What does global evidence tell us about the optimal way to distribute COVID-19 vaccines?

COVID-19 vaccines

The International Network for Government Science Advice has been monitoring different countries’ prioritisation models for the vaccine rollout – from the arts and culture sector in Serbia to indigenous populations in North America and New Zealand, and politicians in Turkey

Kristiann Allen and Tatjana Buklijas

After a year of battling the COVID-19 pandemic with only behavioural and policy tools at our disposal, the regulatory approval and rollout of first-generation vaccines is an historic turning point. Global attention is now focused on vaccine distribution strategies to achieve herd immunity, estimated to be around 70% of the population.

Yet the global vaccine supply remains insufficient, requiring countries to prioritise certain populations and phase their distribution efforts. This situation highlights the significant policy complexity in balancing multiple competing imperatives.

The World Health Organization (WHO) anticipated this problem last September when its Strategic Advisory Group of Experts on Vaccines, Immunization and Biologicals issued high-level guidance for allocating vaccines between countries and in priority populations within counties. Its prioritisation roadmap is underpinned by a values framework which highlights:

  • human wellbeing;
  • equal moral status;
  • global and national equity;
  • reciprocity (protecting the protectors); and
  • legitimacy (making decisions transparently).

But the WHO also anticipated the complexities of weighing and balancing these values in the context of incomplete science, stating that it would issue vaccine-specific guidance as each product became available, which it has now begun to do.

A more sophisticated approach to prioritisation

Now a new research article by Kate Bubar and colleagues has proposed a more sophisticated approach to prioritisation. Building on the work of others, they use mathematical modelling to stratify age groups and integrate variables such as rates of transmission, the pharmacological efficacy of the vaccine, supply limitations, roll-out rates, and characteristics of the target population to help optimise prioritisation decisions for greatest impact in protecting the vulnerable (for which they use the admittedly simplistic proxy of age).

Their model suggests that to reduce mortality, prioritising older (60+) adults is a robust strategy in general – however, specific local conditions may prompt ‘dynamic updating’ of priorities. For instance, they found three situations that would justify prioritising younger adults (20-49):

  • when there are other transmission-blocking interventions and vaccines are 80+% effective;
  • where rollout is at most 0.2% daily or supply is at most 25% of the population, but there are other transmission blocking interventions; and
  • where available vaccines are of low efficacy in older adults (60+).

It goes without saying that the sophistication and dynamic responsiveness of such recent modelling efforts are extremely useful to assist rapid decision-making. However, in doing so, the researchers assume that it is no more than a matter of choosing between two main approaches to prioritisation: (i) directly vaccinate those at highest risk for severe outcomes; and (ii) protect them indirectly by vaccinating those who do the most transmitting.

Yet this classic binary choice is not as straightforward as it may seem. Instead, as with all policy decisions – and especially in the context of resource constraints – choices derive from the fundamental policy aim.

What is the key policy aim?

Whereas the public health aim is to reduce the mortality and the burden of disease, the health policy aim may be to protect health services from being overwhelmed (especially where these are publicly funded), while the government’s overarching policy goal is very likely to ‘return the country to normality’ as soon as possible.

Privileging one or another of these aims means adopting a slightly different vaccine prioritisation strategy, even if we can all agree with the WHO’s roadmap and reciprocity principle that frontline health and safety workers should be vaccinated first. But who is next for the limited supply? Do we prioritise those at risk of severe outcomes, those at risk of transmitting, or those we need most to get the economy and society functioning again?

And while improving speed of vaccine roll-out is non-negotiable in countries where the numbers of cases are high and health services under pressure, those with low case numbers (thanks to successfully implementing a strategy of elimination supported by stringent border closure policies) are facing complex choices. Should they be good global citizens and let countries with high death tolls have priority in vaccine supply? Or should they too insist on speed in the hope of opening their borders and rescuing those industries, such as tourism, that have been devastated by the pandemic’s preventative policies?

Different countries prioritise different communities

Ethical frameworks such as those of the WHO or jurisdictions such as Ontario, Canada and the US state of California’s (based largely on the US National Academies framework), provide moral and scientific orientation but leave a gap with respect to such broader policy choices and implementation issues. The US Centers for Disease Control addresses this gap by explicitly considering the separate goals of mortality and morbidity prevention with protecting ‘societal functioning’ in its phased approach.

Yet even striking this balance is not obvious, because what counts as ‘societal’ functioning and who is considered ‘vulnerable’ will each be highly context dependent. For example, whereas some jurisdictions might see transportation workers high on their list, Serbia has prioritised the arts and culture sector by immunising actors so that live theatre can reopen.

Similarly, the definition of a ‘vulnerable’ population can be constructed from multiple different angles and according to context. While the physiological vulnerability of age and co-morbidity is uncontroversial, we know that pandemics thrive on social vulnerabilities, which exist for multiple and interacting structural and historical reasons. In settler states, this has meant prioritising indigenous populations, as the federal government of Canada has done, while the US state of Alaska has overlooked the gaps in vaccine testing regimes to prioritise indigenous youth.

In New Zealand, which has successfully pursued a COVID-19 elimination strategy and where the vaccine rollout began in mid-February with the immunisation of its border personnel, calls have been made to prioritise the indigenous Māori population. Yet the real centre of vulnerability appears to be the immigrant Polynesian population concentrated in the country’s largest city, Auckland. This group suffers the interacting burdens of poorer health, socioeconomic deprivation and workplace exposure, while also reporting higher levels of vaccine hesitancy. With the latter threatening to impede vaccination efforts globally and even where the supply is relatively good, some countries such as Turkey have decided to prioritise immunisation of groups with high visibility and authority – namely politicians – in the hope of alleviating public concerns.

Some countries have had to reset their priorities

As with all complex policy problems, even in the context of perfectly equitable and scientifically supported vaccine distribution, implementation challenges can be significant.  But when the challenges are created by the distribution protocol itself, a serious reset is needed. Such was the case in California, where a highly sophisticated distribution framework risked leaving vaccine doses to spoil when the precisely targeted individuals were not immediately accessible. Ultimately, a reset was needed under new management.

A similar adjustment was required in Belgium recently, which had originally made the decision (along with France, Germany, Poland and Italy) to set priorities among all but the elderly because the AstraZeneca vaccine had insufficient data for over-55s. Belgium’s policy about-face was forced by reports of the resulting ‘untouched stockpiles’ and public demands to speed up the rollout and not waste doses. In Germany, similar issues are prompting some to rethink its system of priority groups across the population, and instead focus on border hotspots such as the southern states.

Allocating limited resources in any situation is challenging. But when it is a matter of life or death, with both economies and people on life support, it is a challenge of a different order altogether. While many people are, understandably, anxious to know whether their country has chosen the right approach, the only thing that can be measured reliably now is the speed of initial rollout. In this, the investment into vaccine development and production (US), efficient approval process (UK), and diversified approach to acquisition (Chile) appear to be important elements.

Yet with no country approaching the 70% herd immunity target so far, the effects of different prioritisation choices (that is, the impact of different allocation strategies upon transmission, morbidity and mortality rates) remain unknown. And for the moment at least, the same is true for the pressures faced by health services and dormant economies.

Kristiann Allen and Dr Tatjana Buklijas lead the International Network for Government Science Advice (INGSA) Secretariat, based in the University of Auckland’s Centre for Informed Futures