Everything You Need to Know About COVID-19

COVID-19 is no longer a nationally notifiable disease in the United States, most testing happens at home and is never reported, and the surveys built during the emergency have been retired one by one. That has a specific consequence for anyone looking for an answer online: there is no current US case count, no current national case-fatality rate, and no current population hospitalisation rate. Sites that publish one are dividing one modelled estimate by another.

What still exists is a set of narrower measurements that are genuinely good, and this site is built around them rather than around the numbers that no longer exist. CDC's most recent full-season modelling — for July 1, 2024 to June 30, 2025 — estimates 10.4 to 16.7 million illnesses and 34,000 to 53,000 deaths in the United States, and puts the case-fatality rate among symptomatic illnesses at 0.2% to 0.51%. Those are ranges, not points, and we publish them as ranges.

Globally the position is thinner still. WHO estimates that about 14.9 million excess deaths were associated with the pandemic worldwide in 2020 and 2021 (range 13.3–16.6 million) — but that is a cumulative two-year figure, and since global surveillance wound down WHO no longer publishes a clean annual worldwide case or death count. Our COVID-19 Risk Perspective leaves those cells blank, with the reason, rather than charting a two-year total beside a single US season as though they were comparable.

WHO has assigned no new Greek letter since Omicron in November 2021. There is no Pi variant and no Omega variant. As of WHO's tracking page, last updated 28 July 2026, there is no currently circulating variant of concern at all; the single variant of interest is JN.1, and four lineages are under monitoring — PQ.16.1.1, NB.1.8.1, XFG and BA.3.2.

In the United States, CDC's most recent published estimates cover the four weeks ending August 1, 2026 and list 18 lineages, led by SW.2 at 20.7% (95% CI 16.4–25.8%) and XFG.1.1 at 15.8% (12.0–20.5%). All 18 carry CDC's own flag for estimates based on fewer than ten sequences or otherwise unreliable by NCHS standards — two years earlier, 21 of 43 did. CDC's model-based Nowcast projections stopped in April 2026. The honest summary is that several JN.1-descended lineages each account for somewhere between a twentieth and a quarter of US infections, and no one can currently say which leads.

Read the full Variants page — the complete lineage table with confidence intervals, what recombination changed, and why the Greek letters stopped →

Paxlovid's EPIC-HR trial cut hospitalisation or death from 7.01% to 0.77% among unvaccinated high-risk adults treated within three days — a 6.32-percentage-point absolute difference (95% CI −9.04 to −3.59). Remdesivir's PINETREE trial found 2 of 279 versus 15 of 283 (hazard ratio 0.13, 0.03–0.59). Molnupiravir's MOVe-OUT reported a 6.8-point benefit at its interim analysis and 3.0 points in the full population, with a confidence interval running to −0.1 — which is why it sits third and is authorised only when other options are unsuitable.

All three enrolled unvaccinated adults in 2020–2021. Almost no one treated today matches that description, so the absolute benefit now is smaller. Dexamethasone's RECOVERY trial showed the same lesson in reverse: a large mortality benefit in ventilated patients (rate ratio 0.64), a smaller one on oxygen (0.82), and a point estimate favouring usual care in patients needing no oxygen at all (1.19, 0.92–1.55).

Read the full Treatment page — the trials side by side, the molnupiravir number that halved, and why "remdesivir vs Paxlovid" has no trial answer →

Every figure on this site traces to a primary source read directly: a CDC page or public dataset, a WHO document, an FDA decision, or a peer-reviewed paper cited by DOI. Where a source is silent, the page says so rather than filling the gap. Where two good sources disagree, both appear with an explanation of why.

Four things are genuinely not known and are stated as such throughout: how many people in the United States are infected in a given week; whether the lineages now circulating differ from one another in severity; what proportion of infections are asymptomatic today; and how common Long COVID currently is. Each has a reason — a retired surveillance system, a sequencing sample too small to support the analysis, an instrument that stopped collecting.

Our How We Research page shows the method on three worked examples and states what we refuse to compute. When we get something wrong, the corrections log records what was published, what it says now, and the source that settled it — including three corrections made during this September 2026 review.

Sources & References

1. CDC. Respiratory Illnesses Data Channel — weekly national COVID-19, flu and RSV activity; page dated September 11, 2026. cdc.gov
2. CDC. NSSP Emergency Department Visits — COVID-19, Flu, RSV, Combined (public dataset; national weekly percentages through the week ending September 5, 2026). data.cdc.gov
3. CDC. RESP-NET Rates and Clinical Data — COVID-NET cumulative hospitalisation rates per 100,000 by season. data.cdc.gov
4. CDC. National Wastewater Data for Respiratory Viruses — activity-level thresholds and the median method; page dated September 11, 2026. cdc.gov
5. CDC. SARS-CoV-2 Variant Proportions (public dataset; most recent period ending August 1, 2026, published August 28, 2026). data.cdc.gov
6. CDC. Respiratory Virus Activity, United States, July 1, 2024–June 30, 2025. MMWR — the illness, death and case-fatality ranges quoted above. cdc.gov
7. World Health Organization. Tracking SARS-CoV-2 variants — current VOC, VOI and VUM listings; content last updated 28 July 2026. who.int
8. World Health Organization. 14.9 million excess deaths associated with the COVID-19 pandemic in 2020 and 2021. who.int
9. U.S. Food and Drug Administration. COVID-19 Vaccines (2026-2027 Formula) for Use in the United States Beginning in Fall 2026. fda.gov
10. Trial evidence cited above is set out with full citations on the vaccines, treatment, symptoms and Long COVID pages — Polack 2020 and Baden 2021 (NEJM), Dunkle 2022 (NEJM), Hammond 2022, Gottlieb 2022, Jayk Bernal 2022 and RECOVERY 2021 (NEJM), Menni 2022 (The Lancet) and Thaweethai 2023 (JAMA).