Recommendations under empirical uncertainty

The work of the ethics advisory group was based on the evidence available in the autumn of 2020. With the assistance of the other groups under the Norwegian Institute of Public Health, the ethics advisory group had access to up-to-date information on the COVID-19 vaccines, the risk group, infection risk within occupational groups, and logistical challenges involving the vaccine candidates that were on the way. Meanwhile, however, information concerning a number of other factors was scarce at that stage of the pandemic. There was, for instance, uncertainty concerning the efficacy of the vaccines, their potential adverse effects, which groups the different vaccine candidates would have been trialled in and approved for, public uptake of the vaccines and, not least, what course the pandemic would take.

The expected efficacy of the vaccines became a key factor. The vaccine candidates in question proved effective in reducing morbidity and mortality, while their effectiveness for infection control remained unclear (4). The evidence as it existed in November 2020 indicated that the vaccine would have the most benefit as a means of preventing severe illness and fatalities as opposed to preventing community transmission.

Values, goals, and priority groups

To develop an ‘ethics framework’, in the sense of a set of principles and values to inform policymaking, the ethics advisory group largely based the defining values on those otherwise applied in Norwegian healthcare. This is discussed in several Norwegian official reports (5–6). Further, Norway has a long-standing tradition of transparency in its priority setting in health care, and both the national legislation and official guidelines are drawn up in consultation with priority setting committees (7–11). Three formal criteria are applied in guiding priorities in the Norwegian public health service: The health-benefit criterion, the resource criterion, and the severity criterion.

Preventive public health measures cannot generally be compared with prioritising individualised and (potentially) curative healthcare. For this reason, a recent Norwegian priority setting white paper recommended that priority setting of public health measures be addressed as a matter in its own right (6,12). Indirect health benefits are important: a vaccine benefits others in addition to the vaccinee in that it reduces community transmission. In the programming documents for severe communicable diseases, such as pandemic influenza, for example, several prioritisation factors are identified: The main premise is that the risk group, practising healthcare personnel and defined key individuals should be given priority over the general population (13).

The ethics advisory group also consulted international literature on priority setting of vaccines. A number of publications have documented different countries’ priority setting recommendations (3). In particular, the WHO SAGE values framework for prioritization of COVID-19 vaccination is intended as decision support for national strategies (14).

The ethics advisory group advocated that the Norwegian National Corona­virus Immunisation Programme should be predicated on the following values: equal respect, welfare, equity, trust, and legitimacy. In the Norwegian context, the principle of equal respect is a norm of particular relevance, being linked both to official standards of non-discrimination and needs-based equitable access to healthcare, and to the goal of reducing social inequalities in health. Further, the health-benefit value is focal: Scarce health resources should be distributed to maximise their benefit and increase overall national (health-related) welfare. At the same time, procedural considerations are important. The Norwegian National Coronavirus Immunisation Programme should be founded on trust between the public and the authorities, and the allocation of vaccines should be defensible and justifiable with respect for the entire population, including those who do not receive the vaccine (15).

Based on this, the ethics advisory group proposed the following goals for the Norwegian National Coronavirus Immunisation Programme: reduce the risk of severe illness and death, maintain essential services and critical infrastructure, protect employment and the economy, and reopen society.

Vaccinate to gain life years or save lives?

The most common approach in healthcare priority-setting is to quantify the benefit of an intervention in terms of quality-adjusted life years (QALYs). For priority setting of COVID-19 vaccines, however, QALYs are not a viable option due to the scarcity of evidence-based data.

The ethics advisory group discussed whether the vaccine should be aimed at gaining life-years or preventing fatalities. All things being equal, many people will assert that it is more important to prevent deaths among young individuals than among elderly persons. This then suggests life years as a reasonable measure. However, the infection fatality ratio, indicating the risk of dying if infected, is extremely low among the young healthy individuals in Norway but high among the oldest. The large age-dependent difference in the infection fatality ratio implies that it is reasonable to assume that prioritising life or life years will not constitute a significant difference when weighing up benefit^{* See also Ole Frithjof Norheim’s example calculation. This calculation shows that the choice between life or life years has no material significance in the main priority setting strategy for the risk group versus healthcare personnel: “Slik valgte vi hvem som får vaksinen først”[How we decided who should get the vaccine first] Morgenbladet 04/01/21. https://morgenbladet.no/aktuelt-ideer/2020/12/slik-valgte-vi-hvem-som-far-vaksinen-forst (Last accessed 01/03/21).} : older individuals should be given priority over younger individuals both if life years are to be gained and if fatalities are to be prevented by immunising large groups in society.

The ethics advisory group was divided on the question of whether to use life years or life as a measure but ultimately settled on life – with the key assumption that life years and life would result in a fairly equal prioritisation of groups. Life is a transparent measure in the sense that it is easily understood. Life is also helpful for data modelling. Life years, on the other hand, can be modelled as a measure in several ways, and it could be argued that the life-years goal embodies a number of theoretical value assumptions (such as the quantitative norm for lost QALYs defined for each age group). At the same time, the ethics advisory group proposed that a switch might be made to life years gained if it subsequently emerged that this measure yielded significantly greater health-benefit value from the Norwegian National Coronavirus Immunisation Programme.

These assessments were made in light of the evidence available at that time. The ethics advisory group had no evidence-based data on vaccine efficacy for reducing viral transmission and concluded that case fatalities could best be reduced by vaccinating the risk group first.

Should healthcare personnel be given first priority?

Some have advocated giving healthcare personnel first priority for the COVID-19 vaccine, regardless of risk profile, the overall infection situation and the infection control efficacy of the COVID-19 vaccines. In principle, this could be implemented in several ways. For example, all or a selected proportion of the healthcare personnel group could take lexical priority, i.e., come before the first individual in the risk group. Alternatively, healthcare personnel and the risk group could be prioritised concurrently in (at least) two ways: either by ensuring that healthcare personnel are guaranteed a larger share of the vaccine doses or by healthcare personnel (smaller group) having a greater likelihood of being vaccinated than people in the risk group (larger group).

There are many reasons for prioritising healthcare personnel. One argument is reciprocity: society has called on certain occupational groups – such as frontline healthcare personnel – to expose themselves and hence their immediate relatives to infection risk.^{* A reciprocity argument may in principle apply to any group in society exposed to a significant risk of infection, including shop assistants, hospital laundry workers or taxi drivers. Preliminary figures for Norway (for the year 2020) show that cleaners (1.66%) and paramedics (1.71%) were the groups with the highest incidence, excluding those infected abroad. As expected, nurses (1.55%) and doctors (1.39%) had a somewhat higher incidence (excluding those infected abroad) than the Norwegian general population of working age (1.11%). However, health workers such as clinical psychologists (0.69%) and physiotherapists (0.89%) had a lower incidence than the general population (16).} The reciprocity argument dictates that those who run an extra risk are entitled to extra protection. Those who ‘give’ something extra deserve something extra in return. None of the three Norwegian priority criteria – health-benefit, resource, or severity – justify, even in principle, that the reciprocity argument should carry any weight. Meanwhile, it could be argued that frontline healthcare personnel in frequent contact with patients should be protected against communicable disease on the grounds of health-benefit alone – to at-risk patients, to themselves and to their immediate relatives. Moreover, there is justification for prioritising those who would otherwise be difficult to replace if they succumb to illness or have to quarantine: as ‘insurance’ to prevent the collapse of the health service during the pandemic. These last-named factors are covered by the principles of the health-benefit criterion and were emphasised by the ethics advisory group.

Should geographical priority setting be rejected?

The equal respect principle has a strong standing in the Norwegian health service and also includes a commitment to geographical equal respect. The Patient and User Rights Act is intended to ensure equal access to high-quality services. Nevertheless, the ethics advisory group judged that geographical factors could be weighted in order to prioritise parts of Norway with a particularly high and persistent transmission rate, in order to reduce both the risk of disease in the individual and the risk of overloading the health service and in order to reduce the need for other infection control measures. Meanwhile, the group responsible for the socioeconomic cost-effectiveness of COVID-19 immunisation concluded: “If we do not prioritise the population in parts of the country that are most heavily affected, costly infection control measures will need to remain in place for longer than strictly necessary, with implications for the whole country’s economy” (18: 5).

The ethics advisory group advocated that geographical prioritisation should be considered under certain conditions. ^{* One issue that was scarcely addressed concerned who should actually benefit from geographical priority setting. Such priority setting would presumably target the risk group (i.e., reduce morbidity and mortality), healthcare personnel (i.e., guarantee capacity) and epidemiological interventions to control infection.} On 2 March 2021, the Norwegian Institute of Public Health recommended a slight geographical bias.^{* For some time, the Norwegian Institute of Public Health judged that the conditions for geographical targeting were not met: on the one hand, because the course of the pandemic is unpredictable and the effects of the vaccine are not achieved until several weeks after it is administered, and on the other hand because the Institute’s advanced data modelling indicated that vaccine benefit is greatest in geographical regions with either a particularly low or particularly high transmission rate (4). The last-named finding no doubt seems counterintuitive if the assumption is that the vaccine has the most benefit in regions where the transmission rate is highest. Increasing use of and confidence in the findings of advanced data modelling makes the priority setting debate less transparent.}

The assessment of geographical prioritisation exemplifies how empirical uncertainty impacts the trade-off between benefit and equal respect: if the benefit of geographical prioritisation is unequivocal, geographically equal access to the vaccine will have to cede.