The COVID-19 vaccine is still relatively new, and researchers are still interested in studying how to maximize its effectiveness.
Typically, people receive the COVID-19 vaccine in the upper arm, which has multidose options. Multidose vaccines can be received in the same or different injection site for each dose. Other examples of multidose vaccines include those for measles mumps and rubella (MMR) and shingles.
A recent study published in The Journal of Clinical Investigation examined whether switching arms for two doses of the Pfizer BioNTech COVID-19 vaccine increased effectiveness.
Participants who switched arms for vaccine doses experienced a higher antibody response than those who received doses in the same arm.
The results showed this response increased over time in the subsequent follow-up visits.
These results point to a simple way to increase vaccine effectiveness. Future research could explore whether switching injection sites for other multidose vaccines could help improve immunity.
COVID-19 vaccination has effectively slowed infection rates and helped reduced severe illness.
The two main COVID-19 vaccines are both mRNA vaccines produced by Pfizer and Moderna.
Current recommendations from the Centers for Disease Control and Prevention (CDC) involve single doses of the Moderna or Pfizer-BioNTech vaccines for people who are not immunocompromised. However, multiple doses are still recommended for people who are immunocompromised.
Previously, other individuals received two doses of the Pfizer-BioNTech vaccine. Researchers of the current study wanted to see if the immune response produced by the Pfizer-BioNTech vaccine differed based on whether or not participants received doses in the same arm or the opposite arm from their initial dose. Study authors note there hasnt been a lot of research conducted in this area.
Researchers included participants from the OHSU COVID-19 Serology study, including almost 950 adults in their analysis. A total of 507 participants received at least two doses in the same arm, and 440 received at least two doses in opposite arms.
Researchers also looked at antibody response in a subgroup of matched pairs, with each pair having similar age, gender, vaccination, and time intervals between blood sample testing.
They were able to follow up on immune response among participants for up to 14 months after boosting.
Overall, researchers found that the group receiving vaccination doses one and two in opposite arms had a better immune response than those receiving doses in the same arm.
They saw higher levels of SARS-CoV-2 specific serum antibodies. They observed this difference more with later immunity testing than with earlier testing.
Study author Dr. Marcel E. Curlin, associate professor of medicine in the division of infectious diseases at the Oregon Health and Sciences University and the medical director for occupational health at OHSU, noted the following to Medical News Today:
In the context of first-time receipt of a 2-dose vaccine regimen, antigenspecific antibody levels resulting from vaccination are higher when giving the second dose in the contralateral arm relative to the first dose. This effect is durable, lasting more than a year after boosting. Contralateral vaccination also results in a broader immune response to challenges slightly different from the original vaccine (for example, to a variant of the original virus). We do not yet understand why this happens, but it is likely related to formation of memory and multiple rather than individual lymphoid centers.
Non-study author Dr. Arturo Casadevall, PhD, a microbiology and immunology expert with Johns Hopkins Medicine, told MNT the study data are strong.
The finding that contralateral arm vaccination results in higher antibody responses suggests that the simple intervention of switching arms during initial vaccination and boosting could produce stronger immunity and perhaps longer lasting protection, he shared.
This is an example of simple medical research with potentially high benefits for the individual and for public health.
Despite the promising implications the new research does have certain limitations.
First, researchers acknowledge the potential bias that could have occurred, though they believe this cannot account for all the results seen. Second, this study looked at a specific type of vaccination among adults and did not examine alternative immunization routes, so the results may not apply or be significant for other areas.
Researchers also did not look at cellular immunity when looking at potential protection from severe illness.
In addition, the cohort was comprised of healthcare workers, a specific population, so more research could also include more individuals in other fields.
Only 23% of participants were male, so its also possible for future research to include more gender balance. There was also a limitation based on how many participants completed all follow-up appointments.
Non-study author Jessica Smith Schwind, PhD, MPH, director at the Institute for Health Logistics & Analytics and associate professor of Epidemiology, Georgia Southern University, said from an epidemiologic perspective, the study has the potential to influence standard practice, but shared a word of caution:
However, a randomized study will be the gold standard to determine if a contralateral administration of the vaccine series would be most beneficial (and to what extent) for mRNA COVID-19 vaccinations. Also, it is important to keep in mind that immunologic response is a multi-faceted, complex process that can be measured in different ways. This study only measured antibody titers, which is only one component that influences a persons overall immune response to a pathogen.
Future research could focus on verifying these initial findings and expanding the data collection, such as looking at additional time points after vaccination.
At a basic science level the observation raises new questions for immunological research since it is difficult to explain how this effect occurs based on current understanding of how immune responses develop, Dr. Casadevall said.
I think the next step would be to carry out a prospective randomized controlled trial to determine if the effect holds. If the findings are replicated, I can imagine that this could lead to changes in clinical practice for how vaccines are administered and would stimulate new basic science research to understand the immunological mechanisms involved, Dr. Casadevall noted.
This research opens the door for future research into maximizing vaccine effectiveness.
We can probably derive greater levels of protection elicited by vaccines, based on the way we provide vaccination, Prof. Curlin noted.
Improved protection would likely be in the form of some degree of decrease in disease severity, particularly those with comorbid illness who are likely susceptible to severe disease.
One area for future research is looking into how switching vaccination sites may apply to other multidose vaccines. For example, the boost in immunity seen in this study may hold true for other multidose vaccines and increase their overall effectiveness.
Researchers note that future research can include pediatric data, as many of these multidose vaccines are usually part of child vaccination regimens.
Professor Curlin noted the potential benefits of this line of research in the future:
This effect, if generalized, could change the way we administer certain vaccine regimens, particularly in children. This effect could [also] have an impact on vaccines in development, particularly those with efficacy near threshold cutoffs for viable vaccine products. [However], it is important to remember that this issue requires additional study to help us better understand the mechanism for this effect and its generalizability to other vaccines.
A new COVID-19 vaccine study published last week linked the vaccine with an increased risk of heart, brain, and blood-related medical disorders.
The new study, which is the largest COVID-19 vaccine study to date, was conducted by the Global Vaccine Data Network (GVDN) in New Zealand and analyzed 99 million individuals who received COVID-19 vaccinations in eight different countries. The study monitored 13 different potential medical conditions in the individuals after they received a COVID-19 vaccination.
GVDN researchers concluded that the COVID-19 vaccine was linked to a small increase in heart, blood, and brain disorders. For example, some of the individuals who received mRNA vaccines had an increased risk of myocarditis, which is an inflammation of the heart muscle, while some individuals who received viral-vector vaccines had an increased risk of blood clots in the brain and an increased risk of Guillain-Barre syndrome, which is a brain disorder that causes the immune system to attack the bodys nerves.
The GVDN study also linked increased risks of spinal cord inflammation to viral vector vaccines and increased risks of swelling and inflammation in both the brain and spinal cord to mRNA and viral vector vaccines.
READ MORE: Fauci admits Covid social distancing not based on science
According to the Centers for Disease Control and Prevention, 81.4% of the U.S. population has had at least one COVID-19 vaccine. Unlike other vaccine studies, the GVDN study was able to identify potential vaccine safety signals due to the large sample size of the data.
The size of the population in this study increased the possibility of identifying rare potential vaccine safety signals, said Kristna Faksov of the Department of Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark, the lead author of the report. Single sites or regions are unlikely to have a large enough population to detect very rare signals.
Following the publication of the link between COVID-19 vaccines and heart, brain, and blood disorders, Dr. Marc Siegel, an NYU Langone Medical Center clinical professor of medicine, emphasized to Fox News that all vaccines have the potential for side effects.
It always comes down to a risk/benefit analysis of what you are more afraid of the vaccines side effects or the virus itself, which can have long-term side effects in terms of brain fog, fatigue, cough and alsoheart issues, he said.
A global study has connected those who received COVID vaccines to an increased risk of some serious medical conditions, but one Loudoun County, Virginia, doctor says theres more to consider here.
A global study has connected those who received COVID vaccines to an increased risk of some serious medical conditions, but one Loudoun County, Virginia, doctor says theres more to consider here.
The study published by the Global COVID Vaccine Safety (GCoVS) Project looked at around 100 million people who were given COVID vaccines around the world and the number of certain medical conditions reported before and after those vaccinations.
Depending on the region of the world, different increases in medical conditions were noted in those taking a variety of vaccines.
The most noted increases in conditions of those vaccinated in the United States were of cardiovascular conditions in those who had taken mRNA shots by Pfizer and Moderna. Those conditions included inflammation of the heart muscle and of the thin sac covering the heart
But Dr. David Goodfriend, director of the Loudoun County Health Department, said while all vaccines come with some risk of adverse reaction, in this case, all of the increases may not be because of the vaccinations.
COVID itself can cause the same infection, said Goodfriend. Its hard to know if its the vaccine thats resulting in an increased level of this, or it was just that we were having COVID.
Goodfriend said studies have also consistently shown the benefits of the COVID vaccines are higher than the potential risks.
I keep getting a COVID booster, he said, My family members get COVID boosters, because I think the risk of the infection by far outweighs any risk of the vaccine.
But he said it will take years for all the research data to come in. Because the COVID-19 vaccine was developed during a pandemic, the process was different from normal vaccine development.
Usually, when a vaccine is developed, there are years of clinical tests, and the recommended usage is altered as new information is learned during those trials.
In this case, while there were short clinical trials, the entire populace is essentially part of the extended study and there already been cases where the administration of a vaccine was altered to adjust to reactions being noted.
In real time, during COVID, when they started seeing an issue with one formulation of the vaccine they put that warning out to folks and were saying, Hey, were seeing heart conditions in young men who get this vaccine. Try to steer them away to a different product,' he said.
Now, with billions of doses administered, the research is getting much more powerful and sensitive at picking up even very rare side effects.
Goodfriend believes that over time, research will grow and define how COVID-19 is treated, and possibly see newly formulated annual vaccines, much like we currently do for flu shots.
As for how bad COVID has been this year, Goodfriend said we know that COVID is still deadly, but it will be a few more months before we know just how many people it impacted this year.
The good news, at least in Northern Virginia, is our hospital systems were not overwhelmed with COVID cases and were likely past the worst part of it for this winter, he said.
The study was funded by the Centers for Disease Control and Prevention, which provided $10 million for the study.
Get breaking news and daily headlines delivered to your email inbox by signing up here.
2024 WTOP. All Rights Reserved. This website is not intended for users located within the European Economic Area.
What initially was referred to as Paxlovid rebounda return of COVID symptoms or test positivity after starting a course of the antiviralis now more accurately referred to as COVID rebound, because rebound can happen regardless of whether someone takes antivirals.
Whats more, its likely not a phenomenon unique to COVID, says virologist Andy Pekosz, PhD, a professor in Molecular Microbiology and Immunology. What is unique to COVID is the technologyspecifically rapid antigen teststhat allows us to track the progress of an infection so closely.
Researchers are continuing to investigate why some people experience COVID rebound, whether people are contagious during this period, and the role antivirals like Paxlovid might play. But one thing is certain, says Amesh Adalja, MD, a senior scholar at the Johns Hopkins Center for Health Security who specializes in infectious diseases and pandemic preparedness: The possibility that someone might experience mild rebound symptoms should not deter them from taking Paxlovid if they are at a higher risk for severe illness.
COVID rebound is typically described as a recurrence of signs or symptoms or a new positive viral test result after initial recovery from COVID-19, according to the CDC. The order of events generally looks like this: A person is infected with and has symptoms of COVID; their symptoms subside over the course of the infection, and they test negative for COVID on a home antigen test; their symptoms return, and they may test positive again on a home antigen test.
Rebound symptoms are generally mild, and no hospitalizations or deaths have occurred as a result of rebound, according to a December 2023 CDC report.
Researchers are still working to understand why rebound happens and how it varies from person to personwith some experiencing a return of symptoms and others only seeing evidence through antigen testing. But at a molecular level, the fact that rebound occurs is not surprising, says Andy Pekosz.
We think about an infection as starting as nothing, going up to a peak, and then going away, but in reality, your body is much more complicated than that, he says. Using more sensitive tools, virologists have seen that the amount of virus present in someone with COVID goes up and down over the course of an infection, causing different degrees of symptoms.
A CDC review of COVID rebound studies found no consistent association between treatment and rebound. Paxlovid does not directly cause symptoms to return; in fact, only about 1 in 5 people who take Paxlovid experience rebound, and many of those are asymptomatic.
Paxlovid is taken daily for five days, and its possible that the immune response is somewhat blunted during that period. The hypothesis is that after those five days, the drug pressure from Paxlovid is gone and whatever remnants of the virus are still present are able to cause symptoms again, Adalja says. We dont quite understand why it happens in some people and not others, but it likely has to do with the nuance of the immune response and how that's impacted by antivirals.
The risk of rebound should not preclude someone who's at high risk or severe disease from taking Paxlovid, and it definitely shouldnt preclude doctors from prescribing it. Taking Paxlovid clearly reduces the likelihood of an individual developing severe COVID, irrespective of COVID rebound, Pekosz emphasizes.
Yes. In fall 2020 and spring 2021before Paxlovid became availablePekosz and other virologists studied how infectiousness and viral load varied among individuals over the course of a COVID infection. Their results showed numerous instances of what we now describe as COVID rebound: a return of symptoms or positive antigen test.
These results werent unexpected, according to Pekosz. Many of us virologists assumed [rebound] was happening because we understand that these viruses come and go, they're not smooth curves in terms of your responses, he says.
Possiblybut Adalja explains that contagiousness is better determined by test positivity than symptoms. If a person experiences rebound symptoms but doesnt test positive on a home antigen test, they likely arent shedding enough virus to be infectious, he says. However, if they test positive on a home antigen test, that does likely correlate with contagiousness.
A faint line on a home antigen test is still a positive test result. The tests are validated to tell you yes or no to whether it detects viral material, Pekosz explains, not how contagious you are. A dark line does mean more viral load, he says, but theres no system that correlates different shades of red with level of contagiousness.
If you're testing positive following a previous negative test or following a full course of Paxlovid, that's where masking and distancing becomes important, Pekosz says. But again, only at the tail end of an infection, when youre feeling better and symptoms arent severe.
Its likely that rebound occurs with other viral infections, Pekosz explains. It probably happens with influenza, but we don't have the at-home tests to capture that, Pekosz says, adding that he hopes future availability of at-home flu tests will allow virologists to analyze flu cases like they have with COVID.
What we can do in terms of diagnosing ourselves with COVID and following the course of the disease is light-years ahead of what we can do for influenza, Pekosz says. Its a testament to the fact that we've done good science and gotten these tools in place to see these kinds of variations that we've never been able to see with other viruses or infections.
Aliza Rosen is a digital content strategist in the Office of External Affairs at the Johns Hopkins Bloomberg School of Public Health.
Hi, its Bob in New York. The Centers for Disease Control and Prevention is considering revising its Covid isolation guidelines, amid a debate over whats the most practical advice for a Covid-weary world. But first...
Remember Covid-19? While the public largely moved on and tried its best to forget about the pandemic, the virus is still out there. Its highly contagious and continues to mutate. And it is still killing more people than influenza.
Thats the context for the debate over whether the Centers for Disease Control and Prevention should ease up on Covid isolation guidelines that advise people to isolate for at least five days after testing positive. Bloomberg News reported last week that the CDC hasnt yet decided whether to update its guidelines to allow for a shorter isolation period, after the Washington Post reported that such a change was in the works.
The CDCs Covid isolation guideline is more stringent than its recommendations for influenza, which say you should stay home at least 24 hours after fever is gone. As a practical matter, this mismatch can create a disincentive for people with mild symptoms to get tested for Covid.
People who received the Moderna and Pfizer vaccines for COVID-19 were more likely to come down with myocarditis or pericarditis, which are heart-related inflammations, according to a study involving 99 million people who received vaccine.
For example, people were 6.1 times more likely to come down with myocarditis following their second dose of Moderna vaccine. Higher than expected rates of myocarditis also occurred among people who received the Pfizer vaccine, according to the study which can be read in full in the journal Vaccine.
The study, which also looked at side effects among people who received the AstraZeneca vaccine, further found elevated risks of side effects including Guillain-Barre syndrome, a disease of the immune system, and blood clots in the brain.
Dr. Marc Siegel of NYU Langone Medical Center told Fox News, This study does not really change anything; it just provides much further evidence of what we already know.
Siegel and other experts further said the risks of those conditions and others including death are still far higher following an actual COVID-19 infection. Seigel, who is a regular Fox News contributor, called the side effects rare and said other studies show that the vaccine decreases the risk of myocarditis from COVID itself dramatically.
Forbes quoted biotech CEO Jacob Glanville as saying, The odds of all of these adverse events is still much, much higher when infected with SARS-CoV-2 (COVID-19), so getting vaccinated is still by far the safer choice.
The study, which was funded in part by the U.S. Centers for Disease Control and Prevention, looked at people in eight countries including Canada, but not the United States.
Previous studies have found the same side effects, and experts have noted that nearly everyone who experiences conditions such as myocarditis recovers quickly.
The new study, done by the Global Vaccine Network, which is part of the World Health Organization, focused on side effects that showed up within about two months of a vaccine dose. It contained no information regarding severity of illness, recovery time or deaths.
Doctors and the U.S. government have long cited a slight risk of myocarditis following COVID-19 vaccination, with the risk highest among young men.
One study involved members of the U.S. military who received a total of 2.8 million doses. Twenty-three men came down with myocarditis, with most coming after their second dose. In most cases symptoms went away within a week; none died.
All told, 13.53 billion doses of COVID-19 vaccine have been given around the world, with about 70% of people having received at least one. That includes 677 doses million in the United States as of May 2023, and 28 million in Pennsylvania.
A lot of viral illnesses can completely wipe you out, and, as weve learned throughout the pandemic, COVID-19 is no exception. But as conversations about the impact of long COVID continue to swirl, scientists are exploring just how common chronic fatigue might be after a bout with the virus.
Turns out, the risk is higher than you might think, according to a new study from the CDC published in the journal Emerging Infectious Diseases. Researchers analyzed the electronic records of 4,589 adults who received health care in the state of Washington and were diagnosed with COVID between February 2020 and February 2021 (meaning they were infected with earlier strains of SARS-CoV-2). They compared them to the records of people who didnt test positive for the virus during the same period (those with a suspected case or a history of COVID were excluded from this control group). On average, the researchers followed up with people for about a year post-infection.
The results are sobering: People who were sick with COVID had a 68% higher risk of developing incident fatigue, which refers to intense tiredness that develops after whats considered enough time to be recovered or the post-acute period. (There is no single definition for this, but the National Institutes of Health notes that COVIDs post-acute period has generally been defined as three weeks after symptoms first hit.)
The study found that folks in the COVID group were also 4.3 times more likely to develop chronic fatigueenduring exhaustion, in simplest termsthan those in the control group. The researchers noted that, in past studies, people who had post-COVID fatigue experienced symptoms that were similar to the profound fatigue thats signature to myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), a serious and potentially disabling condition that is often, but not always, triggered by an infection.
In general, women and people with certain underlying conditions, like diabetes or COPD, were hit by fatigue the hardest. Physicians should be aware that fatigue might occur or be newly recognized [more than a year] after acute COVID-19, the study authors concluded.
Long COVID is still a bit of a mystery: Theres a lot that experts dont understand about the condition, including why it develops in the first place.
But there are some theories: In people with long COVID, the immune response to the virus may cause direct or indirect damage that disrupts other body functions, senior study author Quan Vu, MD, an epidemiologist at the National Center for Emerging and Zoonotic Infectious Diseases at the CDC, tells SELF. (This would help to explain its vast and ever-growing list of potential symptoms.)
Another biggie: We believe, in long COVID, that many people have this ongoing inflammatory response, Thomas Russo, MD, professor and chief of infectious disease at the University at Buffalo in New York, tells SELF.
When your immune system reacts to a possible threatin this case, an infectionit uses an army of specialized cells and proteins to fend off whatevers making you sick. With long COVID, its as though the army keeps fighting even though it should go back to the barracks, William Schaffner, MD, an infectious disease specialist and professor at the Vanderbilt University School of Medicine in Nashville, tells SELF. That inflammation, experts believe, continues to smolder for a long period of time, he explains, triggering a slew of potential symptoms, including persistent fatigue.
Response variable
COVID-19 unvaccinated percentage (UP) is our chosen response variable. UP is calculated as the partial vaccination rate (PVR) subtracted from 100%, where the PVR is defined as the percentage of people in a county who had taken at least one dose of Pfizer (Comirnaty) or Moderna (Spikevax)16,24. As our goal is to deepen our understanding of vaccine hesitancy or vaccine refusal, we chose to define our variable as the percent of the population that did not receive any COVID-19 vaccine doses, rather than the percent fully vaccinated. Defining our variable as the percent fully vaccinated would complicate the interpretation of the variable as vaccine refusal, since it would exclude those that were willing to get the first dose but did not get the second before our time cutoff. The PVR data used to compute UP is sourced from Georgetown Universitys U.S. COVID-19 Vaccination Tracking website, which primarily relies on CDC data, supplemented with vaccination data from local health departments where CDC data is incomplete16. Vaccination data is not available for 69 counties in Alaska, Nebraska, Georgia, and Virginia, so these counties were excluded from the analysis. Additionally, since the Johnson & Johnson (J&J) vaccine only requires one dose to be fully vaccinated, the PVR excludes individuals who got the J&J vaccine. However, only 3.3% of vaccinations administered were J&J as of December 15, 202125, so the exclusion of J&J vaccinations should have minimal impact on our conclusions.
While our response variable measures lack of vaccine uptake, not hesitancy directly, we believe that this work will still provide insight on factors correlated with vaccine hesitancy. Previous work found that COVID-19 vaccine uptake is strongly correlated with vaccine hesitancy, as measured by survey data6. Further, vaccine uptake rates reflect real world vaccination behavior at the population level, in contrast to vaccine hesitancy surveys which are available for a subset of locations around the U.S. and suffer from sampling biases. In addition, we selected the time cutoff of December 15, 2021 in order to minimize the impact of non-hesitancy factors on uptake, such as vaccine eligibility and accessibility. Therefore, our response variable serves as a reasonable proxy for vaccine hesitancy, and we think it is the best choice possible based on the data that is currently available.
All demographic and socioeconomic variables are sourced from publicly available datasets at the county level, from the U.S. Census Bureaus 2020 Decennial Census17 and U.S. Department of Agriculture (USDA) Economic Research Service18. The percentage of Black people and the percentage of Latinx people represent the self-identified proportion of those races in each county. Postsecondary education percentage is measured as the percentage of adults with educational attainment more advanced than completing high school. The uninsured percentage is the percentage of people who reported not having health insurance. Additional metrics include median age, average number of vehicles per household and the median household income.
The political affiliation variable, defined as the percentage of voters who chose Donald Trump as their presidential candidate during the 2020 presidential election, is sourced from the MIT Election Data and Science Lab19. Previous work has found this data to be associated with vaccine hesitancy9,10. Compared with other political indicators, such as other election results, voter registration data, or public opinion polls, the presidential election has the highest voter turnout and the most policy influence. We hereby adopt this metric as a proxy of the political affiliation. It is referred as Republican presidential vote percentage (%) in the study.
In efforts to explore whether a county that experienced more burden from COVID-19 may be more willing to adopt preventative measures such as vaccination, we incorporate a variable to capture a county's historical COVID-19 infection rate. Specifically, to measure historical COVID-19 burden, we use the cumulative number of COVID-19 cases per 100,000 people as of December 15, 2021 from the Johns Hopkins University Center for Systems Science and Engineering (CSSE) COVID-19 GitHub20. In order to remove outliers, values that were more than 4 standard deviations above the mean were excluded.
Since MMR vaccination coverage is an indicator of vaccine acceptance before COVID-19, we hypothesize that higher (pre-pandemic) MMR vaccine uptake rates may be associated with higher COVID-19 vaccine coverage. Previous work has shown a strong association between MMR coverage and vaccine hesitancy26,27. In most U.S. states, the MMR vaccine is required for children to attend public school, making MMR coverage a strong indicator of anti-vaccination behavior. While this data reflects the results of parents making vaccination decisions for their children, in contrast to measuring COVID-19 uptake among mostly adult populations, the decision to forgo mandatory childhood vaccinations is indicative of strong hesitancy that we hypothesize may transfer to other vaccines. To test this hypothesis, we incorporate a variable in this analysis that is based on the MMR vaccination coverage rates of children in kindergarten in 2019, data that we gathered in a previous study21.
A set of variables intended to capture the potential role of information consumption on vaccine choice includes four television viewership rating variables and a Twitter misinformation variable. The county level viewership ratings (RTG) % for four major channels, namely FNC (Fox News Channel), CNN (Cable News Network), MSNBC (Microsoft National Broadcasting Company), and local news, are sourced from Nielsen Media, where RTG is measured by the estimated percentage of households tuned to a specific viewing source, e.g., news channel. The four viewership variables were computed as the average of the monthly viewership ratings for each channel from February to November 2021. January 2021 data were excluded due to anomalies caused by the January 6th U.S. Capital Attack. Nielsen data is not available for several counties in Virginia and Alaska and all counties in Hawaii, so these 72 counties are excluded from the analysis. The analysis also excludes outliers, defined as those counties with cable viewership values that are more than 4 standard deviations away from the mean. Additionally, within the model each of the cable TV viewership variables was standardized to a mean of 0 and standard deviation of 1, to provide a more interpretable understanding of the relative position of each countys rating.
Another information consumption variable included in the model is the Twitter misinformation variable. This variable is intended to capture the prevalence of COVID-19 vaccine misinformation in circulation on Twitter during a time that likely influenced behavior during the study period. The variable is based on a previous study by Pierri et al., who provided a variable that is representative of the percent of COVID-19 vaccine-related tweets that contain links to low credibility sources at the county level15,23. This variable has some limitations, as it is based on only the set of Twitter accounts that can be geolocated. To ensure a large enough sample size for a reliable estimate, counties with less than 50 geolocated accounts are not included, which results in a data set that includes 904 counties.After excludingoutliers, defined as values that are more than 4 standard deviations from the mean, we have 855 counties. An analogous data set is also available with a minimum of 10 and 100 geolocated accounts, but we opted to use the cutoff of 50 to balance having a more representative sample size of accounts per county with the number of counties we can include in our analysis. Due to the limited number of counties that this data is available for, a separate sub-analysis is conducted that includes this variable (Fig.2).
Various land-use variables are sourced from the U.S. Census Bureau, namely the population size and the number of residents in rural or urban areas for each county28. These variables are used to cluster counties for the sub-analyses, which are further described in the methods section. For the cluster-based analysis we categorize counties into mutually exclusive sets based on (1) population quartiles and (2) a binary rural or urban classification. For the binary classification, a county is classified as rural if the majority of the population is designated to live in areas classified as rural and otherwise classified as urban.
We use a Generalized Additive Model (GAM) to explore the relationship between each county's unvaccinated percentage and the aforementioned variables. GAMs provide a balance between model complexity and interpretability, and critically, they can reflect the relative importance of different features29,30. Specifically, GAMs model the response variable as the sum of unknown smooth functions of covariates, and unlike Generalized Linear Models (GLMs), GAMs offer the capability to model nonlinear relationships between variables. For example, a linear regression model may show an overall positive correlation between an input variable and the response variable, but using a GAM on the same data may reveal a more nuanced relationship, like a strong positive trend in some ranges of the input variable and a weak negative trend elsewhere. Due to the complex nature of relationships between our variables and vaccination uptake, GAMs are a more appropriate choice than linear models, since they can capture both linear and nonlinear relationships.
The proposed GAM is fitted to the unvaccinated percentage as the response variable, which is assumed to have a Gaussian distribution, and a log link. REML (restricted maximum likelihood) is used to estimate smoothing parameters, which returns relatively reliable and stable results. Specifically, the model in our primary analysis has the following form:
$${Y}_{i}sim Gaussian(mu )$$
$${text{log}}left(mu right)sim {s}_{1}left(cumulative ,COVID-19, case, rateright)+ {s}_{2}left(percentage, of ,Black, peopleright)+ {s}_{3}left(percentage, of ,Hispanic, peopleright)+ {s}_{4}left(postsecondary, education, percentageright)+ {s}_{5}(median ,household ,income) + {s}_{6}(median ,age) + {s}_{7}(vehicles, per, household) + {s}_{8}(uninsured, percentage) + {s}_{9}(MMR, coverage) + {s}_{10}(std(FNC, viewership) + {s}_{11}(std(CNN ,viewership)) + {s}_{12}(std(MSNBC, viewership) + {s}_{13}(std(Local, News, viewership) + {s}_{14}(Republican ,presidential ,vote ,percentage)$$
where Yi denotes the unvaccinated percentage for each county i. The model is a sum of smooth functions ({s}_{i}), and each smooth function consists of a number of basis functions (k). Sensitivity analysis that varies the number of basis functions was conducted. A value of k=3 for each smooth function was found to provide the optimal balance between preventing both underfitting and overfitting of the model and maximizing interpretability of the results. Additionally, the GAM model is weighted to prevent the highly imbalanced county population distribution from skewing the results. The weight is computed by normalizing each countys population by the average county population, taking a log transformation to adjust for the skewness. The weight implemented in the primary analysis is defined as:
$$weigh{t}_{i}=frac{{text{log}}(po{p}_{i})}{mean({sum }_{i}{text{log}}(po{p}_{i}))},$$
where i is the county index. The primary model is run for 2885 counties (reduced from the full set of counties due to missing data and data quality issues referenced previously).
As noted previously, the Twitter misinformation variable, ({s}_{15}(Twitter ,misinformation)), is only available for 855 counties, and is therefore run as a separate model using the same general function and weights as the primary model, but with the additional determinant included.
In addition to the primary model presented above, we conduct sub-analyses to determine how the relationship between unvaccinated percentage and its associated factors varies across urban versus rural counties. In the U.S., vaccination uptake was substantially lower in rural areas31. Multiple studies have examined the reasons for this discrepancy. Some factors associated with lower vaccination in rural areas were captured in our study, including lower educational attainment, voting for Trump in the 2020 election, and lower insurance coverage32. However, other relevant factors could not be incorporated in our study, including that rural residents have a lower perceived risk of COVID-19, higher vaccine hesitancy, and are less likely to adopt covid risk mitigating behaviors32,33. In order to better understand the influence of different factors in rural versus urban counties, we complete two cluster-based sub-analyses: one clustered by rural versus urban counties and another clustered into four quartiles based on population size, as described below. Due to the difficulty of neatly separating counties into urban or rural, we provide the population size cluster analysis to confirm that our urbanrural analysis is accurately capturing differences between urban and rural counites, which broadly aligns with higher versus lower county populations.
Land-use cluster-based analysis: Counties are clustered into two groups based on their primary land use pattern, namely as urban or rural counties. Two independent weighted GAM models are run, one for each group. The rural model includes 1,835 counties, and the urban model includes 1,050 counties.
Population cluster-based analysis: Counties are grouped into quantiles based on their population size. Four independent GAM models are generated, one for each distinct quantile. The respective models contain 664, 721, 739, and 761 counties ranging from the smallest to largest population size groups. GAMs are implemented without weights for each group in this sub-analysis, because the weighting is based on population size.
We evaluate the goodness-of-fit by conducting a diagnostic analysis for each model and sub-model. These evaluations include the Q-Q plots, histograms of residuals, mapping of residual values versus predicted values, and mapping of response against fitted values. The diagnostic analysis outcomes for the primary model are presented in the Supplementary Material (Fig. S2). The concurvity in the primary model is also measured to ensure pairwise values remain below 0.8 and avoid cases in which one variable is a smooth function of another. The outcomes of the diagnostic analysis demonstrated consistency in fit and performance across all models.
Dr. Anthony Fauci visited University of Illinois Chicago on Tuesday to discuss his decades-long career in medicine and public service.
Marie Lynn Miranda, UICs chancellor, spoke with Fauci about his work battling two major public health crises the HIV and AIDS epidemic and the COVID-19 pandemic the need for more robust local public health systems to better manage future outbreaks and restoring trust in health care professionals and scientists.
The 83-year-old retired in 2022 after leading the National Institute of Allergy and Infectious Diseases for nearly 40 years. He served as an adviser on domestic and global health issues to seven U.S. presidents.
About 2,100 people attended the chat at UICs Isadore and Sadie Dorin Forum, including former Mayor Lori Lightfoot, who did not speak. At three separate points, pro-Palestine demonstrators with the organization Behind Enemy Lines interrupted Fauci and Miranda and were removed from the auditorium.
Pro-Palestinian demonstrators disrupted Dr. Anthony Faucis talk on Tuesday at UIC.
During the chat, Fauci discussed the countrys failures in responding and managing the COVID-19 pandemic. He pointed out that before the outbreak, the Johns Hopkins School of Public Health ranked the U.S. as overwhelmingly the best country for pandemic preparedness.
But at the end of the fourth year of the pandemic, we had 1.7 million deaths and the outbreak isn't even over yet, Fauci said. We are now in our fifth year of this outbreak, and that is more deaths per capita than virtually any other country in the world, including low- and middle-income countries with maybe one exception.
The countrys diminishing public health system was one of the reasons the response to the virus was weakened, especially given that most local public health departments are critically underfunded, Fauci said.
So when you heard us always talking about how we can control [COVID-19] by identification, isolation and contact tracing, all of that takes place at the local level, Fauci said. And if you don't have local public health infrastructure, no matter how devoted and committed the people are, if they don't have the resources, it's not going to happen.
Dr. Anthony Fauci, who is the former director of the National Institute of Allergy and Infectious Diseases at the National Institutes of Health and former chief medical advisor to Pres. Joe Biden, speaks to University of Illinois Chicago Chancellor Marie Lynn Miranda during Chair Chats at the UIC Isadore and Sadie Dorin Forum, Tuesday, Feb. 20, 2024.
At the federal level, responding to COVID-19 was complicated by the ever-changing virus, Fauci said.
When you're dealing with the moving target, then you have to make decisions based on the information, the data and the evidence that you have, Fauci said. So what we knew in January of 2020, was very different from so many standpoints than what we knew in July of 2020 and in April of 2022.
COVID-19 wasnt the first outbreak Fauci handled. During the early 1980s, he was on the forefront of responding to the growing HIV and AIDS epidemic. He started working for the National Institutes of Health in 1981 after finishing fellowships in infectious disease and immunology.
Nobody else wanted to get involved in treating and learning about the disease, Fauci said, and some didnt think it was even worth addressing.
Fauci said his undergrad degree in classical studies made him a more empathetic doctor, enabling him to see patients such as those with AIDS as people outside of their infections.
I just felt this compelling empathy towards these young gay men who were being stigmatized, not only for being gay, but stigmatized because they had a disease that was a mysterious disease that was likely sexually transmitted, Fauci said.
I felt an almost fundamental ethical responsibility to say I'm going to change the direction of my career and start studying this disease that didn't have a name.
Dr. Anthony Fauci, who is the former director of the National Institute of Allergy and Infectious Diseases at the National Institutes of Health and former chief medical advisor to Pres. Joe Biden, speaks during Chair Chats at the University of Illinois Chicago Isadore and Sadie Dorin Forum, Tuesday, Feb. 20, 2024.
Calculation of excess deaths
The number of excess deaths in each period and age-sex-geography stratum is calculated as the difference between the observed and expected number of deaths:
Where E-hat[i,t] is the estimated number of excess deaths, d[i,t] is the observed number of deaths and d-hat[i,t] is the expected number of deaths in age-sex-geography stratum i in period t.
The estimated total number of excess deaths in each period is obtained by summing estimated excess deaths across age groups, sexes, and geographies:
This "bottom-up" approach ensures additivity throughout the aggregation structure. For example:
estimated excess deaths by age group for males in a particular UK country will sum to the total estimated excess deaths across all age groups for males in that UK country
estimated excess deaths for males and females in a particular UK country will sum to the total estimated excess deaths for both sexes combined in that UK country
estimated excess deaths in individual UK countries (England and Wales combined, Scotland, and Northern Ireland), including deaths among non-residents, will sum to the total estimated excess deaths for the UK
Temporal additivity between monthly and annual estimates of excess deaths is also achieved by summing the estimated excess deaths obtained from the monthly model to derive annual totals. However, weekly estimates will not necessarily sum to annual estimates, as weeks may straddle calendar years at the beginning and end of each year.
It will always be the case that the number of excess deaths in a period is an estimate rather than a known value, because the number of expected deaths is a counterfactual quantity that must be estimated from observed data using statistical techniques. To reflect this uncertainty inherent in expected and excess deaths estimates, 95% confidence intervals are constructed around the excess deaths estimates using the following formula:
Where E-hat[i,t] is the estimated expected deaths in age-sex-geography stratum i in period t, and SE(E-hat[i,t]) is the standard error of the estimate, which is the square root of the variance of the estimate, V(E-hat[i,t]). See our Uncertainty and how we measure it methodology for more information on confidence intervals and standard errors.
The number of excess deaths is estimated as the difference between the observed and expected number of deaths, so the variance of the estimated excess deaths is a combination of the variances of both these quantities. However, the observed number of deaths is a known quantity rather than an estimate, so it has no variance. Therefore:
Where d-hat[i,t] is the expected number of deaths in age-sex-geography stratum i in period t and V(d-hat[i,t]) is its variance, approximated through the Delta method.
The overall variance of the expected total number of deaths across age groups, sexes and geographies in each period can be found by summing the stratum-specific variances within periods:
The population denominators used to calculate mortality rates in each period and age-sex-geography stratum are derived from mid-year population estimates. These population estimates are not timely enough to feed into contemporary estimates of excess deaths. For example, estimates relating to mid-2022 were not published until August 2023 for Northern Ireland, on the Northern Ireland Statistics and Research Agency (NISRA) website, and November 2023 for England and Wales in our 2021-based National population projections bulletin. They have not yet been published for Scotland.
In the future, the Dynamic Population Model and resulting admin-based population estimates may provide more timely estimates (see our Admin-based population estimates: local authorities in England and Wales article). For the time being, the mid-year population estimates are extrapolated with population projections in each age-sex-geography stratum. Historical estimates of excess deaths will be revised whenever population projections for a given year are replaced by the mid-year population estimate.
National population projections are typically updated once every two years but subnational projections (needed for population denominators in the English regions) are only updated once every four years. These are published several months after the corresponding national update. For example, our 2021-based National population projections bulletin was published in January 2024, and before this our 2020-based National population projections bulletin was published in January 2022.
However, our latest available Subnational population projections bulletin (2018-based) was published in March 2020. Therefore, contemporary population sizes for the English regions are obtained by applying the regional proportions from the latest mid-year population estimates to the latest available national population projections for England. This ensures that the population denominators used for calculating mortality rates across the English regions sum to the national population denominator for England.
Population estimates and projections relate to the estimated population size at the mid-point of each year, but population denominators are needed on weekly and monthly bases for excess deaths calculations. Therefore, weekly and monthly population estimates are linearly interpolated between the mid-year estimates.
The pandemic saw a large increase in death registrations, particularly in certain weeks and months that coincided with "waves" of infection (for example, when new COVID-19 variants became widespread in the population). To avoid these periods affecting estimates of expected deaths in subsequent periods, they are removed from the dataset when the model is fitted so that they do not contribute to the mortality baseline. This means that estimates of excess deaths in subsequent periods relate to the additional deaths registered in the period, over and above what would be expected from previous periods had they not been extraordinarily affected by the pandemic.
We define periods extraordinarily affected by the direct mortality impacts of the pandemic as being those where COVID-19 was given as the underlying cause of death for at least 15% of all deaths registered in the period across the UK. This threshold gives the greatest coherence between the weekly and monthly data in terms of periods excluded from the model fitting. These periods are April and May 2020, and November 2020 to February 2021 for monthly data; they are Weeks 14 to 22 of 2020, and Week 45 of 2020 to Week 8 of 2021 for weekly data.
The annual calendar on which we report our weekly mortality statistics usually comprises 52 seven-day weeks and is 364 days in length. By contrast, the Gregorian calendar year (used by most countries across the world) is 365 days long for non-leap years and 366 days long for leap years. This means that the reporting calendar slips out of alignment with the Gregorian calendar by one or two days each year. To avoid this misalignment becoming too severe, there is international agreement that a "Week 53" should periodically be added to the reporting calendar.
Week 53 occurs infrequently (it was last added to the mortality calendar in 2020, and before that in 2015), so it is not practical to estimate a separate seasonal term for it when fitting models to five years of data. Instead, any instances of Week 53 are re-labelled as Week 52 when fitting models and obtaining expected numbers of deaths. This assumes that the mortality rate in a typical Week 53 is similar to a typical Week 52.
In the future, it is anticipated that we will publish estimates of excess deaths in each of the four UK countries as well as the total excess deaths in the UK as a whole. National Records of Scotland (NRS) and NISRA will also separately publish estimates of excess deaths for Scotland and Northern Ireland, respectively, using the same methodology as the Office for National Statistics (ONS). This will ensure consistent and comparable estimates across all parts of the UK.
For consistency with the death registrations data we publish and the devolved administrations, the following models are fitted to estimate excess deaths:
deaths registered in England or Wales, including those for non-residents
deaths registered in Scotland, including those for non-residents
deaths registered in Northern Ireland, including those for non-residents
deaths registered in England or Wales among residents of England
deaths registered in England or Wales among residents of Wales
The total number of estimated excess deaths across the UK is then derived by summing the outputs from the first three models listed. The fourth model listed includes English region of residence as an explanatory variable.
In practice, 10 models are fitted to obtain estimates of excess deaths: five for weekly data and five for monthly data. In addition, five models are fitted to the annual data to obtain standard errors and confidence intervals around the annualised estimates (monthly excess deaths estimates can be summed within years to obtain annual estimates, but this is not possible for the standard errors because of the existence of correlation between successive monthly estimates, which is generally the case with any time series data). To obtain the variance of the annualised estimate, we assume that its coefficient of variation is the same as that of the estimate from the model fitted to annual data.
The models fitted to annual data include age group, sex, English region (only in the model for deaths registered in England or Wales among residents of England), a trend component and the number of weekdays in the year.
In our current approach to estimating excess deaths in England and Wales, and that of the devolved administrations of Scotland and Northern Ireland, the expected (baseline) number of deaths is estimated as the average number of deaths registered in a recent five-year period. In contrast, our new methodology is based on age-specific mortality rates rather than death counts, so trends in population size and age structure are accounted for. Furthermore, the five-year average mortality rate is adjusted for a trend, so historical changes in population mortality rates are also accounted for.
Before the pandemic, the five-year period used in the current methodology was the five years preceding the current year. For example, the expected number of deaths in 2019 was estimated as the average number of deaths registered from 2014 to 2018 (inclusive). Weekly and monthly expected deaths were estimated as the average number of deaths registered in the same week or month over the past five years. For example, the expected number of deaths in Week 1 of 2019 was estimated as the average number of deaths registered in Week 1 from 2014 to 2018 (inclusive).
The expected number of deaths in 2021 was estimated as the average of deaths registered from 2015 to 2019 rather than 2016 to 2020, to avoid the pandemic distorting the excess deaths calculation. The expected number of deaths in 2022 was estimated as the average of deaths registered in 2016, 2017, 2018, 2019 and 2021.
In contrast, individual weeks and months, rather than whole years, that were substantially affected by the immediate mortality impact of the pandemic are removed from the expected deaths calculation under the new methodology.
Other improvements brought about by the change in methodology include:
use of a statistical model means that multiple demographic, trend, seasonal and calendar effects can be included simultaneously in the estimation of expected deaths, and confidence intervals can readily be obtained
a "bottom-up" approach to aggregation means that estimates of excess deaths are additive across age groups, sexes, and high-level geographies, and between months and years
having a common methodology for all four UK countries means that estimates of excess deaths are consistent and comparable across all parts of the UK, and the new methodology is largely coherent (though not identical) to that used by the Office for Health Improvement and Disparities (OHID) to estimate excess deaths in English local authorities.
