By Nicola Oliver and Josephine Robertson
COVID-19 Actuaries Response Group – Learn. Share. Educate. Influence.
Around the globe we are seeing the delivery of vaccinations, in some countries at pace, in a bid to control (and hopefully end) the pandemic. It is not enough, however, that the vaccine is theoretically efficacious; it needs to get into arms to be effective.
Part of the vaccination decision-making process will include an individual’s access to information and understanding of randomized control trial efficacy and real-world evaluation of effectiveness.
Vaccination rates – take-up, hesitancy and behaviour
In our last bulletin on vaccines, we started to look at vaccination rates and the benefits of vaccination (in the box below). Randomised Control Trials on vaccine efficacy so far point to benefit b) but there is no evidence yet for benefits a) or c). To derive benefit d), higher vaccination take-up is better.
What causes people to be hesitant about vaccination take-up is complex. A recent study indicated appreciation of the following key issues:
- the potential collective benefit (benefit c and d)
- the likelihood of COVID-19 infection (benefit a)
- the effectiveness of a vaccine (benefit a, b, c and d)
- its side-effects (vaccine safety)
- the speed of vaccine development (vaccine safety)
Vaccine hesitancy fluidity
Vaccine hesitancy does not necessarily mean refusal and an individual’s choice may be revisited with time. As events occur, more information for or against key issues affecting decision-making will surface.
Below is a small sub-set of the whirlwind of dialogue in the media that may affect decision-making.
The ONS have undertaken ongoing online surveys which provide insights into the changing acceptability and so fluidity of vaccine hesitancy in the UK (Table 1). Though not directly comparable, vaccine acceptability was lower in alternative studies conducted before vaccine approval.
Caution should be taken when inferring meaning given the speed of information emerging that may affect vaccine decision-making. In addition, the ‘not yet offered a vaccine’ population changes over time because of the priority group system. The groups should (by design) correlate to elements of an individual’s perceived benefits from vaccination and so willingness to accept. For example, those at highest risk of COVID-19 and/or in greater need of health services may perceive greater benefits (b and d) than those at lower risk and/or in lower need.
From those unlikely to accept a vaccine (highlighted blue) and of their many reasons provided, some of the most common relate to vaccine safety and vaccine effectiveness (highlighted yellow). Real-world evaluation may with time provide evidence to allay worries and so improve vaccination take-up.
Diffusion theory is often applied in the adoption of new health interventions and in health promotion. This theory suggests that the population can be segmented into those willing to be the first to adopt (innovators then early adopters), those who will adopt later after seeing the first movers do so (early majority then late majority), and some who take longer (laggards).
Figure 1: Diffusion Theory on take-up of health innovations 
An individual’s vaccine decision-making is a choice between the perceived benefits of being vaccinated and risks of remaining unvaccinated. From this, vaccine hesitancy can be appreciated as a spectrum of willingness to accept risk. For the new COVID-19 health intervention of vaccination, the RCT volunteers are innovators; willing to accept a degree of risk to achieve the benefits of vaccination. Due to this willingness the trials were able to accumulate data on cases and controls to create safety and efficacy statistics. As the UK progresses in the initial stages of vaccination delivery, there are people willing to be vaccinated and so provide real-world evaluation statistics (early adopters). As delivery continues these individuals and the early majority provide evidence on those key issues affecting vaccine decision-making for the late majority.
We are grateful to the innovators and early adopters for providing us with the statistics herein and to come, such that we are able to communicate the benefits of vaccination.
This bulletin provides further information on the effectiveness of the vaccine by providing an overview of randomized control trial efficacy and the planned real-world evaluation.
Efficacy within randomized control trails
All vaccinations in the UK have to undergo an approval process and subsequent development of guidelines to assist clinicians in implementing a vaccination programme (see bulletin 98). The final stage of a clinical trial before approval, phase 3, is to test efficacy and it is this data that is presented to regulators. In general, phase 3 trials include many thousands of participants, randomised into two groups, half of whom receive the active treatment, and half a placebo (e.g. saline).
Vaccine efficacy is the percentage reduction in symptomatic COVID-19 in a group of people who received a vaccination in a clinical trial. It differs from vaccine effectiveness, which measures how well a vaccine works when given to people in the community outside of clinical trials.
|AstraZeneca/Oxford||23,848||64% after one dose
70% after two doses
Divergence seen at 14 days
Follow-up study suggests 76% after dose 1 to 90 days and higher efficacy after a delayed second dose
|Pfizer/BioNTech||43,448||95% after two doses
Divergence seen at 10 days
|Moderna||30,420||94% after two doses|
Effectiveness in real-world application
After the RCT, if a vaccine is approved for use, then a process referred to as surveillance begins. This ongoing monitoring is a standard process in the UK for vaccinations and enables decision-makers to take informed future choices on public health policy. Importantly, this is undertaken independently from vaccination manufacturing.
Public Health England announced on 11 January the proposed post-implementation surveillance plan for COVID-19 vaccines in the UK. This proposal includes monitoring vaccine uptake, vaccine effectiveness, population impact and vaccine safety. Surveillance will monitor the effectiveness of the vaccine in deriving benefits a), b), and c) and the duration of the protective effect which will in turn affect benefit d).
Effectiveness is estimated by comparing rates of disease in the vaccinated compared with the unvaccinated. Just as with RCTs there is a period between vaccination dose one, dose two and the ability to evaluate effectiveness. An example vaccination timeline for an individual is shown below (for AstraZeneca). However, not all are vaccinated on the same day or with the same vaccine so there is a need to accumulate a material vaccination rate within a group in order to assess the impact on population health (benefit a and b) and hospitalization rates (benefit d) with sufficient power.
As we have seen during the pandemic, there are different health outcome indicators which lag infection. Evaluations of effectiveness by each metric will therefore emerge at different (successively longer) times. PHE will be monitoring: asymptomatic and symptomatic disease using PCR tests, hospitalization, and mortality. This will provide evidence toward vaccination benefits b) and d) and provide evidence to later adopters on vaccine effectiveness.
However, the extent of all vaccination benefits is contingent on the duration of vaccine protection. PHE plans to assess effectiveness at 3-month intervals during 2021, 6-month intervals in 2022, and annually thereafter.
Importantly, PHE’s planned surveillance will evaluate the vaccine effectiveness on markers of infectiousness and transmissibility through viral load and culturable virus, and also on onwards person-to-person transmission. This will go beyond the evidence created through the RCTs to provide evidence toward vaccination benefits a) and c). Preventing infection would reduce transmission in society and increase the impact of the pharmaceutical relief during the pandemic. It will also provide an evidence base for public health policy decisions such as vaccine priority groups (at risk or those that transmit).
Expected variations between RCT and real-world
RCTs are studies undertaken in ideal conditions to improve the clarity of the efficacy ‘signal’. The trials, and therefore the results, often exclude people with underlying medical conditions, the immunosuppressed, children and the elderly. In addition, RCTs ensure strict adherence to protocol conditions on storage and delivery timeframes which cannot easily be followed in the real-world. We may therefore expect real-world effectiveness to vary from that seen in RCTs.
It is well accepted that vaccine effectiveness reduces with age and as the first priority groups are older, it should not be surprising to see a lower real-world effectiveness rate. This is primary vaccine failure, where an individual fails to develop an immune response. In addition, the rates following the second dose may well vary given the difference in RCT schedules (aligned to study protocols) and real-world delivery scheduling (aligned to population health outcomes). This should not be confused with secondary vaccine failure, which will also be monitored, where an individual develops an immune response but also the disease. These ‘breakthrough infections’ can be a signal of ‘enhanced disease’.
Vaccination safety within real-world application
Bulletin 98 provided information regarding the RCT safety and the approval process for real-world application. Safety is continuously monitored during surveillance to provide evidence of rarer or more delayed adverse events. Yellow card reports are used in the UK to report any adverse reactions. In addition, PHE will assess any increase in specific health service uses. Signals in these two surveillance areas trigger a rapid assessment to inform an epidemiological hypothesis for testing. This surveillance process will continue to provide evidence on vaccine safety.
The outcome of surveillance will support vaccine and non-pharmaceutical intervention policy recommendations in the continued pandemic response. It is important that real-world safety and effectiveness information is portrayed clearly and with care so as not to detrimentally influence people still making their vaccine decision.
Vaccination may provide a degree of unfounded confidence by those vaccinated, and so non-pharmaceutical interventions (hands, face, space) will still be required to keep R down.
The distribution of the population across different parts of the diffusion theory curve is not equal and vaccine hesitancy is known to cluster. In our next bulletin, we look at vaccine hesitancy causes, inequalities in take-up rates and the potential impact on health inequalities in society.
 https://www.gov.uk/government/publications/regulatory-approval-of-pfizer-biontech-vaccine-for-covid-19 and https://www.england.nhs.uk/2020/12/landmark-moment-as-first-nhs-patient-receives-covid-19-vaccination/ and https://www.express.co.uk/news/world/1373769/covid19-vaccine-pfizer-biontech-jab-allergic-reaction-alaska-uk and https://www.bbc.co.uk/news/health-55413666 and https://www.bbc.co.uk/news/health-55388846 and https://www.gov.uk/government/news/oxford-universityastrazeneca-covid-19-vaccine-approved and https://www.bbc.co.uk/news/uk-54763956 and https://www.gov.uk/government/news/first-people-to-receive-oxford-universityastrazeneca-covid-19-vaccine-today-4-january-2021 and https://www.gov.uk/government/news/moderna-vaccine-becomes-third-covid-19-vaccine-approved-by-uk-regulator and https://www.england.nhs.uk/2021/01/hundreds-of-thousands-invited-to-new-nhs-vaccination-centres / and https://www.bbc.co.uk/news/health-55734257 and https://www.theguardian.com/society/2021/jan/31/daily-record-as-600000-people-in-the-uk-receive-covid-jabs-on-saturday
 https://pubmed.ncbi.nlm.nih.gov/24366021/ and https://pubmed.ncbi.nlm.nih.gov/16955861/
A point from the above assessment
‘Expected variations between RCT and real-world’
‘It is well accepted that vaccine effectiveness reduces with age and as the first priority groups are older, it should not be surprising to see a lower real-world effectiveness rate. This is primary vaccine failure, where an individual fails to develop an immune response. etc…’
Being elderly/old  I found this an interesting comment which, presumably is based on scientific evidence The Actuaries have put that their reference was Appendix 20 [PHE approach to Surveillance] in which I was unable to find any mention of the above. However I’m not fussed, less so about their comment as it’s not unreasonable. Indeed I think I have read elsewhere that the WHO believe this also, presumably from recorded experience.
Thus, it was with interest that I read in today’s Metro [page 5] ‘Who knows best’ that the WHO rejected M. Macron’s claims that the ‘Oxford’ was ‘quasi-ineffectual in older people’. WHO have asserted ‘The immune response in people over 65 is almost the same as in younger people and this makes us confident this vaccine is protective. WHO also support UK’s decision to ignore the manufacturers’ three week duration since 12-weeks is sufficient [what happens if you can’t get the follow-up for,say, 15 weeks?] Perhaps medical knowledge currently is based more on political science than medical, especially as it is Politicians who usually fund such efforts.
It will be interesting to see if the Actuaries respond to the above in this week’s Issue.
Thanks Tim, I’ve passed your comment on to the Covid-19 actuaries. Vaccines are certainly getting political!