From Evidence to Action: The Coordination Work Canada’s Drug Agency Is Positioned to Lead

A Portfolio of Coordination Challenges

The developments of the last year fall into a recognizable shape. In each, the agency holds part of the answer, several other parties hold the rest, and there is a clock running. The tiles below name the challenge, the structural trap that keeps it open, and the trigger that makes it live now.

The Lecanemab Reconsideration

CDA-AMC issued a draft recommendation against public reimbursement of lecanemab in February 2026, even though Health Canada had granted the therapy a conditional approval the previous October⁹. The manufacturers responded publicly⁸, and a reconsideration meeting is set for June 24, 2026¹⁰. Three mandates meet here and none can overrule the others: the regulator assesses safety and efficacy, the agency assesses cost-effectiveness and system readiness, and provincial payers hold the budgets. There is no single forum where that discrepancy is resolved, which is what makes the outcome precedent-setting for the wider neurodegenerative pipeline.

What makes the case more than a single difficult decision is the precedent it sets. A growing class of disease-modifying therapies for neurodegenerative conditions will arrive carrying the same profile: regulatory approval on clinical grounds, genuine patient demand, meaningful uncertainty about real-world benefit, and a price that strains public budgets. If each arrives into the same unstructured standoff between regulator, agency, and payer, the pattern will repeat. The open question is whether the parties can agree, in advance and together, on a pathway — managed access, conditional funding tied to evidence collection, or an outcomes-based arrangement — that converts a binary yes-or-no into a sequence the system can actually carry.

Caregiver Impacts in Economic Evaluation

In March 2026 the agency published draft guidance expanding economic evaluations to reflect informal caregiver health-related quality of life and productivity, with the pilot scheduled to conclude in July 2026⁵. Counting caregiver burden changes the arithmetic of cost-effectiveness: a therapy that strains a drug budget can look efficient once societal productivity is included. The actual payers, provincial ministries of health, work within budgets that do not capture those cross-sector gains, so a methodology can be accepted in principle while the willingness-to-pay thresholds it implies go unadopted in practice.

The phrase that captures the trap is agreement in principle. A provincial payer can endorse the broader-value methodology sincerely while its finance officials, working a different ledger, decline to move the threshold that would actually fund a therapy on those grounds. Nothing is being obstructed; the two positions simply live in different parts of government and never have to be reconciled in one place. Closing the pilot in July 2026 without that reconciliation would leave a sophisticated methodology that changes the arithmetic on paper but not the decisions in practice.

The 180-Day Reimbursement Timeline

Reimbursement review enhancements taking effect in August 2026 set a 180-calendar-day target for final recommendations, following consultations earlier in the year⁷¹¹. A fixed clock is meant to speed access. But it compresses sponsors, internal reviewers, and patient and clinician contributors into a tight sequence, and absent a shift toward earlier presubmission alignment and faster data exchange, unresolved uncertainty at day 180 can produce deferrals rather than approvals — the opposite of the intended acceleration.

Regenerative Medicine at Scale

The April 2026 Watch List forecasts rapid near-term growth in highly personalized platforms such as adoptive cell therapies and bioengineered tissue, while flagging real limits in domestic manufacturing capacity and workforce readiness⁶. Delivering these therapies depends on specialized treatment centres, advanced bio-logistics, and local clinical expertise operating together. Today those pieces do not sit under a shared national framework, and the per-patient costs strain smaller provincial systems.

Rare Disease Registry Funding

A 2026-2027 funding opportunity launched in March 2026 sustains rare disease registries on an annual grant basis¹². Reassessing high-cost orphan drugs requires multi-year real-world data, but no single party has committed to capitalizing the registries permanently, leaving the evidence backbone exposed each time a grant cycle ends.

Newborn Screening for Metachromatic Leukodystrophy

Early in 2026 the Health Technology Expert Review Panel recommended adding metachromatic leukodystrophy to the pan-Canadian newborn screening list, conditional on a costly gene therapy being publicly funded first so that screening would not create a two-tiered system¹³. The condition is principled and produces a circular hold: screening waits on therapy funding, while provinces are cautious about funding the therapy before a screening pipeline can size the patient population.

Imaging No-Shows and Diagnostic Capacity

CDA-AMC's March 2026 report on no-shows in medical imaging found that proactive, automated interventions reach success rates as high as 88% in reducing missed appointments, against a backdrop in which missed appointments across CT, MRI, X-ray, and ultrasound absorb an estimated $110 million in directly unrecoverable overhead annually across the Canadian system — staff paid, equipment running, slots wasted¹⁴¹⁵. The report followed the establishment of the pan-Canadian Medical Imaging Task Force in September 2025³. The evidence is settled and the better intervention is named; what is missing is a mechanism to deploy it, since the agency cannot direct procurement at the hospital or provincial level.

The report also documented how uneven the picture is across the country, with the sharpest capacity strain falling on territorial and remote sites. That unevenness matters because a solution adopted in well-resourced urban centres while smaller sites are left behind would widen the very access gap the work is meant to close. It is the cleanest illustration in the portfolio of a problem where the answer is genuinely known and the only open question is coordinated deployment, which is why it serves below as the worked example.

National Bulk Purchasing and pCPA Capacity

Fulfilling a federal mandate, an advisory panel recommended expanding the pan-Canadian Pharmaceutical Alliance to handle national bulk purchasing, while explicitly leaving logistics, delivery, and storage outside its scope²¹⁶. The upstream strategy is federally guided; the downstream execution risk sits with provincial supply chains, and there is a narrowing window to build the alliance's capacity before standard negotiations back up.

The Essential Medicines List

A twelve-member advisory panel delivered a proposed essential prescription drugs list to Health Canada following a 2025 consultation, with the final report deferring to existing public and private coverage rather than replacing it¹⁷. The list is clinically grounded and widely supported, but the budgets that would fund it universally are provincial. Without a binding agreement on how federal transfers reach provincial drug plans, the list remains an advisory baseline rather than coverage a patient can use.

Validating AI-Generated Real-World Evidence

The agency's October 2025 position statement accepts that AI and natural-language processing will increasingly generate evidence for complex submissions, while naming concerns about transparency and appropriateness¹⁸. Auditing that evidence calls for specialized computational and biostatistical capacity, and a multi-jurisdictional means to examine proprietary algorithms does not yet exist, so the expectation is set on paper ahead of the tools to enforce it.

The bulk purchasing advice has the same shape from the opposite direction: a federal strategy that leans on the pan-Canadian Pharmaceutical Alliance to execute, with the operational scope deliberately bounded and the supply-chain reality left to the provinces. In both the medicines list and the purchasing strategy, the design is sound and the boundary is the problem. The work that would close them is not more design; it is a negotiated agreement among the federal government, the provinces, and the alliance about who funds what, who carries which operational load, and on what schedule.

Across the portfolio the same five features recur: multiple constraint owners, none of whom can act alone; agreement reached in principle that does not harden into binding commitment; dependence on a wider set of actors who must move together; a real deadline or a funding sunset; and a gap between a sound framework and its implementation. When that combination appears, the limiting factor is coordination.

Why the Usual Tools Reach Their Limit

CDA-AMC already uses a strong set of convening instruments — public consultations, advisory panels, expert committees, and a flagship annual symposium¹⁹. These are well-matched to the stage of work they are built for. Consultations confirm where stakeholders already agree and surface where they do not. Panels assemble expertise and produce rigorous analysis. Symposia raise shared understanding of causes and, at their best, advance mid-level analysis of solutions.

What these formats are not designed to do is force trade-offs among parties who control different constraints, or test options against the people who would have to implement them. Those people are rarely at the table; senior leaders attend in their place, and agreement reached at that altitude stays general. The result is a familiar sequence: a clear analytical product, broad endorsement of its direction, and then slow, uneven movement at the boundaries the agency does not own.

There is a second reason the usual formats stop short. They are organized around presenting and reacting — a body of work is shared, attendees respond, and the most articulate or senior voices tend to shape the conclusion. That format rewards the appearance of agreement. Saying yes in principle costs nothing in the room and defers the hard part to later, when the parties are no longer together and the specifics that would have forced a real negotiation never get raised. The agreement that results is real but shallow, and it does not survive contact with the budget cycle, the procurement process, or the competing priority that surfaces three months on.

The missing instrument is a decision-forcing process in which the room itself does the work — the actual constraint owners present together, the real trade-offs surfaced and resolved, and a plan built that those same people then own. The mechanics matter less than the principle: a phase that surfaces the constraints honestly and aligns everyone on the same facts; a phase that widens the set of options by reasoning across boundaries rather than within them; and a phase that converges on commitments with named owners and resolved trade-offs. Crucially, the people who will implement are present, not represented, because a commitment made by the person who has to keep it behaves very differently from one made on their behalf. The sequence below shows what that looks like applied to one of the portfolio challenges.

From Evidence to Committed Action: A Three-Stage Process

The structure is simple to state and demanding to run: Analyze, then Diverge, then Converge. Its discipline lies in keeping the stages separate — surfacing the real constraints before generating options, and generating options before forcing a choice — and in insisting that the people who control the constraints are physically in the room. What follows applies the structure to two of the challenges above. The mechanics are identical; the content differs.

Worked Example A — Imaging no-shows and the cost of the empty slot

Analyze: Putting the Real Numbers in Front of the Real Owners

Begin with what the Analyze stage would put on the table, because in imaging the facts are unusually concrete. A missed appointment is not a neutral event; it is false capacity — a slot that appears productive on the schedule but yields nothing, while the staffing and overhead behind it are spent regardless. The national picture, assembled from the March 2026 CDA-AMC report and Canadian microcosting data, is large enough to command the attention of a finance lead, not only a clinician.

Canada performs roughly 6.42 million CT exams a year. At the surveyed national no-show rate of about 5%, that is some 321,000 missed CT appointments, and at an audited sunk cost of $128.31 each, roughly $41 million annually from CT alone. Adding the country's 2.2 million MRI exams at the same rate brings the CT-and-MRI floor to about $55 million²⁰. That figure is conservative: it covers only advanced cross-sectional imaging and only direct staffing cost. Canada also performs an estimated 18.2 million X-ray exams a year at about a 4% no-show rate — some 728,000 missed appointments — which at $24 to $39 of sunk cost each adds $18 to $28 million. Diagnostic ultrasound carries the highest no-show rate of any modality at roughly 7%, generating about 700,000 missed appointments against an estimated 10 million exams, adding some $40 million at $57 to $61 per missed appointment²¹. The national averages also conceal sharp local variation — ultrasound no-show rates exceed 25% in some jurisdictions — which means the burden is concentrated, and therefore addressable, in specific sites.

The point of the Analyze stage is not to admire these numbers but to use them to surface what each actor in the room actually controls. The figures make plain that no single party owns the problem. A regional imaging director controls scheduling and reminder protocols but not capital. A provincial capital-planning office controls equipment budgets but not the clinical criteria that decide which backlog is most urgent. CDA-AMC controls the assessment methodology but not procurement. Radiologists control read capacity but not patient access. The Analyze stage invites each of these to state, in front of the others, the constraint they own and the one they do not — a conversation that the symposium format, for all it accomplishes, more often leaves to the corridor than to the agenda.

Diverge: Generating Options That Cross the Boundaries

The Diverge stage then generates options that a single panel, left to itself, would not reach — because the options that matter cut across the boundaries the actors arrived with. A reminder-and-overbooking protocol tuned to the highest-no-show modalities is one lever; a shared capital-prioritization rule that ranks investment by recovered effective capacity rather than by headline machine count is another; a vendor-neutral procurement standard that lets provinces buy against common criteria is a third. None of these belongs to a single mandate, which is exactly why none has been built.

Crucially, the Diverge stage is bounded by the constraints established in Analyze, which keeps it from drifting into options no one in the room can deliver. An idea survives only if some party present can own it. This is the quiet discipline that separates a structured divergence from a brainstorm: the goal is not the longest list of possibilities but the set of genuinely actionable combinations that no participant would have proposed alone, because each sits across a boundary that only the assembled group can cross together.

Converge: Decisions Built to Survive the Return Trip

The Converge stage is where the structural difference between a solving sprint and a standard working group becomes most visible. Participants work through a structured sequencing of trade-offs with the constraint owners present when each choice is examined. The capital office representative is in the room when the question of procurement prioritization across high-variance sites is worked through. The provincial health authority lead is present when the question of which no-show protocol gets piloted at which tier of site is forced onto the agenda. The Task Force methodology lead is at the table when the accountability criteria those commitments will be measured against are agreed. The process does not strip these actors of their organizational authority — they return to their agencies to champion and implement what emerged. What it does is ensure that the plan they carry back was shaped by their own constraints, tested against every other actor's constraints in the same room, and co-created with the people who must act on it simultaneously.

The imaging data lends itself naturally to this discipline. Because the sunk cost of a missed appointment is concrete and auditable, and because the no-show burden concentrates in identifiable high-variance sites, the Converge stage can attach each commitment to a measurable baseline — making it straightforward to evaluate whether the moves participants carry back are being honoured. A decision that can be measured is a decision that can be defended, which matters greatly to an audience of evaluators and methodologists, and is one reason this challenge is a particularly clean demonstration of the method.

Worked Example B — Regenerative medicine and the capacity no single system owns

Analyze: Naming the Constraints and Who Holds Them

The 2026 Watch List names regenerative medicine — adoptive cell therapies, gene therapies, bioengineered tissue — as the fastest-moving category in the near-term pipeline, while flagging that Canadian manufacturing capacity, workforce readiness, and treatment infrastructure are not keeping pace with the clinical pipeline. The Analyze stage starts from that gap and makes it concrete: what, precisely, does each actor in the system control, and which of those things does everyone else depend on? In regenerative medicine, those dependencies run in unusual directions, and the Analyze stage is essential for surfacing them before the group moves to options.

The cost of each therapy — frequently exceeding $1 million per patient — means that a single coverage decision by a small province can materially strain its drug budget for a year. That arithmetic makes each province cautious and makes national alignment attractive in principle, while the federated structure makes it structurally difficult. No province can be directed to cover a therapy, and no national body can fund it on their behalf. The result is that a therapy approved by Health Canada, assessed by CDA-AMC, and negotiated by pCPA can still stall at the point of actual patient access — held up by a specialized treatment centre that does not exist in the patient's province, a bio-logistics chain that cannot reliably ship a living product, or a clinical workforce that has not been trained for the delivery protocol. The Analyze stage makes plain that the gap between a positive recommendation and a treated patient spans at least four distinct ownership boundaries:

Diverge: Building Options Across the Delivery Chain

The Diverge stage then generates options that no single actor — not CDA-AMC, not a provincial payer, not a treatment centre — would propose on its own, because each option requires crossing a boundary the proposer does not control. A risk-pooling arrangement in which smaller provinces contribute to a shared access fund, smoothing the per-patient shock across a larger budget base, is one such option; it requires the provinces to agree on a contribution formula and a governance structure for claims, which no province would design alone because doing so cedes budget control before the terms are set. A national specialized-treatment-centre accreditation standard — defining the infrastructure, staffing, and quality requirements a centre must meet — is another; it requires Health Canada, the professional colleges, and the provincial health authorities that fund the centres to align on criteria that will bind all of them. A shared bio-logistics protocol, negotiated with air carriers and customs authorities before a product is in transit, is a third.

The Diverge stage is bounded by the same discipline as the Analyze: an option survives only if some party in the room can own a piece of it. The risk-pooling idea stays on the table only if at least one provincial payer is present to say what contribution terms they could defend to their finance ministry. The accreditation standard stays only if a professional college or Health Canada representative can commit to a process. The logistics protocol stays only if someone from hospital operations and someone with supply-chain authority are both present and can sketch the first two moves. What the stage prevents is the wishful option that sounds good because it is general and survives only because no one with implementation responsibility is in the room to test it against reality.

The value of this stage in regenerative medicine is particularly high because the delivery chain is long and each link in it has been built independently, by different actors, for a different era of therapy. The standard working group tends to optimize one link at a time — better logistics, or more centre capacity, or a cleaner registry design — without ever examining how the links depend on one another. The Diverge stage is the moment when that interdependence becomes productive rather than paralyzing, because the people who own each link are in the same room and can see how their individual commitments combine.

Converge: Sequencing Commitments So the Chain Holds

The Converge stage forces the decision that the advisory-panel format is not built to extract. A provincial payer names a per-patient cost threshold above which it would participate in a risk-pooling arrangement, and the terms on which it would enter. Health Canada commits to a timeline for finalizing the accreditation framework for specialized treatment centres, and the professional college commits to beginning the curriculum design that framework implies. A hospital logistics lead from a treatment centre names the bio-logistics protocol changes that would bring their institution into a shared cold-chain agreement. CDA-AMC commits to the registry minimum dataset that will satisfy its real-world evidence requirements — the specification that every other actor needs before they can design their own data-capture workflows. Because the people who have to keep these commitments are present when they are made, the plan that leaves the room is not a recommendation directed at someone else. It is a set of named actions, each owned by the person who stated it.

The sequencing of those commitments is itself a central output of the Converge stage. In regenerative medicine, the dependencies run in a specific order: a registry minimum dataset must exist before provinces can design compliant data capture; a treatment-centre accreditation standard must be in place before new centres can be funded against it; a risk-pooling framework must be agreed before smaller provinces will commit to coverage; and none of these moves is worth making without bio-logistics protocols that ensure the product actually reaches the patient. Mapping that dependency chain in the room — with all the relevant owners present to confirm the sequence — is what produces a 30/60/90-day plan that holds together once people have left. The usual alternative is a series of commitments that each look sound in isolation and then stall at the moment the first actor, waiting on another actor's prior move, discovers that move was never actually made.

The regenerative medicine example also illustrates something the imaging example does not: the problem does not yet have a settled evidence base to anchor the Analyze stage, which makes the discipline of that stage even more important. In imaging, the CDA-AMC report had already quantified the cost of the empty slot; the Analyze stage starts from a known number and works outward. In regenerative medicine, some of the critical facts — the realistic domestic manufacturing timeline, the true cost of centre accreditation, the minimum follow-up period for a credible real-world evidence dataset — are themselves contested, and part of the Analyze stage is agreeing on which facts are settled and which are legitimately uncertain. That agreement is not trivial. Organizations that have been working on different assumptions about the manufacturing timeline, for example, will have drawn different conclusions about which policy moves are urgent. Making the uncertainty explicit, and getting the constraint owners to agree on a shared set of working assumptions, is a precondition for the productive divergence and convergence that follow — and it is work that a panel presentation, however rigorous, cannot reliably accomplish.

The Choice

There is a straightforward way to tell which of these challenges needs a different instrument. If an issue keeps reappearing on successive agendas — raised, discussed, broadly agreed, and then raised again the following year in much the same form — that recurrence is the signal that the limiting factor is coordination rather than analysis, and that a decision-forcing process would change the outcome. If instead an issue is being resolved and closed, the existing tools are doing their job and nothing new is required.

Applied to the current portfolio, the test is easy to run. The questions worth asking are which of these items were also on the agenda a year ago in a similar form, and which are genuinely closing. The ones that recur are the candidates for a different approach.

The Question Before the Deadline

Each of these challenges carries its own clock — a reconsideration meeting, a pilot's end, a timeline taking effect, a funding cycle running out. The useful question to hold against each is simple: a year from now, will this still be on the agenda, or will it have been closed? Where the honest answer is that it will likely still be open, the reason is almost never that the analysis is incomplete. It is that the right people have not yet been in the same room with the authority and the structure to decide. That is a solvable problem, and solving it is a matter of process.