Three years in, the impacts of the covid-19 pandemic continue to reverberate around the world. The widescale, if unequal, distribution of vaccines has surely saved millions of lives and lessened people’s misery since the beginning of 2021. But thousands of people worldwide are still dying, and millions are getting infected weekly, while newly emerging variants threaten to undermine the progress made with our vaccines and treatments.
There also remain many enduring puzzles about the virus and our relationship to it. One simple question has proven incredibly hard to answer with much clarity: How often is long covid happening?
A quick browse through media headlines and the scientific literature will turn up wildly different answers. A review last October, for example, estimated that more than half of all survivors experience postacute sequelae of covid-19, or PASC, six months after their initial infection. Yet another paper by the nonprofit FAIR Health released this June estimated that nearly 25% of patients experienced still had symptoms possibly linked to covid 30 days later. Other studies have pegged the prevalence to be 10% or lower.
Part of the struggle in studying this issue is that the definition of long covid remains imprecise. Patient advocates were the first to document lingering issues following the initial illness by mid-2020, but it took public health agencies much longer to formally recognize the condition. The World Health Organization, for instance, only released its clinical case criteria for “post-covid-19” in October 2021. That criteria defines the condition as “occurring in individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset of covid-19 with symptoms that last for at least 2 months and cannot be explained by an alternative diagnosis.” The very nature of long covid, which is likely to be caused by more than one mechanism, further complicates matters.
“Nobody ought to be surprised that covid can have these effects, because we know about the consequences of acute viral illness from a lot of other things. Long flu is a thing; long Epstein-Barr virus is a thing,” Bill Hanage, an infectious disease epidemiologist at Harvard University’s Center for Communicable Disease Dynamics, told Gizmodo by phone. “The difficulty is in measuring it.”
Read More: The Challenges of Unravelling Long Covid
Even today, there is no simple test or biomarker for identifying someone with long covid, meaning that diagnosis is a process of elimination. There are commonly agreed-upon symptoms, such as fatigue, respiratory problems, and cognitive dysfunction or “brain fog.” But there remains much debate about the extent of health problems that can conclusively be linked to long covid. One study last February, for instance, failed to find a link between infection and the later increased risk of an inflammatory condition around the toes and fingers known as chilblains—a usually rare ailment that seemed to become so synonymous with the pandemic that it was dubbed “covid toes.”
Another stumbling block has been the relative lack of data that could be representative of the average person’s risk of long covid. Many studies at first relied on analyzing the outcomes of hospitalized patients post-infection, in part because their medical records would be easiest to find and track over time. Several studies have since shown that people with more severe illness initially (an important but ultimately small slice of all covid-19 cases) are more likely to develop lingering complications than those with mild to moderate illness. So trying to judge the overall prevalence of long covid from these studies would almost certainly lead to overestimates. Importantly, the earlier mentioned October review looked at studies where nearly 80% of patients in total were hospitalized.
One of the most serious hurdles in learning about long covid has been that many studies haven’t included control groups for comparison. Unfortunately, humans will experience many of the symptoms associated with long covid for many different reasons at any given point in time. So simply counting up the percentage of covid-19 survivors who report these problems doesn’t necessarily tell us whether their infection was the source of it. Using a control group as a baseline alone doesn’t directly establish that covid caused people’s post-infection symptoms, but it helps eliminate much of the noise in the data.
Researcher Ziyad Al-Aly and his colleagues have been able to rely on a unique and large source of data to work around some of these issues: people who get their health care covered by the federal government through Veterans Affairs, the largest integrated health care system in the U.S. Last November, his team published a study trying to gauge the burden of long covid symptoms on patients. Their data made it possible for them to compare nearly 200,000 covid-19 survivors to 5 million people who never tested positive for covid-19 during the study period, and to compare their symptoms across different levels of initial illness.
Even patients who only had mild illness at first, the team found, had an increased risk of health problems beyond the first 30 days after infection, compared to controls, though the risk was highest for severe cases. Subsequent studies of theirs with the same dataset have found that long covid patients have a greater risk of heart problems. But they also estimated that the overall prevalence of prolonged symptoms in survivors that could be connected to covid was around 7%, and around 4% for initially milder cases.
“Having a control group and trying to account for baseline health conditions is very, very important,” Al-Aly, the chief of research and development at the VA Saint Louis Health Care System, told Gizmodo by phone. “That’s how we can start to scientifically and rigorously disentangle the consequences of SARS-CoV-2 from everything else that may be happening in people’s lives.”
The twisting path of the pandemic has added more confusion to calculating the risk of long covid. Last year saw the emergence of multiple variants of the coronavirus, particularly the Delta and Omicron variants. Delta is thought to have caused more severe illness on average than other forms, which in theory would have raised a person’s risk of long covid. Omicron, on the other hand, is likely less severe than Delta, but it has been much more transmissible and able to infect people with some preexisting immunity from vaccination or past infection. So Omicron might be less likely to cause long covid on an individual basis than Delta, for instance, but it still could have led to an overall high number of cases this past winter.
Meanwhile, the vaccines appeared to be very capable of preventing infection (and presumably long covid) early on, and they’ve remained very protective against severe illness, but their potency against infection has waned due to time and the appearance of Omicron. A recent UK review of the data found that vaccines can cut the risk of long covid by half, but another study by Al-Aly and his colleagues estimated that the risk reduction may actually be closer to 15%.
There are also the related questions of long covid severity and duration. Some people’s most noticeable symptom may be a reduced or permanently gone sense of smell. Others may experience fatigue and brain fog so crippling that they can no longer function at work or in their day-to-day lives. And some people’s symptoms may dissipate entirely over time, while many others may have improved but are still not fully recovered.
Some researchers have been able to take a more proactive approach to estimating long covid prevalence. The UK in particular has been much better at tracking covid-19 than other countries, and that’s included asking people in real-time if they’ve been experiencing lingering symptoms post-infection. As of June 2022, nationally representative survey data suggests that 2 million UK residents are currently experiencing self-reported long covid symptoms (symptoms lasting more than four weeks after infection), which is about 3% of the entire population. About two-thirds of these people said that their symptoms were adversely affecting their daily life. Their other data has found that about 8% of double-vaccinated people infected with the Omicron variant reported longer-term symptoms, compared to about 15% during the Delta era.
The U.S. Centers for Disease Control and Prevention recently released data from a similar representative survey of Americans, conducted in early June. They found that nearly one in five people who reported having covid (19%) were experiencing symptoms lasting longer than three months that appeared not to be present before infection. And they estimated that the current prevalence of long covid among all Americans was around 7.5%.
Last April, SolveCFS, a nonprofit advocacy organization focused on myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), released its own report estimating the impact of long covid in the U.S. For decades, ME/CFS advocates have been trying to raise awareness of the condition, which has been linked to post-infection complications. Their modeling assumed that 30% of unvaccinated cases would lead to long covid, based on UK survey data. But they also took into account the likely reduced risk in vaccinated people as well as the emergence of more breakthrough infections during Omicron. They additionally took the unique step of separating out more severely disabling cases and created two models for reported infections and total estimated infections, based on surveys of people’s antibodies to the virus.
They estimated that anywhere from 22 million to 43 million Americans had developed long covid as of late January this year, or 7% to 13% of the total U.S. population They also estimated that 7 million to 14 million Americans, or 2% to 4% of the total U.S. population, would suffer long-term disability as a result.
“I think the main takeaway here is that even the most conservative estimates of long covid prevalence utterly blow the capacity of our healthcare, disability, and other safety net systems out of the water,” Melissa Smallwood, one of the report authors and a science policy researcher at Arizona State University, told Gizmodo.
No single study is going to be the final word on a given topic, and every piece of research has caveats. People who get VA care, for instance, tend to be older and sicker than the U.S. general public. Self-reported surveys can over- or underestimate the scope of a problem, depending on the questions you’re asking and how you’re asking them; they also typically don’t include a control group. And there is a laundry list of other factors that could influence the scales one way or another, such as the role of false test results, positive and negative, in skewing who should be considered a long covid case. Differences in the average age and pre-existing health of a population might make long covid more common in some countries over others. (Children seem to have a lower but non-zero risk of longer-term symptoms than other age groups.)
It’s safe to say that the earliest and largest estimates for long covid prevalence have overshot the mark, though. And there’s a decent case to be made that the true figure is on the lower end. In a preliminary review released this past May—which included the input of dozens of researchers around the world, including Ziyad Al-Aly—the authors analyzed dozens of studies and estimated that the global prevalence of long covid in 2020 and 2021 was around 3.7% of all infections. But this lowest of estimates wouldn’t mean that long covid isn’t a major issue.
“Even if the numbers are single digit, this is still a crisis,” Al Aly said. “It inherently has profound consequences on people’s lives.”
One of the many lessons of the covid-19 pandemic should be that a relatively small percentage of a big number is still a big number. Over 15 million people have likely died over the past three years from a virus that kills less than 1% of those it infects—which makes it the deadliest pandemic seen in over a century. If the above review’s math is accurate, it would still mean that at least 144 million survivors worldwide have endured mysterious and lingering problems during the pandemic. And the authors further estimated that 15% were experiencing symptoms a year later, amounting to over 20 million people. For some context, that’s about twice as many new cases of dementia estimated to occur worldwide annually.
Whatever range of numbers you want to have the most confidence in, none of the figures are negligible. But scientists like Bill Hanage have noticed disbelief and at times fierce pushback from some people on social media to the idea that long covid may not be as common as the highest projections suggest (anecdotally, so has this reporter). This criticism has even led to accusations that citing lower estimates of long covid is akin to downplaying the pandemic or outright claiming that the condition isn’t real. Even if misplaced, though, Hanage argues that this reaction comes from an understandable position.
“The people who’ve had post-viral fatigue syndrome—they’ve spent decades struggling to find a way to improve the attention that doctors give to them. And so the people with long covid are not talking about a thing which doesn’t exist, they really are not. They have reason to feel upset. And the dizzying back and forth we’ve had over the last few years from politicians and some public health outlets about the effect of the pandemic has led to a huge loss of trust,” he said. “I think that we need to approach people with compassion, and listen to them, and take them seriously, while also recognizing the larger estimates are not necessarily going to line up with the science.”
As other writers like Ed Yong have noted, the backdrop to all this is that we’re living in the shadow of defeat to the coronavirus.
Many countries, especially the U.S., have failed time and time again to respond to the pandemic with the urgency needed to prevent mass illness and death. Millions have died who didn’t have to, and many more have needlessly suffered, while existing cracks in our healthcare systems have only widened during this time. Even though we’ve been able to develop countermeasures like vaccines and antiviral drugs in record time, we haven’t distributed them equitably throughout the world—a delay aided by profiteering pharmaceutical companies that might have helped these newer variants gain a foothold. Bad faith actors, including the former U.S. President, have also continuously denied the severity of the pandemic and the need for mitigation. And some people are similarly eager to jump on any data that might help them dismiss the toll of long covid.
It’s important to be careful in how we interpret and communicate scientific data, especially when its implications can affect so many. All of us deserve the most accurate and up-to-date context for understanding the risks of covid, and to know when there are gaps in that understanding. People deserve to know that their risk of long covid isn’t zero, but they also shouldn’t be frightened into believing that their odds of avoiding it are no better than fifty-fifty.
That said, while covid-19 has become less likely to cause severe illness or death over time, thanks in no small part to the vaccines, it is still causing major societal harm. There are covid survivors in persisting anguish right this moment, while more will follow, so long as we do the bare minimum to “live” with the coronavirus. Even today, countries like the U.S. are dragging their feet on allocating enough funding to secure resources like widely available testing or updated vaccines better suited to current and future variants. And just this week, the White House stated that the newest Omicron variant BA.5 posed a serious threat to the country, yet simultaneously assured people that its current, mostly hands-off strategy would be enough to keep it at bay.
It’s reasonable for people to still want to avoid infection (which includes wearing a high-quality mask) and to want our leaders to do more. Those currently afflicted with long covid similarly deserve recognition and cooperation from the medical community to help them figure out this debilitating condition. That remains true no matter what the true risk of long covid may be.
“One of the main things that is needed right now is a push to recognize and research long covid, ME/CFS, and related conditions as biological illnesses,” Smallwood said. “In some ways, the covid-19 pandemic has set the stage for more research and recognition of the biological underpinnings of post-viral illness—ideally resulting in detectable biomarkers, diagnostics, clinical trials, and treatments.”