When an outbreak headline reports a disease’s death rate, two very different numbers are often hiding behind that single figure: the case fatality rate (CFR) and the infection fatality rate (IFR). Both describe lethality, but they answer different questions, and mixing them up is one of the most common ways non-specialists misread epidemiological data. Understanding the gap between them — and why it can be enormous during a fast-moving outbreak — is one of the more useful pieces of interpretive literacy for anyone trying to follow a disease event responsibly.
Two Denominators, One Number of Deaths
Case fatality rate is the proportion of diagnosed, confirmed cases who die from a disease: deaths divided by confirmed cases, expressed as a percentage. It measures lethality among people who were sick enough, or tested enough, to enter the surveillance system.
Infection fatality rate is the proportion of all infected people who die, including everyone who was infected but never tested — the asymptomatic, the mild, the people who never sought care. Its denominator is the true number of infections, not the observed number of cases.
Because the IFR’s denominator is always at least as large as the CFR’s denominator, and the numerator (deaths) is the same, the IFR can never be higher than the CFR. The two numbers converge only when testing is thorough enough that almost every infection becomes a confirmed case — which almost never happens at the start of a novel outbreak.
Why the Gap Opens During an Outbreak
Several forces push CFR and IFR apart, especially in the early weeks of a new pathogen’s spread:
- Undetected infections. A large share of infected people may have mild or no symptoms and are never tested, so they never enter the CFR’s denominator — but they are still counted, once identified, in the IFR.
- Testing capacity and strategy. Early in an outbreak, testing is often limited and reserved for severe cases or high-risk groups, which concentrates the confirmed-case pool with people more likely to die and inflates CFR.
- Reporting lag. Deaths are reported with a delay relative to diagnosis. A CFR calculated mid-outbreak, before all current cases have resolved one way or the other, can be an unstable, moving target rather than a final figure.
- Evolving case definitions. As understanding of a pathogen’s full clinical spectrum improves — including recognizing genuinely asymptomatic infection — later estimates tend to be more complete than first-week figures.
None of this means early CFR figures are wrong to report. It means they should be read as an upper bound on severity among detected cases, not as the disease’s true lethality across everyone it infects.
A Documented Example: COVID-19’s Moving Number
The COVID-19 pandemic is a well-documented case of this divergence playing out in real time. Early data from China suggested a case fatality rate around 3.4%, and the World Health Organization later noted that early country-level CFR estimates varied enormously — from under 0.1% to more than 25% — reflecting differences in testing capacity and case definitions rather than differences in the virus itself.
As seroprevalence studies were conducted — testing population samples for antibodies to estimate how many people had actually been infected, symptomatic or not — infection fatality rate estimates settled much lower. One widely cited meta-analysis put the overall IFR for COVID-19 at roughly 0.68%. A separate review of seroprevalence studies across 51 locations produced a median IFR closer to 0.23%–0.27%, and for people under 70 the median figure was near 0.05%. The headline "case fatality rate" figures reported in the first weeks of the pandemic, in other words, were describing a meaningfully different, more severe population than the one an infection fatality rate eventually captured.
CFR and IFR at a Glance
| Question | Case Fatality Rate (CFR) | Infection Fatality Rate (IFR) |
|---|---|---|
| Denominator | Confirmed, diagnosed cases | All infections, detected and undetected |
| Available early in an outbreak? | Yes, but unstable and testing-dependent | Rarely; needs population studies (e.g. seroprevalence) |
| Tends to run | Higher | Lower |
| Best used for | Clinical severity among detected/treated cases; hospital planning | True population-level lethality; comparing across diseases or waves |
| Main risk of misreading | Overstating how deadly a disease is for the average infected person | Can understate urgency if infections are still being counted |
Reading These Numbers as a Non-Specialist
A few habits make outbreak death-rate figures easier to interpret responsibly. First, check which measure a report is actually citing — "fatality rate" alone is ambiguous and outlets frequently use it loosely. Second, treat any CFR reported in the first weeks of a novel outbreak as provisional; it will typically fall as testing broadens and as still-open cases resolve. Third, remember that IFR is the more meaningful figure for understanding a disease’s true burden across a population, but it is also the slower, harder number to produce — it usually requires the kind of population-based studies that only arrive well after the initial headlines.
For field and NGO teams making resourcing decisions in real time, this distinction has practical weight. A high early CFR from a testing-limited setting may reflect who is being tested, not the underlying severity of the pathogen — a pattern worth flagging before it drives disproportionate resourcing decisions.
Conclusion
Case fatality rate and infection fatality rate are not competing estimates of the same thing; they are answers to two different questions, and an outbreak can legitimately produce very different numbers for each. Recognizing which one is being reported — and how early-outbreak testing gaps inflate the CFR relative to the eventual IFR — is one of the simplest ways to read epidemiological headlines more critically, without needing a background in biostatistics.