In September 2021, NHS England announced what it called a world-first “population health agreement” with Novartis. The deal would give up to 300,000 people with high cholesterol and a history of cardiovascular disease access to inclisiran, a twice-yearly injection, over three years. It was framed as more than a purchase: a collaboration to find and treat patients at scale through primary care.
Two years later, the Pharmaceutical Journal reported, using NHS Business Services Authority data obtained through a freedom of information request, that 4,549 people had been prescribed inclisiran in primary care between October 2021 and July 2023.
The price had been agreed. The supply was there. What was missing was everything a sales contract cannot buy: clinicians who were persuaded, general practices with time to run injection clinics, and professional bodies comfortable with the evidence. The Royal College of General Practitioners and the British Medical Association had urged caution in late 2021, noting that long-term cardiovascular outcome data were not yet available.
That episode is worth studying by anyone designing a “partnership beyond sales.” Here is the thesis: most pharma–health system partnerships fail, or disappoint, because they are still structured as sales. They measure uptake of one product. The ones that create value measure the health problem, and accept that the answer may not always be the partner’s drug.
The two meanings of partnership
When a pharmaceutical company and a health system say “partnership,” they can mean very different things.
The first meaning is commercial. The company offers a better price, faster access or a risk-sharing arrangement, and in return the system commits to volume or to a place on the formulary. This is valuable. It is also a sale, however it is described.
The second meaning is capability. The company contributes something the system lacks: analytical capacity to find undiagnosed patients, funding for staff to run a care pathway, logistics, training or data infrastructure. The system contributes access to patients, clinical governance and the credibility of public institutions. Both sides pursue a shared outcome such as fewer heart attacks, earlier cancer diagnosis or better-controlled asthma.
Most real agreements blend the two. The trouble starts when a commercial deal is dressed as a capability partnership, because the system then expects support it will not receive, and the company expects uptake the system cannot deliver.
The uncomfortable incentive
The central tension is simple. A health system wants a population outcome. A company wants its product used. Those overlap, but they are not identical.
Consider a program to improve cholesterol control. The most cost-effective first step for many patients is a generic statin, taken consistently, perhaps at a higher dose. A company selling a newer therapy has limited reason to fund a program whose main effect is better statin adherence, even though that may deliver most of the population benefit.
This is why health systems should be wary of partnerships where the sponsor’s product is the only intervention on the pathway. And it is why companies that want to be taken seriously as partners should be willing to fund pathway work that benefits the whole class, including generics and competitors. That is a genuine cost. It is also what distinguishes a partner from a vendor with a nicer slide deck.
A worked example of the gap
Walk through what a target like 300,000 patients implies at the practice level. Divide it across several thousand general practices in England and each practice is being asked to find, counsel and inject somewhere in the tens of patients. That sounds small. But each patient needs a records search to identify them, a conversation to explain a new drug with outcome data still pending, a first injection, a second at three months and then one every six months. For a practice already struggling to offer routine appointments, that is several hundred clinical contacts over three years for a single drug, with no dedicated funding for the time.
Now look at the incentives inside the practice. The drug’s price was negotiated nationally, so the saving, if there is one, appears in a national budget. The workload appears in the practice’s appointment book. The GP who spends a morning on lipid clinics is not doing something else, and nothing in the agreement paid for that morning. On those terms, 4,549 prescriptions in nearly two years is not mysterious. It is what happens when the benefit of a deal is booked in one place and its cost in another.
The general lesson is that a partnership should be costed at the level where the work is done, not the level where the contract is signed.
Where capability partnerships earn their keep
The partnerships that tend to hold up share several features.
They fund a bottleneck the system names. If primary care lacks staff to run clinics, fund the clinic time. If records are unsearchable, fund the data work. The system, not the sponsor, should identify the constraint, because a sponsor that picks the bottleneck will tend to pick the one standing between it and its sales.
They measure outcomes, not scripts. Success should be defined as patients reaching treatment targets, or events avoided, not units dispensed. That aligns incentives and makes the partnership defensible in public. It also changes behaviour inside the company: a medical affairs team judged on patients at target will fund statin optimisation, while a commercial team judged on units will not.
They are transparent. Publishing the agreement’s purpose, funding and governance protects both sides. Health systems that sign opaque deals invite suspicion that clinical decisions were bought.
They have an exit. A partnership should say what happens when the funding ends. Programs that collapse the day the sponsor leaves were never integrated into the system. They were rented.
What the inclisiran case teaches
It would be unfair to read the inclisiran rollout as a simple failure. It was a new drug, a new delivery model and a new kind of agreement, launched while primary care was under intense post-pandemic pressure. Some of the caution from clinicians was entirely appropriate; outcome data were still pending.
But it illustrates a pattern. The agreement solved a problem the NHS had, which was access to a new therapy at an acceptable price. It did not fully solve the problems the front line had: time, capacity and clinical conviction. When a partnership is negotiated centrally and delivered locally, the local constraints decide the result.
The strongest counterargument
The best case against this argument is that it asks companies to behave like charities. A pharmaceutical firm has shareholders and a duty to sell its products. Funding generic statin optimisation, or pathway work that helps a competitor, is not a partnership obligation. It is a gift. Asking for it will simply shrink the number of companies willing to sign anything, and health systems will lose the discounts they do get.
There is something to this. A clear, honest commercial deal is better than a muddled one pretending to be something else, and a discount on a useful drug is worth having. Nobody should sneer at procurement.
But the objection misreads the commercial logic. A company that funds the full pathway is not only being generous. It is building the infrastructure that finds eligible patients, and patients who do not reach target on a statin are exactly the ones who may need the newer drug. A well-run lipid pathway is a better route to appropriate uptake than a price agreement stranded in overloaded surgeries. The self-interested case and the public-interest case point the same way more often than either side admits. Where they genuinely diverge, the honest answer is to sign a sales contract and call it one.
Pushing back on the cynics too
It is easy to be cynical about industry partnerships, and some cynicism is earned. But health systems also carry blame. Many will not share data, move slowly on governance and treat every industry conversation as a procurement risk. That drives companies back toward the simplest arrangement, a discount in exchange for volume, because it is the only one that can be signed within a year.
A health system that wants more than discounts has to be a more capable partner: clear about its priorities, able to share data safely and willing to publish what it agrees.
Second-order effects
The way these deals are structured shapes more than one drug’s uptake. A partnership that fails in general practice teaches clinicians to be wary of the next one, however well designed. Frontline trust is a shared resource, and a badly built deal spends it for everyone.
Opaque deals also distort the evidence base. If uptake is driven by a centrally negotiated agreement rather than clinical judgement, prescribing data become harder to read as a signal of what clinicians believe works. And if the system comes to depend on sponsor-funded staff, it may find that its own capacity to run pathways has atrophied by the time the funding stops.
What each actor should do
Health ministries and national bodies should publish a standard template for partnerships that sets out the named bottleneck, the outcome measures, the funding and the exit plan. A template lowers the transaction cost of doing it properly, which answers the complaint that capability deals take too long to sign.
Regional commissioners and hospital chief executives should insist that any agreement includes funding or capacity at the point of delivery. If the contract is signed nationally and the work falls on local clinics, the clinics need to see a line item.
Companies should move partnership accountability from commercial teams toward those who can be judged on population outcomes, and accept that some funded work will benefit generics. Founders building data and pathway services have an opening here, as the neutral party that both sides can pay to find patients and track outcomes without either side controlling the result.
Clinicians and their professional bodies should be involved before announcements, not after. The inclisiran caution from the RCGP and BMA arrived once the deal was public. A partnership that had tested its evidence case with them first might have met less resistance, or might have been designed differently.
What to watch
Over the next year or two, watch whether new industry agreements, in the NHS and elsewhere, publish outcome measures rather than volume targets. Watch whether any agreement funds clinical time in primary care explicitly, and whether its uptake curve looks different from inclisiran’s. Watch, too, how long-term outcome data for newer lipid-lowering drugs are received by professional bodies, because clinical conviction was one of the missing ingredients last time. And watch the language: the first time a major company announces a partnership whose headline metric is heart attacks avoided rather than patients treated, it will be worth reading the small print to see whether the incentives really changed.
The takeaway
Before signing a partnership, both sides should answer one question honestly: if this program succeeds and the partner’s product turns out not to be the main reason, will everyone still call it a success? If the answer is no, it is a sales agreement. That may be fine. Just call it what it is, and do not expect it to change care.
Have you seen an industry partnership that improved a population outcome, not just a product’s uptake? What made it work? We’d welcome your examples.
