In the past month, two adenovirus-based vaccines against COVID-19, developed by AstraZeneca and Johnson & Johnson, have been temporarily suspended in 18 countries in Europe, South Africa, and the US pending a review on very rare incidence of specific type of blood clots. Most countries restarted them again soon after. I fear there is a potential for these hasty vaccine suspensions having led to more harm than good.
Flash judgements of caution vs risk-benefit
A crisis situation is difficult for all of us. Since we must react relatively quickly to rapid developments, there is a greater possibility for our natural risk biases to gain the upper hand – and this includes health officials and decision makers. We are all human, after all.
The wish to take time for careful deliberation of risks, in itself perfectly reasonable, immediately lead to decisions to suspend the vaccine while authorities deliberated. To my mind this is indicative of a natural quick and less measured reaction happening in tandem with the appropriate data-analysis. The natural reaction to be particularly averse of man-made, unfamiliar risks (like those from vaccine side-effects) easily overshadows our appreciation of natural ones (like the continuing spread and death toll from the pandemic). It feels more reassuring to actively take a step to avoid a risk (vaccination side-effect), while passively allowing the risk of disease to continue.
Recent opinion piece in the BMJ (British Medical Journal) also warns of this effect, called omission bias – it’s easier to remove an action we fear may cause harm, even when harm from inaction is greater:
Studies have linked this pattern, known as omission bias, to the sense of responsibility from the commission which intensifies the anticipation of regret. 
Defaulting to a “better safe than sorry” -approach, at least temporarily, is a reaction with a large emotional component – and I don’t call it emotional to belittle or ridicule, on the contrary. These feelings affect all of us and can be very hard to ignore. Public health professionals are not immune to such impulses and their responsibilities are likely to increase the pressure they feel. But health authorities should take particular care not to let natural biases in risk perception interfere with more objective guiding principles.
Ignored principles of risk management
The principle of “better safe than sorry” can be useful in some contexts, but it is only one of the many principles important in managing risk and risk communication. The European Center for Disease Control (ECDC) has a A literature review on effective risk
communication for the prevention and control of communicable diseases in Europe. It summarizes a handful of reviews, one of them a paper by Jardine et al and its 10 decision-making principles for risk management and communication. A precautionary “better safe than sorry” principle is one of them (number 7) – with the caveat that a cautious approach should be adopted “when faced with a potentially serous risk, even if the evidence is uncertain.” But hasty vaccine suspensions may be argued to ignore or go against several other principles outlined in the same review. I’ll mention four:
- 1 Do more good than harm (beneficence, nonmalificence). – The ultimate goal of good risk management is to prevent or minimize risk, or to “do good” as much as possible.
When a continued lack of protection is more serious, this precaution does not seem justified. It does not minimize risk.
- 3 Ensure an equitable distribution of risk (equity). – An equitable process of risk management would ensure fair outcomes and equal treatment of all concerned through an equal distribution of benefits and burdens.
Is it fair to expose the elderly to a continued risk of the disease, by putting them on hold instead of receiving the vaccine? The groups at greater risk from COVID-19 and in little risk of clots are not being treated in a way that would ensure lowered risk overall. They are also not allowed to participate autonomously in the decision making (principle 8).
- 4 Seek optimal use of limited risk management resources (utility). – Optimal risk management demands using limited resources where they will achieve the most risk reduction of overall benefit.
Many countries are only at the beginning of their vaccination process, others have not even begun and the pandemic is picking up. Managing the risks through other means are costly for the society and the individuals in many ways. We have no doses to waste or delay.
- 10 The complete elimination of risk is not possible (life is not risk free). – Risk is pervasive in our society, and cannot be totally eliminated despite an oft-expressed public desire for “zero risk”.
The choice to put vaccinations on hold seems to me to be covertly reflective of a spontaneous kind of thinking that we could choose to merely eliminate risks. That’s never true – there are always risks, both to action and inaction, to all the alternatives. As I and Dr Alison Bernstein wrote in our series: Zero Risk Is an Impossible Dream.
What is good risk communication?
The ECDC risk communication review also notes that:
…negative messages should be counterbalanced by a larger number of positive or solution-oriented messages  or, more specifically, ‘risk communications are most effective when they focus on what is being done rather than what is not being done’
It seems to me that a focus on reducing overall risk, while being proactive about surveillance and communication about the possible side-effects of vaccinations would be the best scenario here. Pausing vaccinations is a strong negative message against our best solutions, and in the midst of a pandemic, should not be done lightly.
It has not been suggested that the potentially heightened clotting risk would approach the current risk of continued exposure to an actively spreading pandemic for the general public. There would have been many ways to communicate the potential increase in risk transparently, while also acknowledging and underlining the responsibility to reduce the greater harms from the disease itself by continuing vaccinations. Recommendations could have been made to the target group (say, young healthy women) to consider waiting, or choosing a different vaccine, while allowing the old and the more-at-risk to be vaccinated. The suspensions in Europe came at a time when mostly the elderly and at-risk populations were being granted access to the vaccine.
To borrow from another set of risk communication guidelines, A practical guide to public risk communication (from the Risk and Regulation Advisory Council, 2009, linked here among other good sources):
Understanding a public risk typically involves considering the range of benefits and costs associated with it. All aspects of the risk need to be thought through and explained, or the dialogue about the risk may become dominated by one particular part of the story – often a real or perceived cost.
The cost-benefit in pausing vaccinations was not there. Only the precautionary part was. By inflating a particular part of the story, we risk endangering the public’s trust in vaccines while prolonging the pandemic. The European Medicines Agency has maintained this since the beginning of the investigation:
As previously stated, while its investigation is ongoing, EMA currently remains of the view that the benefits of the AstraZeneca vaccine in preventing COVID-19, with its associated risk of hospitalisation and death, outweigh the risks of side effects.
With update a few days later, COVID-19 Vaccine AstraZeneca: benefits still outweigh the risks despite possible link to rare blood clots with low blood platelets.
I wish the authorities would quickly have gone out with what we know about the sizes of the possible risks and their target groups, while continuing on a path of minimizing harm. The WHO Europe office’s guide to managing the communications response during Vaccine Safety Events recommends that while waiting for data in an evolving situation officials should:
Remind the media of the high value of the immunization programme — provide statistics of the national situation regarding, for example, coverage and the estimated lives saved and illness averted.
Here is a comparison of risks for different age groups and levels of pandemic spread (low, medium, high) that I found to be a good example of a clear communication of the risks, from Winton Centre for Risk and Evidence Communication at the University of Cambridge. Benefits and potential harms in situation of highly active spread in the population:
A comment on the choice of comparison, also elaborated on in their article:
Avoidance of ICU admission was chosen as a benefit comparator because the potential harms being illustrated are equally severe.
Of course for every one of these potential ICU admissions there are many many people who might have had hospitalisation or long COVID.
This is the scenario during medium exposure:
Only in the case of low exposure (below) would the benefits/harms approach each other, and even then only for the youngest age group. For countries with low spread, recommending the young to wait or to choose a different vaccine could be a reasonable choice.
In US where the share of J&J vaccines was only around 5%, which plentiful access to Pfister and Moderna vaccines would easily compensate for, halting the vaccine may not be considered as risky – it won’t delay the vaccination process per se. But the effect of the message it sends to the public is still likely to be net negative. As Steven Novella writes in Science Based Medicine:
…the antivaccine forces on social media would exploit the situation to provoke as much fear and doubt about the vaccines in general as they can. There is no decision that would prevent this, so you might as well do what is best scientifically and then just explain the decision as best you can.
Whether there is damage done to public perception on (COVID-19) vaccine safety in general after suspension of one kind of vaccine in particular may be harder to predict, but a diminished trust in the vaccine in question can hardly be a surprise to anyone. These perceptions don’t necessarily remain within regions where other vaccine choices are abundant. In poorer countries in particular, trust in the vaccines which are easier to store (like the AsraZeneca and J&J ones) may be crucial for putting a stop to the spread.
Health authorities may tell themselves that they know the vaccine is very safe, and their actions are only to optimize the risk-benefit to a greater degree. But the hasty counter-reactions of the authorities are most certainly multiplied in the minds of the public. From a 2017 Australian study on the effects of vaccine scares:
Findings indicated that, for those who recalled the vaccine suspension, there was a lasting sense of uncertainty and confusion and a perceived lack of information.
The communication contained in the action itself, the suspension of the vaccine, gives us a definite feeling of unease which is can be difficult to repair with messages later on that try to convince us that there is no reason for concern. As the BMJ opinion piece reads:
Restoring confidence after the covid-19 Oxford/AstraZeneca vaccine pause will be an uphill battle – The impacts on public confidence are likely to be long lasting and experienced globally
This is also why I thought it apt when the Washington Post described the European response as “chaotically cautious” in their piece Europe’s ultra-cautious vaccine strategy is itself a risky bet. They write:
But the toll has already been high for Europe either way, with expired doses, surging cases and weakened morale. […] In Europe, the biggest problem is not recklessness, but cautiousness. […]
“It was a political choice,” Nicola Magrini, the director of Italy’s Medicines Agency, told La Repubblica newspaper Monday. Leaders saw other countries pausing the drug and made the decision to do so as well, rather than listen to global public health bodies.
There seems to be an element of fear baked into what the authorities would like to believe is merely a rational degree of caution. They may worry what the public may think if the authorities don’t appear decisive and protective enough. There have certainly been many academic discussions of the continuing misuse of the precautionary principle in Europe, used in a way that has more to with “media, stakeholder and member state pressures, rather than scientific evidence per se.” Scientists have gone as far as to argue that:
The evidence summarized … has demonstrated that the precautionary principle holds back technology, innovation, incomes, environmental improvements, and health benefits, while increasing trade disruptions, risks and human suffering. The precautionary principle is a tried but has failed as a risk management strategy. It is time to move beyond it. (Marchant, Al.Felsot, and Griffin 2013, 16)For a detailed discussion, see Looking back and going forward: what should the new European Commission do in order to promote evidence-based policy-making? – Ragnar Lofstedt & Anne Katrin Schlag, 2016
I do not envy the job of health authorities with the task to react in ways that are both fast, measured, transparent, and honest. I understand that risk biases are very central and automatic tenets of human psychology, and that they play in to the pressures of how to reassure the public. But I also think that we need to demand a high degree of responsibility from our institutions that take decisions on public health measures. This added responsibility also means avoiding the dangers of exaggerated caution. Reacting more strongly to man-made, unfamiliar threats or opting for an active avoidance of a single threat while overlooking the larger context may reassure us on an emotional level, whether we are health professionals or not, but when it sways the measures taken by authorities, it can cause harm on the very real and tangible level of public health.
If you are interested, you can find my other pieces on this topic under Vaccines and Health. If you would like to have a discussion in the comments below, please take note of my Commenting policy. In a nutshell:
1. Be respectful.
2. Back up your claims with evidence.
You show evidence that suspending vaccinations is wrong by orders of magnitude.
Deaths from Covid are “natural” events from diverse causes. Deaths from vaccination are clearly due to actions of the medical/governing profession, to whom blame will be assigned.
An engineering approach would be Bayesian; make the best decision now based on the best available evidence now; with new information change actions.