Policy analysis · drug regulation & supply · sources checked

When submission volume exceeds review capacity: what a drug regulator can change without lowering the bar

A regulator that cannot review everything on time has exactly two honest options: change what it reviews first, or change how much it reviews. Everything else is a way of quietly changing the standard. Health Canada has just chosen the first, and said so plainly.

The concession is the most interesting part

Regulators rarely publish the sentence Health Canada published: that it is missing its own performance standards, and why. That admission is what makes the rest analysable. A backlog described as a temporary surge invites patience; a backlog described as a structural mismatch between volume and capacity invites a decision about allocation — and allocation decisions can be argued about, which is the point.

It also draws the line that matters. Prioritisation changes review order. It does not change what must be demonstrated, and Health Canada says so in terms. A reader who takes away "Canada is fast-tracking generics" has read it wrong; a reader who takes away "Canada has decided which sponsors wait" has read it right.

Seven instruments, and what each one costs

01 · Triage by clinical need

Prioritise submissions that relieve a shortage, add a first generic, or restore supply. Health Canada’s stated criteria.

What it costs: Defensible and measurable. It also deprioritises the fourth or fifth generic of a well-supplied molecule — which is where price competition comes from.

02 · Capped response clocks

90 days to answer a deficiency, one extension request, up to 90 further days, 180 total, then withdrawal or cancellation.

What it costs: Ends the open-ended file. The sponsor with a genuinely complex analytical problem and the sponsor who is stalling get the same clock, which is the point and also the cost.

03 · Targeted backlog flexibility

For submissions already past performance targets in Review 2, allow further safety, efficacy or quality questions after reviewing a noncompliance response rather than restarting.

What it costs: Clears files that would otherwise die procedurally. The regulator states expressly this is not a general future practice — which is honest, and also means sponsors cannot plan around it.

04 · User fees

Fund review capacity from the industry that consumes it.

What it costs: Aligns funding with volume; creates a dependence that has to be governed, and prices out small sponsors who bring the marginal competitor.

05 · Reliance on comparable authorities

Use another regulator’s assessment where the standard is genuinely equivalent.

What it costs: Highest-leverage option available. Requires the reviewing authority to accept that equivalence is a judgement it can be wrong about.

06 · Domestic manufacturing preference

Prioritise review for submissions with domestic manufacture.

What it costs: Addresses a real supply-chain vulnerability with a regulatory instrument. Invites reciprocal treatment abroad and imports industrial policy into a queue that is meant to be merit-neutral.

07 · Staffing

Hire and retain more reviewers.

What it costs: The only option that raises capacity rather than reallocating it. Slowest, most expensive, and least announceable.

The question prioritisation cannot avoid

A first generic relieves a monopoly; a fourth generic relieves a price. Both matter, and a criterion written around access and shortage will systematically favour the first over the second. That is a coherent choice about what a regulator is for, and it should be made explicitly rather than as a by-product of triage language.

The sponsor deprioritised for bringing a duplicate has no obvious basis to object on the merits and a real commercial injury nonetheless. Nothing in the announced package gives that sponsor a route to be heard, and a queue without a review mechanism is a decision without accountability.

Domestic manufacturing preference is a different kind of instrument

The other six instruments allocate a regulator’s own scarce attention. Domestic-manufacturing preference imports an industrial-policy objective into that allocation. It is defensible — a supply chain that fails is a clinical problem, not an economic abstraction — and it is still a departure from the principle that a queue should be indifferent to where a compliant product is made.

Two consequences are foreseeable. Other jurisdictions can reciprocate, and a network of mutual domestic preferences slows every cross-border submission. And "domestic" is a fact about a facility, not about resilience: a single domestic plant is a single point of failure regardless of its address.

What this means for the United States

The comparison is unflattering in one specific respect. The United States funds review capacity substantially through user fees and has correspondingly less need to triage, but it has no published mechanism that prioritises a submission because a drug is in shortage — the shortage response sits in separate programmes with separate authorities. Canada has just wired the shortage signal directly into review order.

The reliance question also connects here: the ICMRA collaborative-assessment work is the only instrument on the list that raises effective capacity without either more money or a reordered queue, and it is the one that requires a regulator to trust a judgement it did not make. That trade is examined in regulatory reliance without regulatory surrender.

What KPSGILL does and does not conclude

No position is taken on whether Canada chose correctly. The package is well-designed on its face: it names the constraint, states the criteria, caps the clocks and preserves the standard. Whether it produces faster access to the medicines patients actually lack is an empirical question that will be answerable in a year from Health Canada’s own performance reporting, and this site will read that reporting rather than predict it.

One observation does survive: a regulator that publishes its capacity constraint and its allocation rule can be held to both. One that quietly lets timelines slip cannot.

Sources