PMC COVID-19 Forecast, March 4, 2024
🔹 1.6% (1 in 63) are infectious, falling to 1.0% in a month
🔹 >760,000 COVID cases/day
🔹 >38,000 resulting Long Covid cases/day
Monthly and Long-Term Forecast with Weakening Public Health Guidance:
One Month Forecast
Over the next month, we should see transmission fall from 760,000 infections/day toward more like a range of 350,000-550,000 infections per day, depending on better or worse scenarios.
Note that the one-month projects are higher than last week because transmission fell less this week than anticipated. Alt Forecast #1 (Turtle) basically ignores the most recent week’s data, and shows you the path we would expect to be headed on if this week’s decline in transmission had been as large as anticipated. It’s possible the current week’s data get corrected, and we’re still headed on a path more like the green line, but this week’s numbers temper expectations.
The forecast suggests that a month from now about 1% of the U.S. population will be actively infectious on any given day, which is still very high transmission.
Acute hospitalizations and acute deaths continue to decline over the long-term, but shifting public health guidance has not accounted for long COVID, the development of new chronic diseases, and non-acute deaths (over 3-15 year trajectories) in any concrete data-driven way. High transmission during “lulls” remains a significant public health problem.
Longer-Term Trends (Speculative)
- Knowns - The PMC’s long-term position has been that forecasts beyond two months are highly speculative, and should be characterized as highly uncertain. New vaccines or treatments could emerge, or new subvariants could emerge and more rapidly. This section characterizes a few key knowns and unknowns. The key knowns are that over 4 years, the U.S. has had 8 COVID waves, including 2 per year. The spring or summer wave has typically been smaller than the winter surge. The lulls between each winter surge and the next summer wave have been marked by valleys with slightly increasing transmission levels at the low points. The “best of times” are getting worse. Simultaneously, the CDC has weakened isolation policies, again on Friday. Masking remains rare. Vaccine utilization is diminishing. Utilization of Paxlovid remains low. The protective impact of vaccines against the development of long COVID remains debated due to substantial uncertainty in point estimates as well as variation in timelines considered, as studies often focus on recent vaccination, and much of the public has not gotten vaccinated recently. Public health metrics are inadequately capturing long COVID, non-acute hospitalizations, emerging trends in the development of new diseases, and post-acute deaths.
- An Uncertain Summer Ahead - What appears likely is that barring any unforeseen changes in vaccine development or subvariants, the U.S. will see at least two waves during the remainder of the year. The timing of a summer wave is variable. The timing of the winter wave, or more likely “surge,” has thus far followed a relatively predictable historical pattern. It appears likely that the “low point” of transmission for the remainder of 2024 will be higher than the low point in and prior year, again barring a market-busting new vaccine. One key question is whether we will see three waves in the remainder of 2024. With public health policy very weak at the moment, declining personal precautions, and higher “lull” transmission, there is a reasonable (perhaps 30-50%) chance of a small wave in the late spring or early summer, followed by a more traditional wave in the late summer or early fall, and surge in the winter. In contrast to the 3-wave scenario, we may simply see high “low point” transmission the next couple of months before the summer wave kicks in, followed by the winter surge. Whether we see a “small” late spring or early summer wave is a bit of a moot point in the big picture because transmission may hover around 400,000-600,000 infections per day regardless, and it’s more a matter of whether that longitudinal line of transmission is relatively flat or bumpy. No data on vaccines, infection histories, or “hybrid immunity” speaks against these fundamental trends in merely modeling the data.
- Expect Continued Waves Long-Term with More Volatility - Over the very long term, the decline in mitigation suggests that waves and surges will continue to remain likely, barring market-busting changes in vaccines. If there have been two waves each year, a reasonable bet is that there will be two waves this year and two waves next year. Due to the decline in mitigation, however, it could be argued that the time point and size of waves may become less predictable. Low mitigation increases the risk of a BA.1-style surge or surge half that magnitude. Low mitigation increases the probability of off-cycle waves, multiple waves, or post-peak hills in transmission. Low mitigation also reduces the chances of making COVID a truly seasonal (i.e., winter-surging only) virus because governments are not pushing down the summer wave through more frequent vaccination, better isolation policies, more available testing, and better air cleaning.
- Weakened Isolation Policies Mean Stepping Up Other Mitigation - The decline in the isolation period suggests that wise organizations and individuals should increase other precautions. Hospitals and other public settings should improve masking policies to assume non-isolating contagious individuals will be present. Organizations and individuals should upgrade ventilation and filtration where possible. Exclusively in-person only events should offer remote options, as highly-contagious individuals are no longer expected to take reasonable isolation precautions if asymptomatic or symptoms are subjectively improving. Testing programs should be reintroduced at events to further reduce transmission risk.
PMC COVID-19 Forecast, March 4, 2024
You can access the full PMC dashboard and report online at
Please share anything you like across platforms. Never a need to ask permission. 🙏pmc19.com/data
• • •
Missing some Tweet in this thread? You can try to
force a refresh
31 Reasons Why the New 1-Day COVID Isolation Policy is Wrong
#1
Experts in modeling and testing know that people are infectious with COVID for an average of 7 days, with substantial variability around that average.
#2
People use defense mechanisms to temporarily avoid the death anxiety evoked by thinking of COVID. The too-short 5-day iso was an example of this (see final example).
Such defenses provide temporary relief and are almost always harmful long-term.
PMC COVID-19 Forecast, Week of Feb 26, 2024
1 of 5 🧵
Forecast for the next month
Over the next month, we should see transmission fall from 790,000 infections/day toward more like a range of 200,000-450,000 infections per day, depending on better or worse scenarios.
That's "good" news in the relative sense for those putting off medical appointments the past 6 months, though still extremely high transmission in any objective sense.
See the online report for details on the models.
Surge in Context
At this point in the surge, it is clear that the peak transmission day was around December 27 (1.92 million/day), and the midpoint of “surging” infections (>1 million/day) was around January 9.
We are estimated to have had 85 total days with >1 million infections per day (November 28 through February 20) during the surge, though these numbers may still fluctuate with corrections the next few weeks.
The low-point leading into the surge was October 18 at 547,000 infections/day. Infections have been at “wave” levels (>500,000 infections/day or higher) since the onset of the late summer wave surpassed that milestone on July 27. We are estimated to dip below 500,000 infections/day around March 6.
This is very unfortunate timing because the medical facilities that enacted universal masking may end policies on March 1. Many were hoping for a period of lower transmission before such policies ended. As of today, the estimated low point for transmission is March 27 (348,000 infections/day), but the level and date are subject to much uncertainty.
PMC COVID-19 Forecast, Week of Feb 26, 2024
2 of 5 🧵
Current State of the Pandemic
🔹73 million infections in the U.S. in 2024 (so far)
🔹790,000 daily infections
🔹1.66% (1 in 60) actively infectious
🔹40,000+ resulting #LongCOVID cases/day
Deeper Dive
Transmission is finally starting to decline again, and expect major declines the next four week.
U.S. wastewater levels indicate that COVID transmission is higher than during 58.4% of the days of the pandemic (down from 85.9% a week ago). Transmission is lower than 41.6% of the pandemic.
As we noted the past two weeks, we believed the post-peak hill was itself peaking on around February 7th and that last week’s slightly higher values might get retroactively corrected downward. That was, in fact, the case (the peak was the 7th), and transmission has fallen further since.
We are still at very high “wave” levels, but no longer “surging” at over a million infections/day. The big picture remains very bad, but this is good news for people putting off medical appointments for months.
PMC COVID-19 Forecast, Week of Feb 26, 2024
3 of 5 🧵
Risks in Group Settings
Although transmission is falling, it's easy to get distracted by the relative changes and ignore that the absolute risk remains high, especially in large groups with limited or no mitigation.
In a group of 10, there's a 15% at least one person is actively infectious. In a group of 30, it's a 40% chance, and so forth. Almost nobody would take those chances of a serious illness if informed and capable of grappling with the seriousness of that risk without becoming defensive. Unfortunately, a lot of institutions are pushing minimizer narratives if not directly forcing students and workers into more dangerous settings.
Dr. Moriarty & other modelers know people are infectious for an average of about 7 days, per high-quality studies. Many for much longer.
Dr. Mina's pinned Tweet lays out a sample timeline.
Sending kids to school on Day 2 positive will essentially maximize infections. 2/4
The consequence of the California 1-day isolation policy is that many parents and grandparents will develop serious health conditions and too often die prematurely.
Bad for families. Good for inspiring the next generation of bereavement workers.
1 / 6 🧵
PMC COVID-19 Forecast, Week of Feb 5, 2024
We are seeing escalating transmission in a post-peak hill.
🔹2.8% (1 in 36) actively infectious
🔹1.3 million infections/day
🔹Hill peaks in 2 days 🤞
🔹>65,000 resulting #LongCOVID cases/day
2 / 6 🧵
PMC COVID-19 Forecast, Week of Feb 5, 2024
We began to see evidence of a post-peak hill 3 weeks ago. I was skeptical. The real-time data now bear this out. Hopefully we are not in a Terminator-style scenario where the technology outsmarts the maker. 🤣
3 / 6 🧵
PMC COVID-19 Forecast, Week of Feb 5, 2024
Zooming out to the full pandemic, we see transmission higher than a week ago, and higher than 87.8% of the pandemic.
Nearly 10 million infections/week. >50 million estimated infections so far in 2024 in the U.S. alone.
Transmission in the U.S. is extremely high & stable from last week on the back of the 2nd largest surge all-time.
🔹2.5% (1 in 39) actively infectious
🔹1.2 million infections/day
🔹>60,000 resulting daily LC cases
🔹0.8-1.3 million daily infections the next month
Detailed Forecast:
We are in the 2nd largest U.S. surge of the pandemic. Over the next month, we should see 0.8-1.3 million infections per day.
Biobot has made considerable post-hoc corrections to each of the prior 4 weeks of data. This essentially moved the peak earlier, with real-time estimates corrected downward in hindsight the past several weeks, including last week’s numbers getting corrected downward by 3%. The details of the forecasting models can be found in the Technical Notes section in the online report.
Post-Peak Hill? Reviewing the graph of the whole pandemic (Tweet #3), you’ll notice that after about half the waves, instead of a straightforward decline, there is occasionally a temporary deceleration, or even a small hill. Currently, this is what the forecast is showing. It’s unclear whether we will see a very small hill (rebound), a couple weeks with a flat plateau, or just descend at a gradual pace, but it looks like there will continue to be more transmission on the back end of the mountain than the start, as the models have continued to forecast for a long time now.
Good News: The model suggests that around Valentine’s Day (Feb 14), transmission should begin to fall for rapidly. The good news, if any, is that we can begin to discern the “end” of the surge toward late February, when daily infections will likely drop below 1 million nationwide.
Bad News: The bad news is that even the apparent "end" will still have a long ways to go down. For example, note that the starting point of the blue line on the left of the graph is actually the peak of the late-summer wave, which was quite bad.
Technical Note: If you go to the Biobot Dashboard, you'll see levels down 4% from last week, but our model shows levels marginally higher. For Biobot's Jan 27 data point, we enter that as Jan 24 (it's actually the average for the preceding week), assume that data point is a very slight underestimate (pattern the past several weeks), and model that we're already starting to tick up into that little hill. I'd avoid over-interpreting, and just take this to mean transmission is pretty stable at the moment.
2/ What's with the Post-Peak Hill?
We're anticipating a post-peak hill. However, I wouldn't be surprised if it's instead a temporary plateau (flat) or simply a deceleration.
The best explanation is that transmission remains high with considerably geographic heterogeneity.
Examples of heterogeneity: 1) Single peak. Many places appear to have a single winter peak this year. In Western Mass, they peaked with 38% of the population actively infectious. Nearly unbelievable. It's hard to imagine a 2nd hill after that. Many other places peaked around 10% actively infectious on the worst day.
2) Late peak: Some places haven't peaked yet. This could account for a post-peak hill.
3) Rebound: Some places had an apparent winter peak but are starting to rebound. The rebound may be marginally higher or lower than the existing apparent "peak." Likely, the earlier peak was lower than in places experiencing a true single peak. They "flattened the curve" a little, but it's a tough fight against in-school transmission.
4) Rollercoaster. More extreme example of the Rebound case. Several peaks and valleys.
5) Flat. Some places have looked pretty flat on transmission for months. Perhaps they are wiser on engineering controls, or perhaps they are approaching a late peak.
3/ Zooming out to the full pandemic:
🔹Stable transmission relative to last week
🔹Transmission remains higher than during 85% of the entire pandemic, i.e., not "over"
🔹We are still in the 2nd-highest wave all-time with >1.2 million U.S. infections/day