Source: US Centers for Disease Control.
The upshot of this posting is that flu doesn’t much matter, in terms of crowding of hospitals during the COVID-19 pandemic. When flu season comes around this year, if we’re still in the COVID-19 pandemic, the flu will add a bit of stress to the hospital system. But only a bit.
Best guess, based on a variety of sources, the impact of the peak of a bad flu season, on hospital inpatient resources in any one state, will be maybe 5% of the size of the impact of the peak of the COVID-19 outbreak. So far. That’s based on number of hospital admissions, concentration of those admissions within states, and average hospital resource use per admission. A bad flu season certainly won’t help things. But it’s not really a make-or-break issue in this context.
In terms of things we need to worry about, I’m putting the influence of seasonal flu far, far down my list. That, no matter what offhand remarks I may hear by public health officials about the dangers of a simultaneous COVID-19 and flu season.
Nationally, the raw numbers make it look like flu matters materially.
This post started out as an analysis of the effectiveness of flu vaccine. But before I can get to that, I needed to know whether flu mattered in-and-of-itself. How does the peak of the flu season compare to the peak of a COVID-19 outbreak, in terms of stressing hospital inpatient resources.
A bad flu year can fill a lot of hospital beds. This is among the reasons why, prior to the recent outbreaks (CA, AZ, TX, and others), responsible government officials were pointing toward renewed danger of overrunning hospital capacity this winter. They are worried what may happen if the peak of flu-season demand for hospital resources occurs at a time of great COVID-19 demand for those same resources.
Over the past decade, the worst flu season peaked with just over 76,000 hospital admissions per month for flu, in January 2015, for the US as a whole. That’s in addition to roughly a third of a million hospital OPD/ER visits.
How does that compare to the rate at which we’ve seen admissions for coronavirus? COVID-19 hospitalization data are spotty (some states report, some don’t). Here, I’m relying on the CDC’s surveillance network, which consists of all hospitalizations from a well-established sample of hospitals, inflated to give US total estimates. Here’s there estimate of hospitalizations per 100,000 per month:
The graph isn’t impressive until you do the math. That peak admission rate of 10 per 100,000 population per week, in April of this year, works out to about 140,000 COVID-19 admissions per month, at the (prior) peak of the US pandemic in April 2020.
The upshot is that the peak month, of a bad flu year, can add to US total hospital admissions, at just over half the rate we’ve recently seen for COVID-19, at the national level. And, because these are respiratory infection admissions, they use more-or-less the same medical personnel and resources as the COVID-19 cases do.
Those raw numbers — national totals, simple count of admissions — certainly make it look like flu matters. It makes it look like a bad flu year would greatly complicate the hospital inpatient picture in the middle of the COVID-19 pandemic.
But those raw counts, of national data, are substantially misleading in that regard.
Getting from national raw counts to state-level resource-adjusted counts.
The analysis above exaggerates the implications of flu, for hospital resource use, relative to COVID-19, for two key reasons.
First, flu cases have much shorter stays, on average. The average length-of-stay for influenza discharges is just over 4 days, with a median (typical case) of just 3 days (calculated directly from the AHRQ HCUPnet database, at this link). By contrast, the US experience (here, California and Washington), suggests that the typical COVID-19 hospitalization takes about three times as long. A typical flu discharge takes one-third the hospital inpatient days of a typical COVID-19 discharge.
Consistent with that, flu cases appear far less severe than COVID-19 cases and so less resource-intensive overall. (I.e., not just bed-days, but like other factors as well.) The simplest measure of that is deaths per admission. In the US, about 1.7% of flu admissions result in death (AHRQ, calculated, same source as above). Whereas about 22% of persons admitted to the hospital for COVID-19 die there (California and Washington study, again).
Arguably the best way to compare overall hospital resource use per case is to get some estimate of cost per admission. For discharges with principal diagnosis of flu, the AHRQ HCUP data (referenced above) shows about $9000 mean cost and $5000 median cost in 2014. Inflating those to 2020 at about 2 percent/year yields a cost estimate of $10,100 mean and $5,600 median cost per discharge. Then, FWIW, here’s a theoretical estimate of a median cost per hospitalized COVID-19 case of about $14,000. (Or a little less than 3x the median flu case). A separate way to get at this is to look at the average Medicare payment per case of $23,000 per discharge for COVID-19 cases (from this source). Adjusting for negative overall Medicare inpatient margin of about 9.3% (MedPAC, 2020) yields an estimate of cost per discharge of about $25,500. (Or, again something less than 3x the average flu cost cost). This latter estimate does not account for differences between the COVID-19 cases and other cases in the Diagnosis Related Groups to which those COVID-19 cases are assigned.
As a good working number, then, the average flu discharge is only “worth” a third of an average COVID-19 discharge, in terms of hospital resources consumed.
From that perspective, at the US national level, the peak month of flu season, in the peak year of recent years, added about (76,000 * (1/3) /140,000 = ) 18% of the total COVID-19 hospital burden during the April peak of the US pandemic.
Second, and perhaps more importantly, the national flu numbers above reference a disease that hits all the states, at roughly the same time. The US peak flu hospitalization rate reflects cases that are reasonably evenly spread across the country. So the US peak matches the typical state peak. Here’s our most recent flu season, showing start, beginning, and end flu prevalence, from the US CDC. The brown middle map shows that flu was widespread in all 50 states, at the same time.
Flu discharges in New York State: Roughly speaking, then, to get from that national peak number (76,000 flu admissions) to a state number, you can more-or-less just pro-rate it based on state population. If I were do to that for (e.g.) New York state, I would estimate that the peak month of flu cases (nationally) generated just 6,400 flu admissions per peak month in New York State. And, unsurprisingly, the most recent record-breaking flu outbreak in New York State resulted in hospitalizations at roughly that rate — around 6,200 flu admissions in February of this year. (Estimated from the peak weekly rate cited in this reference.)
By contrast, peak outbreaks of COVID-19 have been highly geographically concentrated. For the April COVID-19 hospitalization peak referenced above was, New York State accounted for roughly 55,000 hospitalizations (estimated from this source), or close to 40% of all COVID-19 hospitalizations during that peak. In other words, within New York State, peak observed flu hospitalizations per month (6200) amount to about 11 percent of observed peak COVID-19 hospitalizations per month. So far.
But if each flu hospitalization is “worth” only a third of a COVID-19 hospitalization, then the peak impact of flu, on hospital resources in New York, was only about 4% of the peak COVID-19 impact.
This is a rough cut, New York was the worst-hit state so far, so in the spirit of no false accuracy, let me round that to 5%. Best guess, if peak flu season corresponds to a COVID-19 hospitalization peak, it will add about 5% to overall hospital inpatient resource use, over and above the COVID-19 burden. In a state that is facing a seriously large peak in COVID-19 cases relative to hospital capacity. As was the case in New York, in April.
Sure, the simple national numbers make it look like flu is in the same ballpark as COVID-19, in terms of stressing hospitals. But if you look a bit further, that’s not even remotely true. That’s not true because the average flu case uses only about a third of the days and resources of the average COVID-19 case. And that’s not true because the peak of COVID-19 hospital use (so far) has been highly concentrate in a few states at a time. While the national data for peak flu hospital admissions reflects, more-or-less, a simultaneous peak in most or all of the states.
When you break it down to the likely impact on hospital inpatient resources, at the state level, flu is not exactly rounding error, but it’s close. The real danger to overwhelming hospital inpatient beds is from COVID-19. Compared to the magnitude of that, the burden of flu barely registers.