providers and publishes the results, with payers agreeing to pay and review based on the grade. If an accrediting body could validate effective outcomes that would be a game changer. The advantage is one scorecard instead of a dozen across various payers. The risk is that any single body becomes a gatekeeper without accountability to the payers and families who rely on it.

The second is a public-private partnership that starts with Medicaid. State Medicaid programs are rewriting their ABA rules right now through moratoriums, hour caps, and accreditation mandates. A state that is already rewriting those rules could write a grade into them and let its managed care plans pay by it. Commercial plans could adopt this approach if it works.

The third is to let innovation produce the scorecard, and I find this route most promising. Several practice management platforms, outcomes registries, and billing vendors already hold the data an A-through-F grade would run on. The bigger opportunity is what comes next.

Artificial intelligence, a phrase that makes many clinicians wince, may soon do what no reviewer reading a six-month progress report ever could: evaluate quality at the session level. A model trained on what a good ABA session looks like can flag whether a technician is delivering the plan with fidelity, whether the child is engaged, whether reinforcement is delivered as written, and whether the supervising BCBA’s feedback shows up in the next session.

Done in a HIPAA-compliant way, with the data protected and the methods open to scrutiny rather than locked inside one vendor, that is a quality signal a payer could trust more than any form. Payers should adopt the best of these tools as they mature and let competition decide which grading model survives.

The objection I hear first is that small practices cannot produce this data. Five years ago that objection had force. Today, outcomes measurement, supervision logs, and training records live in software most clinics already pay for, and the tools that can read them are getting cheaper every year. The cost of proving quality is collapsing, opening up incredible new opportunities.

Quality is best for children, who get more of what works and less of what does not. It is best for high-quality providers, who finally get paid for what they invest. It is best for payers, who stop buying hours and start buying outcomes. And it is best for an industry that has spent four years being defined by its worst actors. The pipes for a quality signal are being built anyway. Whether anything about quality flows through them is a decision we all, as an industry, have to make.

AT A GLANCE

SOURCES & REFERENCES