Every day, the state releases the percentage of different groups of New Yorkers who are vaccinated against COVID-19, sometimes accompanied by statements trumpeting the achievement of major milestones.
But those percentages — such as the more than 91% of adults the state says are at least partially vaccinated — are probably inaccurate and inflated because of double-counting, population shifts and other factors, experts say.
Some vaccination rates are mathematically impossible to achieve. For example, the state says 153,621 Nassau County residents 65 to 74 had been vaccinated by Friday, even though, according to a state website, only 132,862 people in that age group live in Nassau. That would mean a vaccination rate of 116%. Rates for residents 55 to 64, and 75 and older, also exceed 100%.
Denis Nash, a professor of epidemiology at the CUNY School of Public Health in Manhattan, noted that under former Gov. Andrew M. Cuomo, the state released COVID-19 numbers that were later found to downplay the severity of the pandemic in New York.
"I know we’re under new leadership [Gov. Kathy Hochul] now, which is great, and we have a great new health commissioner coming in," Nash said, "but the state doesn’t have the best track record in ensuring confidence in the numbers they’re putting out there. They tended to try to make it look better than it was. We don’t want to see that happening for something as important as the vaccine numbers."
What to know
The state’s official vaccination rates are likely inaccurate, experts say. Some rates are mathematically impossible to achieve, such as the 116% rate for Nassau County residents 65 to 74.
The problems probably are caused by factors such as some names and birth dates being entered incorrectly, leading to double-counting, and by changes in the population used to calculate rates.
Discrepancies are common outside New York as well. The Centers for Disease Control and Prevention lists a nationwide vaccination rate for seniors of 99.9%, even though some states report rates below 90%.
New York State Department of Health spokeswoman Erin Silk said in an email that numbers on its vaccination data website are "accurately reflecting the department’s tracking of vaccinated individuals."
Silk said the issue may be with U.S. Census Bureau numbers that the state uses to calculate the vaccination rate. State calculations are made by dividing the number of people in a certain group who are vaccinated by projections for 2018 that are based on the 2010 census.
The 2018 projections are often lower than the populations in the Census Bureau’s 2019 American Community Survey estimates, which are based on an annual survey of more than 3.5 million households nationwide. Detailed population data is not yet available for 2020.
Using a lower population estimate leads to a higher vaccination rate.
Silk said the state doesn’t use the American Community Survey because it incompletely counts "congregate settings." But, according to the Census Bureau, the survey has been counting people in congregate settings such as nursing homes, student housing, prisons and homeless shelters since 2006.
Nash said accurate numbers are especially important as COVID-19 case numbers climb with the arrival of colder weather and the holiday season.
"It’s the proportion of people who are vaccinated that we’re really relying on to know how well we’re going to be able to weather another surge this winter," he said. "If that is not reliable, if it’s wildly underestimated or overestimated, and especially overestimated, the consequences are dire."
Melody Goodman, associate dean for research and associate professor of biostatistics at the NYU School of Global Public Health in Manhattan, said accurate vaccination rates are important in determining "the potential population that is at risk, how much potential there is for community spread" of the coronavirus, and what steps are needed to slow the spread.
Discrepancies not just in NY
The vaccination rates over 100% for groups such as seniors and Asians suggest that other vaccination rates the state releases likely are inflated, Nash said.
But the discrepancies in numbers are not limited to New York. Across the Hudson River, multiple New Jersey municipalities report rates of more than 100%. In Hoboken, for example, 114% of those 65 and older have at least one dose, and 247% of kids 12 to 17 are at least partially vaccinated, according to New Jersey Department of Health data.
And the Centers for Disease Control and Prevention currently reports that 99.9% of adults 65 and older have at least one dose — despite individual states that report numbers well below 90%.
Nash said some overcounting likely is from incorrect names and birth dates being entered. For example, someone’s name may be spelled one way when receiving a first dose and slightly differently when getting a second dose. "If they don’t recognize it, I’m being counted twice," he said.
Mistakes are more likely if the person was inoculated at different sites each time, he said.
Double-counting would lead to an inflated single-dose vaccination rate but an undercount of the rate of those fully vaccinated, Nash said.
State data shows a 103% vaccination rate for Long Islanders 75 and older with at least one shot, but a rate of 87.9% for those in that age group who are fully vaccinated.
Sean Clouston, an associate professor of public health at Stony Brook University, said people who received booster shots before they were officially authorized by the federal government may have said they were getting their first shot, to evade scrutiny. "They would be listed as a first vaccine," he said.
Likewise, in the first few months of vaccination, when there was far more demand for vaccines than there were appointment slots, some people trying to get a shot at a vaccination site with a residency requirement gave an incorrect address, said Stephanie Silvera, an epidemiologist and professor of public health at Montclair State University in New Jersey.
"I don’t know if all the sites are checking that carefully in terms of ID," especially early on, when sites were overwhelmed with people, she said.
Varied populations, data terms
Goodman said population estimates that are different from the actual, current population likely are the main cause of the discrepancies and of numbers that exceed 100%.
There always are changes in population, but the pandemic caused an especially large shift, Clouston said. Some Manhattanites, for example, fled to the Hamptons during the spring 2020 COVID-19 surge and never returned, he said. They likely were vaccinated in Suffolk County but were counted as Manhattan residents in pre-pandemic census data.
The population increase in southeast Suffolk is why for months some ZIP codes there have had vaccination rates of more than 100%, Clouston said.
Discrepancies also might stem from the Census Bureau and state using different terminology in asking for demographic data, such as for multiracial people, Goodman said.
Asians are the one racial or ethnic group for which percentages exceed 100%. On Long Island, 106% of Asians 15 and older are vaccinated, according to state Health Department calculations.
Using the broadest 2019 population estimates definition of "Asian," which includes Asians who identify with one or more races in addition to Asian, the vaccination rate for Asians on Long Island drops to 99%.
The discrepancies between the CDC’s numbers and those from various states likely stem from some of the same problems as the New York data, including double-counting and population changes, Nash said.
He said a 99.9% vaccination rate for seniors seems "implausibly high" because of widespread vaccine resistance and because of the large number of older unvaccinated people who have been dying of COVID-19.
The CDC officially lists the nationwide vaccination rate for those 65 and older as 99.9%, but it may be higher than 100%, because 99.9% is the maximum the CDC will report, CDC spokesman Scott Pauley said in an email.
He acknowledged "potential data reporting errors," as well as changes in population data, as possible factors.
The CDC receives only "de-identified" data of vaccine recipients, with a "recipient ID" but without names, to protect privacy, Pauley said. As a result, the agency "is not always able to link first, second and booster doses for people who received doses in multiple jurisdictions or at different providers within the same jurisdiction," he said.
Silvera said she doubts nationwide and statewide vaccination rates "are ever going to be accurate. You’re essentially asking for a complete and perfect census of everybody who’s vaccinated and everybody in the country. Neither one of those I think is going to be feasible."