The Decision Gap

Seventy-two hours. That is the standard for answering a prior-authorisation request in Medicare’s pilot of AI-assisted review, the programme known as WISeR. In one case recorded in the pilot’s first months, the answer took 83 days. We know this not because the system flagged it, nor because a department announced it, but because a digital-rights organisation went to court. Roughly 1,000 pages of records obtained by the Electronic Frontier Foundation through freedom-of-information litigation showed that two vendors denied 5,944 requests in the first three months, and that the vendors were paid more when they denied. A patient-advocacy group, the Medicare Rights Center, then set out what the records meant for people waiting on care. The signal existed from the first day a deadline was missed. The public conversation began months later, in a lawyer’s filing.

That interval, between the moment a problem becomes knowable and the moment someone with authority decides what to do about it, is the subject of this essay. Call it the decision gap. Every government has one. Few measure it, and fewer still are organised to close it.

The anatomy of a gap

It helps to break the gap into parts, because each part has a different cause and a different cure.

First comes the signal: the earliest trace that something is going wrong. It might be a complaint to a call centre, a caseworker’s note, a missed service standard, a spike in appeals, a local news report or a cluster of posts in a regional language. Next comes recognition: the point at which someone inside government sees the pattern rather than isolated incidents. Then ownership: a named person or body accepts that the problem is theirs. Then, finally, decision: a choice is made and recorded, even if the choice is to wait.

Most public debate collapses these stages into one and asks why “the government” did not act. That obscures the real question, which is where the time went. A gap that is mostly signal-to-recognition is a problem of sensing. A gap that is mostly recognition-to-ownership is a problem of structure. A gap that is mostly ownership-to-decision is a problem of politics, capacity or nerve. They need different remedies, and treating one as another wastes effort.

Problems announce themselves

There is a comforting myth that crises arrive suddenly. Most do not. Administrative failures in particular tend to build up slowly and in public. A benefits system that wrongly refuses claims produces appeals before it produces headlines. A procurement failure produces late deliveries before it produces an inquiry. A misfiring algorithm produces complaints from the people it misjudges long before it produces a ministerial statement.

This follows from how organisations work. The people closest to a failure, whether frontline staff, the people it affects, local reporters or specialist charities, see it first because they live with it. Information then has to travel upward and across, through layers that each filter, summarise and soften it. Every layer has a reasonable motive to wait: to check the facts, to avoid raising a false alarm, to let a colleague handle it, to avoid being the bearer of bad news. Each delay is small and defensible. Together they add up to months.

The WISeR records show the pattern in miniature. A 72-hour standard is a built-in tripwire: every breach is, in principle, a signal that the system logs. The question is not whether the information existed but whether anyone was positioned, and obliged, to act on it before an outside party forced it into the open.

Learning from the newspaper

Ask senior officials in almost any country how a problem first reached them, and a familiar answer comes back: they read about it. The press office, not the policy directorate, often serves as a government’s early-warning system.

There are structural reasons for this. Journalists do something that departments find hard: they work across silos. A reporter following a story will speak to the claimant, the caseworker, the local councillor, the supplier and the opposition spokesperson in a single afternoon. A department sees only the slice of the problem that falls within its remit. Media coverage also comes with a deadline and a reputational cost attached, which turns a dormant issue into an urgent one overnight.

But a government that learns about itself mainly from the media has handed its agenda to the news cycle. It responds to what is newsworthy rather than to what is important, and these are not the same thing. Slow, technical failures that hurt many people quietly may never become a story at all. Vivid, isolated incidents may dominate a week. Worse, by the time an issue is in the headlines, the range of options has narrowed. The decision is no longer “what should we do?” but “what must we be seen to do?”

A government that learns about itself mainly from the newspapers has already lost the chance to decide quietly and well.