Scored in the Dark

A father in Ontario wanted to know one thing: how the government decides how much money his autistic child gets for therapy.

He did what citizens are supposed to do. He filed a freedom-of-information request with the Ministry of Children, Community and Social Services for the criteria used to set a child’s core-services budget under the Ontario Autism Program. The ministry searched its files and found exactly one record that answered the question, a 26-page instrument called the Determination of Needs Tool. Then it refused to give him a single page.

The legal ground is worth sitting with. The ministry did not say the tool held another child’s personal information, or that it was a cabinet secret. Its lead argument invoked section 18(1)(d) of the Freedom of Information and Protection of Privacy Act (FIPPA): the clause that lets the government withhold records whose release could be injurious to its financial interests, or to its ability to manage the economy.

Stripped of the legalese, the position was this. If parents could see how their children are scored, some of them would lie. They would coach their answers, inflate their children’s needs, and walk away with more public money than they deserve. In March 2024, an adjudicator at the Information and Privacy Commissioner agreed. The tool would stay secret. The order is public, and it is called PO-4494.

Every family seeking core clinical services goes through the Determination of Needs process at least once a year. The tool holds roughly 90 items, scored by a care coordinator in what the ministry calls a semi-structured interview. The coordinator asks questions and assigns ratings across ten domains, reaching into nearly every part of a child’s life.

The output of that scoring is a funding allocation, sorting a child’s needs into bands: limited, moderate, or extensive. Parents consent to having intimate details about a vulnerable child recorded and evaluated by a stranger. They are not allowed to see the framework their child is being measured against.

The dollars attached to those bands are not trivial, and the spread is enormous. The Financial Accountability Office puts core clinical services funding between roughly $6,600 and $65,000 per child per year, based on assessed need. The secret test is the thing standing between those two numbers.

The heart of PO-4494 is the fraud rationale, and it deserves to be read carefully, because the implication is severe. Disclosing the tool, the ministry argued, could reasonably be expected to let some families misrepresent their children’s needs, compromise the process, and force the province to build a costly replacement.

The father’s response was blunt. He called it reprehensible for the ministry to cast aspersions on already stigmatized parents of autistic children. The adjudicator sided with the ministry.

Here is the part that should give any reader pause. The order records no evidence that a single family had ever actually gamed the tool. Section 18(1)(d) requires only a reasonable expectation of harm, not proof. There was no hearing at which the claim could be tested, no cross-examination, and no expert evidence about the program’s real experience of fraud. The assertion was accepted because the law did not require it to be proven.

Fraud was not the only theory. The adjudicator also accepted that the information in the tool has commercial value from not being known, a separate branch of section 18 that protects government information the way the law protects a private company’s trade secrets. A clinical instrument built to assess disabled children, administered at public expense, was found to hold trade-secret-like commercial value the public may not see. The order never asks the obvious question: commercial value to whom?

The most revealing move in the order is the precedent the ministry reached for. To justify the secrecy, it pointed to an earlier decision about lottery tickets, specifically the security numbers printed on lottery transaction records, where releasing the numbers would undermine the integrity of the lottery and invite fraud. The ministry argued the Determination of Needs Tool was analogous. The adjudicator accepted the comparison and applied it.

The entire case rests on one claim: that disclosure threatens Ontario’s finances because families would extract more money than the program can sustain. The province’s own record makes that hard to credit.

The constraint on the program is not that too many children are taking too much. It is that the overwhelming majority are taking nothing, sitting on a waitlist for years while committed dollars go unspent. The Financial Accountability Office has flagged roughly $174 million in autism funding carried forward rather than delivered as services.

The bottleneck the evidence describes is the determination process itself, the same assessment the tool governs, clogged because every funded family must be re-assessed every year.

There is a way to test whether the blackout was really about protecting public money. Less than a year after PO-4494, the same Commissioner faced an almost identical question about the Passport program, which scores adults with developmental disabilities for funding. That decision did not break from the autism ruling. It adopted it. The only reason anything was disclosed is that the ministry had already published part of the Passport tool online, and you cannot claim financial harm from releasing what is already public. For autism, the ministry published nothing and sealed all 26 pages.

Where the Commissioner has grown genuinely demanding is elsewhere. In a 2025 order about major transit records, an adjudicator rejected the same financial-harm exemption because the claimed harm was speculative and not tied to anything specific in the records. Same exemption, same year, two different burdens. The autism tool was sealed on the lighter one.

The freedom-of-information fight is not the only place this program has been tested. In early 2025 the Divisional Court released companion judicial-review decisions arising from a human-rights proceeding that served as a test case for nine families. The lasting value of those rulings is not the result. It is the record, because to defend its program, Ontario put its own evidence before the tribunal and the court wrote it down.

Under the first version of the program, Ontario’s own evidence was that funding reached only about 15% of eligible autistic children, while the other 85% received nothing. The province nearly doubled the autism budget in 2019, from $361.5 million to $667 million, to fix exactly that. Today’s roughly one in four served still describes the same condition: fund a minority, waitlist the majority.

The admission that turns the story over is the conflict of interest. To justify dismantling the old program, Ontario told the tribunal its central flaw was that the clinicians who assessed each child and set the funding level were the same people delivering the therapy and being paid from it. Assessor and beneficiary were one. The province was right to see the problem. In the system built to replace it, the assessment is run by a paid private consortium using a tool the public cannot see. Whether that eliminates the conflict, or relocates it behind a contract and a sealed document, is the question the secrecy makes unanswerable.

There is a process to dispute a child’s level of need, but the court confirmed there is no mechanism at all to seek an exception to the age-based funding caps. A ministry director testified that allowing exceptions would jeopardize the program as the exceptions start to become the rule. So you may dispute the label the tool assigns, you may not see the tool that assigns it, and you may not appeal the cap even if the label is right.

The care coordinators who administer the secret tool and assign the scores do not work for the ministry. They work for AccessOAP, an intake organization run by a consortium led by Accerta Services, a private benefits administrator, together with Autism Ontario, the care-coordination firm Serefin, and research centres at McMaster University. The single largest non-therapy line in the program was an estimated $57.9 million paid to AccessOAP.

You can file a request to the ministry for records it holds about the consortium. You cannot file one against the consortium for its own internal records: how costs are split, how the operation is staffed, what performance data never reaches the ministry. A private company running a public program sits largely beyond the reach of the law built to make public programs transparent, and no Auditor General performance audit of the program is on record.

A hard line is required here. There is no evidence in the public record that AccessOAP, Accerta, McMaster, Autism Ontario, or Serefin has misused public money or acted improperly. The problem is structural, not an accusation. Whether these organizations are performing well or badly, the public currently has almost no independent way to find out.

Ontario points to a dispute process as a safeguard. Set it beside the sealed tool. Families are told they may challenge the score the tool produces. They are also told they may not see the tool. An appeal right against a hidden instrument is barely an appeal right.

It would be easy to read PO-4494 as a narrow dispute over one form. It is more useful to read it as an early data point. In the spring of 2026, the same government passed Bill 97, a budget bill that rewrote the freedom-of-information law itself and removed all records held by the Premier, cabinet ministers, and their political staff from the system, retroactive to 1988. It wiped out pending requests and nullified a ruling that had ordered the release of the Premier’s personal-phone call logs from the week of the Greenbelt decision.

Ontario’s Information and Privacy Commissioner warned that if records about government business can be shielded “simply because they sit in a minister’s office, on a staffer’s device, or within a political account, public accountability is eviscerated.” The autism tool and the Greenbelt logs are not the same scandal. They run on the same logic, and the autism case is the version aimed at the people with the least power to fight it: parents already exhausted by a years-long waitlist, told that the reason they cannot see the test is that they cannot be trusted not to cheat it.

Most of this story is unanswerable from the outside, by design. That is itself the finding. But it points to specific questions that disclosure could resolve, and that the public is entitled to put to the province. None of them accuses anyone of wrongdoing.