By the end of 2022, six doctors had voluntarily published their fees on the Medical Costs Finder. The site had been running since December 2019 and had cost, on the government’s own account, about $24 million to set up.

By September 2026 the figure had reached 106, out of around 6,300 eligible specialists.

That is the cleanest answer available to a question that comes up in every scheme design I have worked on: will the sector self-report if we ask nicely? The answer here took seven years, $24 million and roughly one in sixty. The government’s position is that specialists were asked to publish their fees voluntarily and almost none of them did.

This is a finding about disclosure design, not a story about doctors. Voluntary disclosure regimes fail in a predictable pattern: when publishing is costly to the person publishing, when the benefit accrues to somebody else, and when the regulator has no fallback if nobody participates. Every one of those conditions held here. Nothing in the Medical Costs Finder’s design required a specialist to do anything, and nothing happened to those who did not.

The Health Legislation Amendment (Improving Choice and Transparency for Private Health Consumers) Act 2026 finally passed both Houses on 16 September, received assent two days later, and its publication provisions commenced the day after that.

It does not compel specialists to disclose anything. There is no obligation on a practitioner to supply fee information and no penalty for withholding it. What the Act does is insert a publication power into the Health Insurance Act 1973, with mirror provisions in the Private Health Insurance Act 2007. The Secretary “may publish information relating to professional services” rendered by or on behalf of medical practitioners, where Medicare benefits are payable. Publishable information expressly extends to practitioner names, qualifications, specialties, locations, hospitals, the Medicare benefits paid and the medical expenses incurred for services by identified practitioners.

So the mechanism works through publication instead of compulsion: the government publishing, at named-individual level, data it was already collecting through Medicare, hospital and insurer billing. The voluntary upload was never the only route to that information, merely the only route the government had given itself permission to use.

There is one more provision, and it is the one nobody is talking about. Under section 124ZZA, the Secretary is not liable to civil proceedings for loss of any kind suffered by any person as a result of the publication of information under that power.

On 6 July 2026 the Digital Transformation Agency brought its Digital Seller Underperformance Policy into effect. It applies to Strategic Digital Contracts, meaning agreements for digital products and services with a whole-of-life value of at least $4 million GST exclusive, or others the DTA determines to be strategically significant. Where a seller is found to have seriously underperformed, the finding is recorded on a Confirmed Serious Underperformance Register.

The DTA is emphatic that this register stays inside government. In its own words it is “internal to the Australian Government and information collected through the DSUP process is not made public by the DTA”, recorded and shared on a needs-to-know basis internally, with disclosure to buyers governed by Commonwealth law and policy. The DTA is equally emphatic that it is not a blacklist: the policy does not prohibit an agency from engaging a seller who is on the register, and a listing becomes one factor in later value-for-money assessments rather than an automatic exclusion.

When the DTA was asked why the register stays inside government, the answer was legal risk to the Commonwealth, defamation among it.

That answer is genuine and I do not think it is a dodge. A finding of serious underperformance against a named firm is an adverse finding about an identified party, made administratively, with commercial consequences. The DTA built real process around that finding: a requirement that buyers attempt direct resolution first, fifteen business days for a seller to respond, a right to request internal review, and a retention period cut from five years to three after industry consultation. The peak body still objected, publicly, before the policy was made.

Hold the two instruments next to each other. Within about ten weeks, one arm of the Commonwealth decided that publishing identified individuals’ commercial information carried a legal risk it could manage, and legislated an immunity so it could publish anyway. Another arm decided that publishing identified firms’ performance information carried a legal risk it could not manage, and kept the information inside government.

The difference between those two outcomes is not a difference in the underlying risk. Both involve publishing something adverse about an identified party who would rather it stayed private, and both invite exactly the same cause of action. The difference is that one of them was being done through legislation, where an immunity clause is available, and the other was being done through policy, where it is not.

The gap gets described as a principled distinction when it is really an instrument-selection one. A surgeon’s fee schedule is now publishable by name. The delivery record of a firm holding a $4 million Commonwealth contract stays inside government. If you think that record should be public, the useful contribution is to specify what evidentiary standard, what review right and what statutory protection would make publishing it defensible. The health legislation has just demonstrated that all three are obtainable when somebody decides to obtain them.

And if you think it should stay private, that position is available too. But it should be argued on its merits instead of on legal risk, because legal risk has now been shown to be a variable the Commonwealth can set.

Two questions, and I would put both of them on the table before anything is built.

The first is the fallback question. If your scheme depends on voluntary participation, what happens at two per cent? Most schemes I have reviewed have no answer, because participation was assumed rather than modelled. Seven years is a long time to discover that the answer is six doctors. The fallback does not have to be compulsion. It has to exist.

The second is the risk question. When somebody says a disclosure regime cannot be public because of legal risk, ask whether that is a finding or a starting position. Occasionally it is a finding, reached after someone worked out what protection would answer the risk and concluded it was unavailable or disproportionate. More often it is where the conversation stopped.

The Medical Costs Finder ran for seven years on the assumption that people would publish information against their own interest if the website was good enough. The website was fine. In its first three years, six doctors used it.

References