COVID Death Discrepancy for NYC Public Hospitals

Two sources that report COVID-19 Deaths for New York City's Public Hospitals in Spring 2020 disagree about when those deaths occurred

A discrepancy between sources that report COVID-19 deaths in New York City’s public hospitals in spring 2020 raises questions about whether the mass casualty event experienced by America’s largest city began sooner than officials have disclosed.

The two sources - a study on early intubation COVID-19 patients and the other a public dataset published by New York State Department of Health - each show time-series data from NYC Health + Hospitals’ (H+H), the agency that manages the city’s 11 public acute-care facilities. The study places the COVID death peak in those hospitals two weeks earlier than the state does.

I’ll show the discrepant data, explain why the discrepancy matters, and share my attempts to reconcile the differences. If the timeline for COVID deaths reported by the study is correct, a lot of officials have a lot of explaining to do. If it’s an error, then the study should be revised or retracted.

The State

I’ll start with the state’s data - specifically, COVID deaths reported by NYC H+H, which can be found in this file.¹ The file shows 2,633 COVID deaths reported between March 26th and May 30th between all H+H facilities,² which includes Elmhurst Hospital Center in Queens, the so-called “epicenter of the epicenter” of the city’s COVID outbreak.

Figure 1 shows the number and timing of the deaths peaking at 631 during the week of April 5th.

Figure 1

This timing and trajectory of the curve is consistent with the COVID and all-cause death peak for all city hospitals, as reported in CDC WONDER (the federal death database) and by NYC Bureau of Vital Statistics. So if it’s wrong (i.e., if the peak was earlier), the federal and state data are probably wrong too.

The Study (Parish, et al, 2021)

The second source for data on COVID deaths in H+H hospitals is Early Intubation and Increased Coronavirus Disease 2019 Mortality: A Propensity Score–Matched Retrospective Cohort Study (Parish, West, Caputo, Janus, Yuan, & Singer, 2021). The authors, all of whom are affiliated with the same H+H hospital, studied death rates among COVID-19 patients who were intubated and placed on mechanical ventilators in all H+H hospitals between March 1 and December 1, 2020. [AMENDMENT, April 16, 2024 - Per the Author Affiliations information in the article, lead researcher Austin Parish is also affiliated with the Meta-Research Innovation Center at Stanford (METRICS) at Stanford University. The co-director of METRICS is renowned scientist John Ioannidis.]

In supplemental table 2, the authors report several variables, including total weekly COVID-19 deaths. Figure 2 below is my graph of that data.

Figure 2

Clearly this is not the same curve as the state’s, despite being for the same hospitals and (presumably) gathered from the same hospital system.

The Discrepancy

Graphing the curves alongside one another makes the discrepancy more obvious:

Figure 3

There is a difference in weekly and total number of COVID deaths, with NYS showing 438 more deaths than Parish, et al. However, the study excluded COVID patients who died within 48 hours of triage, while the state appears to include all deaths occurring at the hospitals, whether inpatient or emergency department (ED). That distinction is significant enough to explain the gap between weekly and total number of COVID deaths.

More concerning, and much harder to rationalize, is the difference in timeline. Parish, et al show COVID deaths as early as the week of March 1st-7th, with a peak toward the end of March. The state’s file starts later, but shows a peak in the second full week of April. In the context of “15 days to slow the spread,” the difference is consequential.

An early March rise and peak like the one reported by Parish, et al, would change U.S. and New York officials’ claims about when the city’s excess-death period began — and about initial cases, hospitalizations, and deaths. Remember that the “first” case in New York City was announced on March 1, the first case hospitalized at a city hospital disclosed on March 2, and the first NYC resident COVID death publicized on March 14.³

Zealous use of mechanical ventilators in NYC hospitals is old news, yet officials still have not disclosed how many patients were placed on ventilators, or how many COVID deaths were people who were intubated and placed on mechanical ventilators at some point during their stay.⁴

State data show report ICU-intubated census for all city hospitals peaking the second week of April. Parish, et al present a different story. In their study of intubation in the public hospitals, most COVID patients who did not have DNI (Do Not Intubate) orders and were intubated, were intubated in mid-March. The city was effectively shutting down the second week of March, Governor Cuomo’s “stay home” order was signed on March 20th. So, if the Parish, et al timeline is accurate, it appears a significant number COVID patient intubations and deaths in H+H hospitals had already occurred. [UPDATE 4/16/24 - The federal government declared “15 Days to Slow the Spread” on March 16, 2020.]

Is it possible that both datasets are right, with each capturing overlapping timeframes?

Combining or stacking the data from each source shows a peak timing between the peaks of the two datasets, but a COVID death toll approaching 5,000 people (Figure 4), which would be roughly a third of deaths attributed to COVID in all NYC hospitals during these weeks. Frankly, that number is hard to fathom. The daily average census across H+H hospitals for the 2020 fiscal year was ~3,200 occupied beds.

Figure 4

From a body-management standpoint, losing over 150% of average census in less than two months is incredible, regardless of what caused the deaths, or the provision of more storage trailers.⁵ In retrospect, even 2,000+ deaths due to a virus with a fatality rate comparable to or lower than flu is questionable.

Attempts to Reconcile the Discrepancy

Toward figuring out which H+H COVID death-curve timeline is correct, I reached out to Dr. Austin Parish, the lead and corresponding author on the Early Intubation study. I showed him my graph of the data from his study versus the state’s data for the same hospitals. Dr. Parish acknowledged the time-shift but (like me) wasn’t sure what accounted for the difference. He was friendly and said nothing to suggest he thought incorrect data were supplied to his team, or that they made any errors with the data they obtained from H+H.

I then contacted H+H’s Office of the Inspector General. A staffer advised me to send my inquiry to a “Covid-19 Research Committee,” which I did three weeks ago, but still haven’t received a response.⁶

Is it possible that neither timeline is correct and/or that both timelines involve inconsistencies or tampering on the part of H+H and/or the state?

Perhaps — especially given the financial incentives involved. Surely, there were other deaths in H+H hospitals during these months that weren’t attributed to COVID. How many were not blamed on “novel virus”?

To find out, I submitted a Freedom of Information Law (FOIL) request to H+H on May 9, 2023, asking for the daily number of hospital deaths in their facilities, regardless of cause, between 1/1/2017 - 12/31/2022. The Deputy Records Access Officer recognized my request a few days later, but it was the last communication I received. No one has responded to my follow-up emails asking for an update on the status of my request. [UPDATE: I finally received data, but for 2020 only. NYC HHC denied my request for the 2017-2019 numbers.]

What's Needed

Anyone watching or reading the news in early 2020 knows that deaths in New York City hospitals were used as propaganda and “proof” that sudden spread of a novel deadly coronavirus was occurring and necessitated drastic, harmful protocols and policies.

If COVID-blamed deaths in the city’s public hospitals happened on a timeline that’s earlier or different from the one New Yorkers, Americans, and the world have been sold, the implications are potentially seismic, from a political and public policy standpoint.

If the presumed timeline is not in need of revision, then NYC H+H should contact Dr. Parish and colleagues, so they can revise their data or retract their study.

Either way, the state’s hospital data are incomplete and a long way from presenting the public with a complete picture and transparent view of what occurred.

1. Initially called HERDS - Hospital Electronic Response Data System Hospital Survey: COVID-19 Hospitalizations and Beds - the dataset was renamed New York State Statewide COVID-19 Hospitalizations and Beds on 11/4/2021. Readers may notice that the series begins March 26, 2020, which is curiously and inexcusably late. The file should start on March 1st, at least, if not January 1st or sooner. According to the dataset overview, “Hospitals began reporting for the HERDS COVID-19 survey in mid-March 2020.” When I inquired with the state health department data team about whether there are plans to report earlier data, staff replied, “There are no plans at this time to report data from earlier in 2020. It is likely the reporting system and the labs were not set up in time to report information from before that start date.”

2. Bellevue Hospital Center, Elmhurst Hospital Center, Harlem Hospital Center, Jacobi Medical Center, Kings County Hospital Center, Lincoln Medical & Mental Health Center, Metropolitan Hospital Center, North Central Bronx Hospital, Queens Hospital Center, Ruth Bader Ginsburg Hospital (South Brooklyn Health), and Woodhull Medical & Mental Health Center

3. At least one COVID death occurred on March 11, 2020, per data I obtained from NYC DOHMH via FOIL request. The first NYC resident COVID death was announced on March 14, 2020.

4. FOOTNOTE ADDED, 4/16/24: We Still Don’t Know How Many People in New York City Were Killed by Ventilators

5. The technical name for these these trailers is Body Collection Points (BCPs).

6. I’ve been unable to determine the current membership of the Covid-19 Research Committee. My understanding is that Dr. Joseph Masci (Chairman of Global Health, H+H/Elmhurst, now deceased) was once the committee chair and Dr. Nicola Davis (VP, Office of Population Health) and Dr. Michael Bouton (Chief Medical Information Officer) the co-chairs.