symptoms of pneumonia and tested positive for Covid-19.

This is Domingo. But not just him. This light represents 500 people who have been infected

with the novel coronavirus in New York City. By May 16th, there were nearly 190,000 known

cases of Covid-19 there. Health officials report these numbers every

day, in cities and countries around the world, but they know that they're incomplete. Because

Covid testing has been like a narrow flashlight in a dark room. Anything we're not pointing

the light at, we can't see.  But now researchers are collecting data that

can capture the pandemic more fully, to try to get a better handle on just how much we've lost.

Domingo Vega died on April 16th at the age

of 45. Originally from Mexico, he came to the US when he was 16, and worked in restaurant

kitchens since then, eventually launching his own business with two locations in Brooklyn.

He and his wife had three children.  Domingo was one of 20,720 New Yorkers whose

lives were cut short by the coronavirus as of May 16th, according to the city's count.

Each blue light here represents 500 known deaths from Covid-19 and 500 families who

don't need data to tell them how dangerous this disease can be.

But when it comes to the statistics, that question -- how deadly is Covid-19 -- has

been difficult to answer. The relationship between the known deaths

and the known cases is called the “case fatality rate.”

At this point in New York City, 1 out of 9 people known to have Covid-19 have died.

That's 11 percent. But that rate varies drastically across cities

and countries. It was over 12% in Sweden in mid May but less than a percent in Iceland.

It also changes over time. For the US, it dropped down to 1% near the end of March before

climbing back up as people who tested positive several weeks prior ended up dying.

When the case fatality rate varies this much, it's saying a lot more about these countries

than about the disease itself.  For example, deaths may be higher in places where

the health system is overwhelmed  or where the population is older. We know

that Covid-19 is more deadly in seniors, and especially those over 75.

But also, on the other side of the fraction, the rate reflects how much testing is happening.

If a country is aware of more non-fatal cases, their case fatality rate is lower.

So this statistic isn't all that useful because we know most countries are missing

cases. We're also missing deaths. According to an estimate by the New York Times,

there have been thousands of deaths that weren't included in the official count for New York

City.  We don't know for sure if it was coronavirus

that killed them. But here's what we do know.

If you look at 2017, 2018, and 2019, and chart the average number of deaths per week, the

line looks like this. It includes deaths from all from all causes.

For 2020 so far, that line of weekly deaths looks like this. The area above the typical

level is called “excess deaths” by researchers. And it gives us a fuller picture of the cost

of this pandemic. We're seeing excess deaths in many places

that have suffered big outbreaks.  In each case the excess deaths are higher

than the official count of Covid deaths.  It includes people who may have died from

other causes but who were unable or unwilling to access medical care because of the pandemic.

But it also includes some people with coronavirus who may have died at home or care facilities,

or were never diagnosed. As this data comes in, it shows that in some

places, the pandemic is even more deadly than we thought.

But the virus itself may be less deadly. Because we're also learning that a lot more people

have been infected than the official tallies show.

Health officials in New York have taken small blood samples of people at grocery stores

to check for Covid antibodies. This is called a “seroprevalence survey” and it helps

capture the substantial number of people who didn't know they ever had coronavirus.

So far these tests show that around 20% of people in New York City tested positive for

antibodies that indicate a previous Covid infection.

If those shoppers are representative of the city's population, that would mean there

were more than one and a half million Covid-19 cases in the city by early May.

Without changing how many people have died, the antibody survey lowers the fatality rate

by identifying more non-fatal cases. Remember the case fatality rate was 11% one

in 9, but the fatality rate for all those infected may fall somewhere between 1 in 60

and 1 in 90 for New York City. So while the death count is higher than we

thought, the death rate may be lower. But a low fatality rate is not all good news.

It paints a picture of a tricky virus that moves undetected through many of us and causes

immense suffering and death in others. We can look for comparisons to try to wrap

our heads around the death toll - more people lost in three months in the US than a year's

worth of car crashes or drug overdoses. Still fewer than annual deaths from cancer or heart

disease.  But the comparisons are limited. Because unlike

car accidents or cancer, Covid-19 is contagious. Human beings are the vector for this disease

and their actions are hard to predict. So even with better data about how many are

infected and dying, we won't know the full death toll of this pandemic

until we find out how it ends.