Discover more from Wood House 76
We Still Don’t Know How Many People in New York City Were Killed by Ventilators
The vent data we need are behind closed doors, and the data we have falls short of disclosing the full truth
Photo cred Medical_Nemesis
How much of New York City’s spring 2020 death toll can be blamed on ventilators — really?
We STILL DON’T KNOW.1
As far as I can tell, there’s no public data showing what daily ventilator capacity in NYC actually was, how many individual patients were placed on the machines, or how many deaths there were among ventilated patients. State and federal agencies are closeting the data we need, and the data we have is obscuring the truth.
You read that correctly.
Despite having absorbed a full-scale assault about the necessity of
missiles ventilators in New York, We the Taxpayers don’t have the most basic information about whether and how they were used.
Pretty strange, given we’re 3+ years past the event.
Many people certainly believe that thousands of patients died in New York City hospitals due to overzealous ventilator use. A “whistleblower” posted a viral confessional-style video on March 31, 2020, with concerns about how vents were being used to treat the disease. A study released in April 2020 showed alarming death rates, as did another study based on records from the city’s public hospitals, published later.2
And yet we never hear journalists, patient advocacy groups, medical professionals, or elected officials demanding that the state or any federal agency disclose how deadly the devices were (or weren’t) in New York City hospitals.
Why is that, when a purported 20,000+ New York residents died in hospitals during The Death Spike? How many were placed on ventilators upon arrival or during their stay? Has there been a Great COVID Ventilator Cover Up, as Michael Senger put it?
I can find only one public time-series dataset related to those questions. It falls very short of telling us what we need to know.
The Data We Have
It’s a daily census of COVID patients intubated in ICUs, shown below for the first five months of 2020.
The census has four prima facie shortcomings (i.e., it doesn’t pass a first-impression “smell test”).
Unclear definition. According to the dataset dictionary, the census addresses the question “Of the confirmed positive COVID-19 patients currently in the ICU, how many are intubated?”
The word “ventilator” isn’t used. Unlike Chicago, whose hospital census data says ventilators in use, New York’s data says intubated, without any connection or reference to a specific device or reason. While this probably refers to COVID patients on mechanical ventilators, it could also include patients intubated for other procedures or purposes. Is the term intentional obfuscation, or a semantics choice of no consequence? It’s not clear.
No comparison group. The file doesn’t report the number of non-COVID patients intubated in the ICU. This is also different from Chicago, which reports both the census for both COVID and non-COVID patients on ventilators. Is the number of ICU-intubated COVID patients higher than the normal number of ICU-intubated patients? We don’t know.
No baseline or context. The census starts on March 26th, which is almost a month after the city found its first COVID case. Without knowing the ICU-intubated census on each day in previous weeks (and months and years), irrespective of patient diagnosis, we have no sense of baseline. How many patients were ICU-intubated in January 2020 — or in January 2018, which was the peak of a “bad flu season”? Data can look scarier or different when it’s disconnected from any historical context. This is a “good” example of that.
Specious, potentially incentivized, start date. March 26th isn’t a random date. The CARES Act was passed on March 25th and gave certain reimbursements to hospitals for beds occupied by patients diagnosed with COVID-19, with a bump-up for patients who were mechanically ventilated. As I suggested in this thread, “pushing” data forward could have been motivated by such incentives.4
The 26th is also the day the New York public hospital system announced it had finished transitioning to a new electronic medical records system. During an emergency doesn’t seem like the best time to such changes. But the timing could be handy in the future, if there’s a need to blame missing data or data anomalies on implementing a “new system.”
The data also fall short mechanistically. That is, what we see is hard to reconcile with other data that speak to what was going on in real time and/or to what seems feasible — the “Could this even be possible?” question.
Patient intake issues. Taking the curve at face value, we see intubation of COVID patients occurred very quickly, plateaued, and then declined gradually. There were 956 patients intubated in ICUs on March 26th. Seven days later, the census doubled to 1,960 – and kept rising to a 2,713 peak on April 14th. A 183% increase is hard to square with emergency department, patient transport, and inpatient admission data that show fewer people coming to the hospital in these weeks than in prior months.
Whence the patients, if not from a mad rush of individuals? New York hospitals began testing in earnest on February 29, 2020. It’s possible that patients who were already in the hospital for other reasons comprise a hefty proportion of the census we see. Officials and media gave the impression that hospitals were overrun patients sudden onset of symptoms that required treatment, including intubation and ventilation. If that wasn’t the case — or if occupancy/census was pushed/pulled forward (i.e., post-dated) — Americans deserve to know.
No way to determine rate. Another shortcoming of the census data is it doesn’t show how many people were intubated and extubated each day. A 150-patient increase in COVID patients intubated from one day to the next could be 50 patients extubated and 200 patients intubated.5 The data we need to discern the total number of unique individuals who were placed on mechanical ventilators are (surprise!) not available. We can only see the daily census change (shown in the graph below).
No corresponding intra- or inter-hospital transfer or discharge disposition data. Finally, there’s no way to tell how many patients were extubated and then moved out of the ICU to Non-ICU beds (or vice versa) – or how many patients were transferred to another hospital, or discharged dead and discharged alive, respectively. Extrapolating rates from studies shouldn’t be necessary and has other limitations regardless. Patient records exist; the relevant data are no doubt resting in a high-security cloud. Where is it and why can’t the public see it?
As I told Will Jones of The Daily Sceptic, the peak census (n=2,713) can be regarded as a minimum number COVID patients ICU intubated that may have died from aggressive or inappropriate ventilator use. However, even a generous percentage of that peak census doesn’t explain the incredible speed and magnitude of New York City’s inpatient deaths.
The Death Data We Have
Hospital inpatient deaths alongside the intubation census paints an incredible picture.6
The above graph shows 19,329 inpatient deaths total between March 16 and May 31. That’s 277% more inpatient deaths than in the same days in 2019, with a 500% increase, from base to peak, in 21 days.7 Nearly all of the spike is attributed to COVID-19, despite the disease having no impact on daily mortality until “15 Days to Slow the Spread” was declared.
Disaggregating the daily COVID and non-COVID inpatient death timelines, we see non-COVID rose only slightly, for a week, which makes no sense in a “locked down” city of 8 million people. From early April through mid-May, non-COVID deaths dropped below baseline, implying that COVID attribution “stole” from other causes of death.
Because there’s a very strong relationship between the rise of the census for ICU-Intubated COVID Patients and the rise of inpatient COVID deaths, it’s tempting to say “See, the ventilators!”
Inappropriate or hasty intubation, Ventilator-Associated Pneumonia (VAP)/Ventilator-Induced Lung Injury, and/or the drugs used to sedate patients could have contributed to many deaths happening so quickly. Martin Neil, Jonathan Engler, Norman Fenton, and I offered such possibilities in a recent article. However, there’s no way to discern how many New York hospital deaths might involve these factors — especially because so many deaths attributed COVID-19 as underlying cause.
It’s also hard to fathom such a high rate of indiscriminate mass intubation, whatever the outcome. A testimony from a Manhattan ER doctor seems to corroborate the speed & intensity, but the tone of her story is odd and data for her hospital’s emergency room contradict her perceptions about ER patient volume.
Is it possible that many deaths occurred upon or very shortly after intubation (within 24-48 hours)? Perhaps sedatives administered in ambulances, emergency departments, or COVID ICUs created immediate casualties that were attributed to the Novel Disease or other causes?
Along those lines, consider what three weeks of inpatient death and intubation census data show:
There were an incredible 4,671 hospital inpatient COVID deaths in 21 days. Intubation census increased daily and reached 2,503 patients by April 7th.
If even half of COVID deaths were ICU-intubated patients on ventilators, it would mean massive turnover on those machines, with enough patients to quickly fill the beds. Again, this is hard to imagine during weeks when other data show volume was low. Without corroborating intake numbers or baseline data, the relationship between intubation and fatalities can’t be estimated.
Could immediate deaths upon sedation, attempted intubation and/or prior to ventilation – in addition to everything else that appears to have been going on — account for not only the magnitude but the speed of sudden deaths?
I’m not sure, but tend to think the answer is NO, which is one of many reasons the entire event should be substantiated with public release of death certificates.
How Many Vent Deaths? Still Hard to Say
So where does this leave us? Did ventilator (mis)use in New York City hospitals play an outsize role in the spring 2020 mass casualty event?
It’s still hard to say.
But they and others implicitly blame a) a sudden-spreading Novel Pathogen with unique etiology and additive mortality risk that was making people sick and compelling them to come to the hospital, and b) panicked doctors making mistakes as they learned how to handle the disease wrought by the Novel Pathogen.
If that’s the case, then where’s the raw data for the “epicenter” city where this pathogen “hit” hardest, with the impact of a bomb? How about an autopsy of what happened with ventilators in each hospital, so that the alleged chaos that ensued can be prevented next time? Surely, even those who believe that a coronavirus from Wuhan is responsible for those NYC hospital deaths would agree an independent review of patient records is warranted.
Remember, no third-party witnesses were allowed on the battlefield.8 The testimonies we’ve heard are from crisis actors, travel nurses, and head-scratching stories of patients going from talking on the phone one minute to being intubated the next.9
It’s a confusing set of scenes. Should we picture a disaster-medicine situation, akin to the one at Memorial Hospital in post-Katrina New Orleans — or dancing nurses and victory celebrations better capture what did (or didn’t) happen?
Again, it’s hard to say.
The Data We Need
What’s not hard to say is what kind of data we need to address basic questions about ventilator use in New York City hospitals. Examples include:
Daily census and inpatient admissions,
Number of patients intubated and extubated daily, with corresponding data on the reason(s) for intubation,
Daily ventilator capacity (by machine type),
Daily ventilator census,
Daily number of individuals patients placed on and taken off mechanical ventilators, respectively,
Intubation and MV status for each discharge, expired and unexpired.
Medication inventories for paralytics and sedative agents (e.g., propofol, ketamine, midazolam,10 fentanyl, phenobarbital, pentobarbital, etomidate, haldol)
Patients’ “COVID status” is less important than having each data point for all patients, as daily time-series, irrespective of how many hospitals or federal agencies labeled as COVID-19.
The only thing worse than the mental picture of thousands of people being killed by ventilators in New York City hospital is the obvious attempt to keep true numbers — which may very well be lower, not higher, than we think — in a closet.
Access my other posts on the NYC mass casualty event on this page:
UPDATE: I posted a corresponding Twitter thread with versions of some of the figures in this article that capture the essence of my observations (pasted below).
“Still” is also a nod to the times I’ve addressed this same issue at length: as a response to Michael Senger’s estimate October 2022 and in a letter to Will Jones on June 1, 2023. Readers with the stamina to read those explanations will note I’ve been consistent and earnest in my attempts to gather and makes sense of available data and evidence.
I wrote about discrepancies between data in the public hospitals study and the state’s data for those hospitals a few weeks ago.
Specifically, by the Office of Primary Care and Health Systems Management. This is the same office I emailed regarding discrepant occupancy data for Elmhurst Hospital but did not receive a response. I wrote about that and other issues with the file here.
Chicago’s hospital dataset begins earlier than New York’s, on March 19, 2020. Earlier data were not available when I requested it from Chicago DPH in 2021.
For an analogy, imagine a restaurant and the rate at which tables turn (i.e., how many times each table changes guests during service shift). The number of people seated in the dining room at the top of every hour during a dinner shift doesn’t tell you how many unique guests were served between 5:00 p.m. and 10:00 p.m., because the parties seated at each table stay for different amounts of time.
There are differences between the data I obtained from the city Bureau of Vital Statistics and federal data (CDC WONDER). I highlight some of those in another post.
Hospital inpatient deaths for March 16, 2019 - 5/31/2019 = 5,132 (Source: NYC DOHMH, Bureau of Vital Statistics, raw data obtained via FOIL)
[UPDATE FOR CLARIFICATION] I am not talking about journalists; I’m talking about patients’ own loved ones/advocates. Reporters were in some of these hospitals and failed to ask obvious questions or follow-up properly. For example, the NYT’s coverage of a panicked doctor at Elmhurst Hospital, which they NEVER fact-checked, even when the data were available. (Here’s my fact-check.) Media also participated in full-fledged films that are better characterized as mockumentaries than documentaries.