How will we know about long term side effects of vaccines: learnings from thalidomide, swine flu and Donald Rumsfeld?

Now that a vaccine has been licensed for use, with regulators (at least in the UK) satisfied about the safety of the Pfizer vaccine; many are asking, especially given the novelty of the approach and the rapidity of the path from discovery to use, how can we be sure the new vaccines  are free from long term side effects?  In this post I consider some of the lessons from history and describe the different approaches that will provide answers to address this concern.

Before addressing this topic keep in mind the two issues in the table below

Historically vaccines are generally safe

  • There are very few instances of vaccines being linked to long term serious illnesses and this is reassuring 
  • I have scoured the medical literature and have come up with these two examples of possible problems from mass flu vaccination programmes


  • Guillain Barré is a rare paralysing neurological disorder (which can be a complication of flu).  The ‘jury is out’ on whether the cases that were seen following that vaccine campaign were due to the vaccine or happened by chance.
  • I was involved in looking at the data on narcolepsy and H1N1 with the GSK vaccine Pandemrix.  To be honest I am still not convinced that the data proves the association was real though the government agreed to pay compensation
  • Rumsfeld’s oft quoted statement about ‘unknown unknowns’ is relevant to drug vaccine safety and underpins the rest of this post
  • The challenge is that if there are no clues as what might be the specific diseases or problems that result from a vaccine, then we are not sure what we are looking for
  • Let me give an example:
  • It is biologically plausible that vaccines could lead to multiple sclerosis*
  • The truth is this has been investigated and the answer thus far, based on a systematic review (ie a careful analysis of  at every research article)  is that they don’t
  • The issue is that without doing the necessary detailed studies, that question could not have been answered.
  • We cannot do studies though for very possible disease that just may be a consequence of a vaccine.

*I am only using this as an example to make the point, I am not aware of any data to suggest that any of the Covid-19 vaccines have been linked with multiple sclerosis!

The need for a comparison group

  • It is impossible to interpret data that might suggest for example a rate of multiple sclerosis in people who are vaccinated of (say) 1/5000 without knowing what the rate of multiple sclerosis is in the population who are not vaccinated
  • But this is not easy, especially if the aim is to vaccinate everyone: who would be left?
  • Such a sample of unvaccinated people might be different in other ways from people who are vaccinated which makes the comparison unfair (or biased) 

So how do we do it! 

  • There are actually 4 different ways of investigating long-term safety of vaccines in the past, and will no doubt be used again now for assessing the longer term safety of the Covid-19 vaccines (I have been involved in all these activities with different products!).  
  • The different ways are:
    1. Long term follow-up of the participants in the trials
    2. A centralised reporting system for unexpected health problems
    3. Linking national datasets
    4. Bespoke in depth epidemiological research study 
  • In the paragraphs below, I give the advantages and disadvantages of the different approaches

Long term follow-up of the participants in the trials

  • Fairly obvious but in the trials of the Covid-19  vaccines we have been hearing about, the participants will be now also followed up regularly for two years and asked to report any new health problem
  • Advantages are that this will be easy data to collect and analyse
  • Disadvantages are:
    • The participants in the trials are not representative of the general population (as mentioned in my previous blogpost)
    • The numbers in the trials are too small to produce robust data on diseases that are rare (but important)

Centralised reporting of unexpected health problems

  • The unexpected occurrence of babies born with limb defects in mothers who had taken the drug, thalidomide, in the 1960’s led to the recognition that a system was needed to collect these reports centrally
  • The UK started its ‘yellow card’ system in 1964 and the USA its MedWatch system as late as 1993
  • Both these systems are based on an individual report to medicine regulators of an unexpected occurrence of a problem following the introduction of a drug or vaccine 
  • In the UK anyone can fill in a report, this can be the patient, a physician, a pharmacist or whoever
  •  (For the UK readers if you are vaccinated and think you have developed a long term problem you can report it online without needing anyone’s permission!)
  • The advantage is that these systems can receive reports of anything and hence can pick up unexpected issues
  • The disadvantages are obvious:
    • The systems rely on voluntary reporting of a problem 
    • There is no requirement for proof of  ‘attribution’, by this I mean the reports are of a health problem that is considered to be ‘unexpected’ but this does not mean that it is vaccine related.  This is more challenging the longer the time interval from vaccination.  There is an inevitable arbitrariness – consider: 
      • A very unusual rash within a few weeks of being vaccinated is likely to be reported
      • A bad kidney infection 8 months after vaccination may or may not be reported
      • Multiple sclerosis symptoms starting 3 years after vaccination are unlikely to be reported 
  • Despite these shortcomings medicine regulators like the system as it can alert them to a problem after just a few cases are reported which were unexpected.  Such a ‘cluster’ can then act as a trigger for more formal assessment by the vaccine company
  • There is also a legal responsibility that any report of a side effect to the vaccine that manufacturers become aware of are passed onto the government regulators
  • FYI I use to sit on the Committee that received these reports and each month we would consider if the reports coming in should trigger a more formal enquiry.  In this meeting from 2014 we considered a nasal flu vaccine:

Linking National Datasets

  • This approach is useful for following up on a lead of a possible serious side effect from the yellow card or similar system as above
  • This may be done by the companies at the request of regulators or be initiated by university researchers 
  • National health record databases are searched linking vaccine history and then the subsequent development of the possible side effect.
  • Data are then analyses comparing the rates of the possible side effect between those who were vaccinated and those who were not
  • The advantage is that this can be done very quickly on very large numbers in the population
  • The disadvantage is that the quality and completeness of the data in national datasets on both vaccine history and subsequent health state may be not very good depending on the disease 

Bespoke in depth epidemiological research study

  • Independent epidemiological research can be planned aiming to collect very  accurate data on
    • On vaccine history
    • Diagnosis of any complications
    • Also they include an appropriate comparison group
  • It can go disastrously wrong.  The now debunked studies on MMR vaccine causing autism had a major public impact and resulted in many children and adults developing and in some cases dying from the complications of measles 
  • Although as a researcher I support the idea of these detailed projects, the studies are really difficult, time consuming and expensive to do
  • Less than perfect studies may get published and publicised with conclusions are not justified!
  • As a result, most safety studies are now done using national databases 


  • It is not possible to guarantee any vaccine is free of any possible long-term complication 
  • The history of vaccine safety is a good one and that is reassuring
  • We can only study a question that we know about, so Donald Rumsfeld has a point!

If you would like to receive email notifications of new posts on this blog just click this link and enter your email at the bottom of the home page:

3 replies on “How will we know about long term side effects of vaccines: learnings from thalidomide, swine flu and Donald Rumsfeld?”

Hi Alan
Epidemiological evidence may emerge from death certification. For example, an increased ‘death association’ or cause emerging following the date of mass immunisation might be a pointer to a smaller, discrete retrospective study. Alternatively a local rise in unexpected morbidity.
Until such time we can believe well meaning regulators and non-reassuring scientifically naive politicians.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s