Isn’t it curious how describing deaths brought forward by a short interval by covid as dry tinder garners disapproval, while the MHRA’s breezy description of Yellow Card deaths associated with covid vaccination as spontaneous combustion — most likely would have died anyway, dear boy — goes quite un-remarked. Be that as it may, Dr No has taken a look at the MHRA’s Yellow Card Scheme for reporting adverse drug reactions, after some discussion in a recent post about a BMJ Rapid Response which suggested significant underreporting of covid vaccination side-effects. What can the Yellow Card Scheme tell us about adverse reactions to the covid vaccines? Is there any evidence for or against significant under-reporting of adverse reactions?

The first, and probably most important, thing about the Yellow Card Scheme (YCS) is that it is voluntary. There is no absolute requirement to report suspected side effects, just a sort of obligation. Anyone can make a report (which can list more than one reaction), either by using one of the traditional yellow cards which gave the scheme its name, or more recently online. The reports are not that onerous to complete, but at least three things have to happen for a report to get filed. First, someone has to suspect an adverse reaction to a drug (the definitions of both adverse reaction and drug are broad), second, that someone needs to know that the YCS exists, and third, that someone, or someone on their behalf, has to take the time to file the report.

In the ordinary course of events, making a link between a drug and a suspected adverse reaction is by no means a given. In an otherwise well patient given a new drug who then develops a rare condition, the possibility of causal link might well spring to mind, but in the majority of otherwise unwell patients, on a raft of drugs, who sometime later develops yet another condition, the suspicion of a causal link can all too easily pass unnoticed. Nonetheless, the covid vaccine is a new drug, and many receiving it will be otherwise well, so the odds of meeting the first requirement, that a suspected link between the vaccine and a reaction exists, are probably on balance weighted in favour of spotting a suspected link, should an adverse reaction occur, but this will diminish in patients with complex co-morbidities.

The odds on the second requirement, that someone needs to know the YCS exists, are less favourable. Doctors certainly know about the YCS, but do those receiving the vaccine? Anecdotal reports (the best we have at the moment) suggest many do not, and it goes without saying that if you do not know about the YCS, then you are not going to use it. Many suspected adverse reactions, despite being noticed, will remain unreported. The third requirement, that someone takes the time to make the report, also contributes to under-reporting. Put bluntly, the yellow card is just another form, and a voluntary one at that, that one should fill in — and the road to hell is paved with good intentions.

For these reasons, the YCS in the ordinary course of events is severely handicapped. In November last year, the MHRA crowed about reaching a “Major milestone – 1 million Yellow Card reports since scheme started over 50 years ago”. This sounds impressive, but needs some context. Cobbling together some community prescription data for England (hospital data is not available, and the other UK nations have less long term data) over the last 50 years from here, here and here, and doing some interpolation for missing years, there have been some 30,456,890,537 (30 billion) prescription items dispensed over the same period. That works out at around one report for every 30,000 items dispensed, or put another way, only 0.0033% of community items dispensed generate a yellow card. Given that adverse reactions are hardly rare, these are not encouraging figures, and because the number of items dispensed does not include hospital prescriptions, the true rates will be even lower. The cards may be yellow, but faces at the MHRA must surely be red.

Or maybe not, at least for the covid vaccines. In 2020, the MHRA received a total of 40,924 reports of suspected adverse reactions. In contrast, it received a total of 160,071 reports from 34,100,000 vaccine doses administered up to the 28th March 2021. The extra 34,100,00 prescriptions generated by the vaccine are but a drop in the ocean compared to the over 1 billion community items dispensed annually in recent years, and yet the reporting rate is all but four times as many reports from the vaccines alone in just 3-4 months, than all reports for any drug for all of 2020. Put another way, around 0.47% (one in just over 200) of all vaccines doses administered generate a yellow card, compared to around 0.0033% (one in 30,000) for the Yellow Card Scheme over its lifetime. These figures represent orders of magnitude differences, and are not to be sniffed at.

We may not sniff, but we might cough. There may be orders of magnitude more reports, but we still don’t know whether they represent over, under or about right reporting. If they are about right, do the numbers mean the covid vaccines have high rates of side effects, or are they just being reported more? Or could the startling rise in Yellow Card use, and the data it provides, somehow amount to a green light, telling us that the vaccines are indeed safe? Unfortunately, the Yellow Card Scheme, which is intended primarily as an alert system, can’t itself answer these questions. To get those answers, we need properly conducted Phase IV (post marketing) trials, and those trials are not going to be published for sometime yet. In the meantime, the great ship Vaccinator continues to sail on uncharted waters in a fog of increasing density and complexity.   


  1. Annie Davenport Turner Reply

    All I can say is that I know, and my friends know, people with headaches lasting 5 days, preventing work, and flu symptoms taking others to bed – with ‘an unbelievable negative cvd test?’ – and who immediately poo-poo any suggestion it’s a side-effect of their vaccination. Plus, an elderly lady who, after time in hospital after a clot after #1 (AZ) had her GP bang on her door to administer #2 (O) on Easter Monday evening, and without knowing about any of what she”d had or been through. Now tired and confused, it’s not being considered connected. I am in increasing despair at the human being…

  2. dr-no Reply

    Annie – the real problem is we simply don’t know. The YCS is a voluntary ad hoc system that has seen a notable rise in suspected adverse reaction reports for the covid vaccines, but we have no way of knowing whether that represents over-reporting (unlikely, see post for reasons), about right reporting (ditto) or under-reporting. This is yet another wasted opportunity to get some real answers.

    Nothing in the post alters that fact the vaccines are still in effect experimental, no medium to long term data exists, and full Montgomery level consent should be obtained, but clearly this isn’t happening. In a n=1 survey this afternoon, Dr No asked someone who had the vaccine in early Feb, and was admitted with a heart attack in early March (and made a full recovery after stents x 2), whether they knew what the Yellow Card Scheme was. They didn’t. Interestingly, they had speculated as to whether the vaccine had somehow triggered the heart attack…

  3. Vannnie Reply

    I’ve started asking people I know who’ve had the vaccine, firstly whether they were given the Patient Information Leaflet before their vaccines, so that they could make a decision based on “informed consent” (albeit rather “cherry-picked” information), and secondly whether they saw the bit on the form about reporting side effects.

    A lot of them were only given the PIL AFTER the event, and most of them didn’t bother to read it. A friend I saw today said he was just “so relieved to get the vaccine” that he didn’t really see the point in reading the leaflet. So it is extremely likely that adverse events are hugely under-reported.

    On the social media groups I’m in where people are giving feedback after their vaccinations, when it’s suggested that they report these side effects to the regulator in their country, most of them have no idea about there even being such a system!!!

  4. Shawn Reply

    I’d add one other thing to your original list of necessities: The person reporting has to believe that the side effect is significant enough to bother reporting it. Yes I know you can report anything for new drugs but let’s face it, if it’s insignificant you’re unlikely to bother. Quite how this plays into the high number of reports I’m not sure, you could argue it makes it more alarming but I think this is probably a constant. There’s clearly some hot-stuff bias at play here I’m sure you will agree. I would suggest that Yellow Cards were never designed to deal with such a novel medication being used at this scale on a population so hyper-aware of their own health.

  5. Tom Welsh Reply

    The basic problem is that everything in our social hierarchy is biased towards preserving the status quo: power first, money second, and everything else (including health) dead last. People get sick due to viruses, bacteria, etc. – never, of course, because of malnutrition, pollution, stress, lack of exercise, etc. – and then it’s the duty of the NHS to “treat” them by administering expensive drugs, tests and equipment. All of this keeps the profits flowing, and as far as humanly possible prevents any thought about the true causes of illness or thinking about how we might do better.

    Medical decisions should NEVER be taken by politicians, civil servants, other bureaucrats, businessmen, or anyone else without the necessary knowledge and experience. That emphatically includes those who were once doctors or scientists, but who have swum in political and commercial waters for so long that they have turned into sharks.

    The NHS was a fine idea and a grand aspiration, but perhaps its most decisive effect has been to take medical decision-making away from those best qualified and to give it to the ignorant, foolish and selfish.

    The admirable efforts and intentions of a million doctors, nurses, dentists, technical specialists, paramedics, etc. can be utterly subverted and frustrated by the selfish, scheming actions of a handful of contemptible managers and politicians.

  6. Helen McArdle Reply

    A brief browse through the MHRA Yellow card vaccine notification reports does immediately show the benefits & limitations of a system open to both public self-reporting and health professionals.

    One obvious benefit is that some very common side effects are immediately evident from the number of reports. These marry up well with the information GPs received in February for the AZ vaccine after the trials of 12021 vaccinated over 18s (of whom 90.3% were aged 18-64 and 75.5% where white).

    For example there have now been >48000 Yellow Card notifications for headache. In the AZ trial data, after injection site tenderness/pain, headache was the number one side effect (52.6%) followed by fatigue (53.1%), myalgia (44%), malaise (44.2%), pyrexia/feverishness (33.6%), chills (31.9%), arthralgia (26.4%) and nausea (21.9%). Most reactions were mild to moderate and usually resolved within a few days, but by day 7, 13% still had at least one systemic reaction. Older people had fewer reported side effects.

    So given that the commonest side effect symptom is also now a potential symptom of the rarest, most lethal side effect, it is little wonder that we are now seeing this:

    One problem with the unfiltered Yellow Card system is it contains a mix of symptoms, signs and diagnoses. Some of the symptoms are hopefully lay reports eg eyelid thickening, glassy eyes, screaming, moaning, foaming at the mouth, daydreaming; others are clinical observations eg faeces soft, pulse absent; some clear diagnoses eg myocardial infarction, DVT and yet others are frankly confusing (food poisoning, inguinal hernia, tooth missing, drowning, apparent death, AIDS, abdomen crushing, dementia, oesophageal cancer, exploding head syndrome). My favourite so far are ‘compulsive shopping’ and ‘sense of a foreshortened future’.

    Clearly, correlation does not always equal causation.

    I’m more interested in the HCP-reported diagnostic notifications and feel a responsibility to file these where there is a plausible link, even when there is uncertain correlation. But not every one does this and I don’t think we can rely on Yellow Card Notifications to give an early signal of a problem. Some HCPs may be tempted to tell a patient that their thrombosis/MI within 2 weeks of the vaccine is definitely not related to the vaccine, because they are ‘not the rare type of thrombosis associated with low platelets’. But what if they can be caused by the vaccine in certain people with certain risk factors, but haven’t been Yellow Carded, so are not known to be due to the vaccine? Unknown unknowns.

    Given these vaccines are being used under Emergency Use Authorisation, surely the Yellow Card system must be just an adjunct for detecting adverse events, not THE system for detecting them? Other studies show it is possible to link databases (eg the recent study widely reported in the press as supporting evidence for reduced Covid-19 transmission in vaccinated healthcare workers So theoretically we could link the GP record database with the vaccination database and detect a higher number than expected diagnoses. We do code the vaccinations within the GP record, so might not even need to do that as all the important diagnostic data may already be in one place.

  7. dr-no Reply

    Vannnie and Shawn – agree on all points. Hot stuff bias is definitely in play (yellow cards going from 40,000pa to 160,000 in three and a bit months) but at the same time many suspected ADRs don’t get reported, for all the various reasons stated. Dr No’s hunch is that the YCS will inevitably on balance under-report suspected ADRs. This is not really surprising, given it was never meant to provide data of rates of ADRs, but instead was intended to act as an early warning system, which it sometimes did rather too well. Back in the 1980s, it was criticised for triggering the removal of drugs from the market on the flimsiest of evidence.

    Helen – excellent observations. A ‘sense of a foreshortened future’ is a corker among corkers, and had Dr No done his homework as thoroughly as you, he might have used it as the title for the post.

    The pre-print you link to also reminds us of the missed opportunity to get some real data on side effects. Healthcare workers were the perfect cohort on which to do a prospective followup: a ‘captive’ population already plugged into databases, an epidemiologist’s dream, with the only limitation that the population by definition would be limited to working age people. But it would still be a very useful start. One can but wonder why use was not made of such a heaven sent opportunity.

  8. dearieme Reply

    I’m a fan of Malcom Kendrick’s blog. Either there, or in his excellent book Doctoring Data, he wrote – pre-Covid – that he was pretty confident that adverse side-effects are substantially under-reported.

    The reason, he suggested, lay in incentives. Who has an incentive to report them? Not the manufacturer. Not the doctor, who doesn’t want to be told that his prescription harmed a patient. Not, all too often, the patient – who doesn’t want the fuss. (And who might, I suggest, fear victimisation of some sort from the NHS.)

Leave a Comment

Your email address will not be published. Required fields are marked *