Norwegian nursing home deaths after vaccination are in line with what is expected by random chance

Updated: Jan 19

The occurrence of nursing home deaths after SARS-CoV-2 vaccination has created quite a stir, causing Norway to reconsider whether to vaccinate the most frail members of nursing homes, and raising concern in people all over the world. This has fed some of the fear-based vaccine hesitancy in many who are think that vaccines are inherently unsafe, or at least are hesitant to be vaccinated right now based on the perceived newness and exoticness of the mRNA technology and the unprecedentedly fast development, clinical study, and emergency approval of the vaccines.

As it became clear that these vaccines were going to earn emergency approval, I was concerned about talk that the most frail members of society in nursing homes would be in the first groups vaccinated. While these are clearly those at most risk of death from COVID-19, their demographic was not included in the phase 1-3 trials -- while the Pfizer and Moderna had reasonable proportions of subjects >65, it is not clear that any of them were nursing home residents. Thus, I was also concerned when seeing these deaths and wondered whether it was reckless to vaccinate this vulnerable population.

According to this article in Bloomberg on January 16, 2021, there have been 29 deaths in Norway in nursing home residents >75 years old from the day of vaccination up to 4 days post vaccination. The article is also careful to note that officials state that it is not clear whether the vaccination was a cause or contributing factor to these deaths, but of course most individuals in the public would naturally think that they must be somehow related to the vaccine and its side effects.

In this blog post, I will evaluate whether this number of deaths is more than one would expect on a given day even sans vaccination. according to the Norwegian Medical Agency, there are ~400 nursing home deaths per week in Norway, which is ~57 per day. According in, as in the table below, there were ~32,000 nursing home residents in Norway in 2019.

Based on these data, this would suggest that ~57/32,000= 1/560 residents of long term care homes in Norway die each day. Thus, if all nursing home residents were vaccinated, we would expect 32000/560=57 of them to die on the day of their vaccination even if the vaccine did not contribute at all. Put another way, based on these numbers, even if we went around and gave every nursing home resident in Norway a glass of water, we would expect 57 of them to die on the day we gave them the water, when clearly the water had nothing to do with their death. This is a type of "excess deaths" analysis that can provide a useful calibration point to assess whether it is likely that the vaccines were causing or contributing to these deaths.

The Bloomberg article states that just over 42,000 Norwegians have been given the first dose of vaccine so far, and that these were given to those at most risk of death from COVID-19, which includes the nursing home residents. It is not clear to me how many of these 42,000 went to nursing home residents, but if >16,000/42,000 went to nursing home residents, we would expect 28-29 of them to die on the actual day of vaccination, even if the vaccine had nothing to do with it, and many more in the few days afterwards. Thus, the report of 29 deaths within 4 days of vaccination is not surprising, and possibly well within the range of what one would expect even if no vaccination were done.

Let me emphasize this clearly -- these people are not numbers or statistics -- every one of them is precious, and their deaths are tragedies that are rightfully grieved. And like I said, each and every case is and should be thoroughly investigated by medical, scientific, government and regulatory officials and if vaccine related, this information needs to be taken into account in future vaccination plans to protect people in the future. But if these deaths are the same number that would occur in a day without the vaccine and if the vaccine in fact had nothing to do with it, misinterpreting their deaths as vaccine-related and then failing to give a potentially life-saving vaccine to others does not honor them or their memory, does not bring them back, and does not help with the grieving process. We can and should grieve them, but need to make an objective assessment of whether their death is likely to be vaccine-related or not if we are going to best protect future lives.

While it is important to investigate any reports of deaths or serious adverse events post-vaccination, it is also important for the public to follow the advice of officials, including those quoted in this case, and not be alarmed and jump to conclusions. But people will naturally be alarmed anyway, since it is natural for people to believe that deaths or serious unexplained side effects occurring just after vaccination must be caused by vaccination. This is especially true:

  1. when so many are on edge with a new technology

  2. whose development seems to be rushed

  3. in an environment when a substantial proportion of the country distrusts the scientists and government to the point that they even doubt the veracity of the pandemic or its seriousness, and

  4. in a society in which the level of statistical literacy is generally poor.

Individuals who considered and understood the basic calculations I demonstrate here would be more cautious about jumping to conclusions about vaccine safety based on reports like these. However, it is understandable that few would think to consider the expected rate of deaths on a given day or know how to compute it, and many would be distrustful of this line of analysis even if presented in an understandable way.

Given the fast and rapidly increasing viral transmission evident in the pandemic that is leading to healthcare systems being pushed to their limits and increasing record number of deaths daily, it is crucial that we use every tool available to us to manage the pandemic. This requires us to clearly and objectively assess the cost and benefits of population-level vaccination. Delivery of a safe and effective vaccine to the population is our best chance to reach herd immunity and bring the epidemic phase of this pandemic to an end and allow society to fully operate once again. If fear based on misinformation or lack of perspective causes many to refuse what may in fact be a safe and effective vaccine, it will not be possible to reach herd immunity and we will continue this current pattern of viral surges and lockdowns/closures indefinitely.

It is important to continue to evaluate vaccines to address the unanswered questions of durability of immunity, ability to prevent asymptomatic and transmissible disease, and safety long term and in special subpopulations. But so far they look promising, with stellar higher-than-expected efficacy data, and with strong safety data with the key documented side effects being temporary fever, headaches, fatigue, and soreness in the few days after vaccination and rare allergic reactions that are typically predictable and treatable if looking out for them. Reports of deaths and other serious side effects continue to be investigated and aggregated as required by the FDA, but the public must be careful not to be alarmed at anecdotal reports of deaths or side effects that occur after vaccination, assuming causation without verification or confirmation even in cases when the rates of death or disease emergence are not out of line with what is typical in the relevant population in a given day.

Additionally, any level of serious side effects needs to be weighted against the risk of NOT vaccinating, which include continuing surging exponential viral growth, a predictable percentage of those leading to deaths and hospitalizations that could overwhelm our healthcare system, not to mention the collateral damage to the economy and people's lives from mitigation strategies including business and school closures. At the end of the day, pandemic management is a multi-factor cost-benefit analysis, and these analyses need to be carefully and transparently done by a multidisciplinary team representing expertise in all relevant areas.

If we can ensure transparent reporting of side effects, careful measurement and characterization of efficacy and durability of immunity, and setting aside biases and fears, we have the potential as a society to gather and aggregate the available information to characterize the efficacy and safety of vaccines and construct the most appropriate vaccine distribution strategies to hopefully bring the virus under control so we can return to our normal, pre-pandemic lives. It will take transparency, trust, mutual respect, and cooperation at a societal level, which may seem impossible in our current politically fractured climate, but our mutual benefit depends on finding a way.

A word on consistency of death attribution after vaccination or COVID-19 infection

I have heard some ask, "If we consider any death with positive SARS-CoV-2 test as COVID-19 related, why shouldn't we consider any death after vaccination as vaccine-related?" This is a perfectly legitimate question, and comes from people that sense hypocrisy of one standard used to judge whether deaths of confirmed cases are virus-related, and another in which deaths after vaccination are dismissed as not vaccine-related. Many of the people who believe the danger of the pandemic is being exaggerated by the media and government are the same who doubt the safety and efficacy of the vaccine, and arguments like the one I am making in this blog post seem like inconsistent double-talk.

My response is that I completely agree with them that the exact same standard should be used in both cases, and I think they are. Not all deaths that occur with positive SARS-CoV-2 test are automatically recorded as COVID-19 related -- the virus has to be considered a contributing factor in order to list it on a death certificate, and this determination is not made by any central body but by individual hospital officials. And the temptation for bias, if any, is for hospitals, states, and nursing home facilities to downplay the number of COVID-19 related deaths since they will suffer PR and political damage from too many of them. In spite of urban legends floating around out there, hospitals are not "paid more" for attributing COVID-19 related deaths, and deaths of individuals who have tested positive for SARS-CoV-2 from clearly unrelated reasons (e.g. car accident) are not listed as COVID-19 related deaths.

Also, the CDC website presents results from an "excess deaths" analysis similar in principle to the one I attempt above for the Norwegian nursing home residents. As can be seen below, the data clearly demonstrate a 10-40% increase in deaths every week since the beginning of the pandemic, for a total of >400k excess deaths so far. This demonstrates that the level of deaths in the USA is far beyond what one would expect by random chance without the pandemic, and provides clear statistical evidence that the pandemic has produced an alarming level of excess deaths.

These net excess deaths were not all directly caused by COVID-19, but also include increased death rates from suicide, drug overdose, and delay seeking medical care because of fear of contracting COVID-19 or overwhelmed hospitals, and decreased death rates from any factors such as car accidents that might have decreased in a year in which most people stayed at home and traveled less than usual. However, the official COVID-19 related death counts have tracked linearly with reported hospitalizations and confirmed cases when appropriate lags in time for infection, testing, hospitalization and deaths are taken into account. This suggests the number of deaths for which COVID-19 was a causative or contributing factor is on the same order of magnitude as the official COVID-19 related death counts, which are currently approaching 400k for the USA and steeply rising, with >100k more expected in the next month.

This affirms what experts believe -- that the danger of COVID-19 is well established in terms of causing deaths that would not have occurred sans virus and in terms of potentially overwhelming health care systems when cases are surging. While more data needs to be collected to characterize the long term safety profile of the vaccines, data to date does not yet provide any evidence of excess deaths due to vaccination, or even establishment of any post-vaccine side effects that have long term repercussions and for which the vaccine has been demonstrated to be causative or a contributing factor. This supports the current population vaccination efforts.

Photo from NY Daily News

1,690 views2 comments

©2020 by Covid Data Science. Proudly created with