The Claim
The claim is that the official U.S. COVID-19 death toll was substantially inflated because people who merely tested positive for SARS-CoV-2 were routinely counted as having died from COVID-19 even when the infection did not cause or contribute to death.
The important word is substantially.
Individual certification errors are inevitable in a mortality system of this scale. The question is whether incidental infections were misclassified often enough to create a large systematic upward bias in the national toll.
Where It Comes From
During the pandemic, public discussion often contrasted dying “with COVID” and dying “from COVID.”
The distinction is medically legitimate. A person can have SARS-CoV-2 infection and die from an unrelated cause.
Confusion increased because U.S. guidance allowed COVID-19 to be certified as “probable” or “presumed” when laboratory confirmation was unavailable but the certifier judged it a likely cause.
Statements about deaths involving multiple comorbidities were also frequently interpreted as evidence that COVID-19 had not caused those deaths.
Why It Sounds Convincing
Death certification is imperfect.
People with severe chronic disease can die while infected with SARS-CoV-2, and deciding how much each condition contributed can require medical judgment.
Cause-of-death forms can contain poor causal sequences or incomplete information.
Those are real weaknesses.
The viral claim becomes much stronger when it asserts that U.S. surveillance simply counted positive tests as COVID deaths.
That is not how the formal system was defined.
The Short Answer
The evidence does not support substantial U.S. overcounting from routine “with COVID = from COVID” classification.
Guidance from the U.S. Centers for Disease Control and Prevention (CDC) and its National Center for Health Statistics (NCHS) explicitly says that a positive SARS-CoV-2 test alone is not enough to classify a death as due to COVID-19.
COVID-19 is counted when the certifier reports it as an underlying or contributing cause of death.
Direct comparison work in Minnesota found high agreement between death certificates and case-based investigations and found undercounting more often than overcounting.
National and county-level excess-mortality studies also generally identify substantial mortality that was not officially attributed to COVID-19.
Certification quality was imperfect, and individual over-attribution occurred.
That is different from evidence that the national death toll was substantially inflated.
How It Would Have to Work
For the proposed substantial overcount from incidental positive tests to be supported, the following should be observed:
- certifiers would need to list COVID-19 as causal in many deaths where infection was incidental;
- those errors would need to be systematically more common than missed COVID deaths;
- the net error would need to be large enough to materially inflate the national tally.
The claim is therefore about the direction and magnitude of population-level misclassification, not whether any incorrect certificate existed.
What Actually Happens
A death certificate records causal judgment
The cause-of-death section is not a list of everything found in the patient.
Certifiers are instructed to report conditions that caused or contributed to death.
If SARS-CoV-2 was present but did not cause or contribute to death, it should not be reported as a COVID-19 cause of death.
“Contributing cause” is not the same as incidental infection
A contributing condition may not be the initiating disease but can still materially contribute to death.
For example, COVID-19 can worsen cardiopulmonary disease or precipitate complications.
Counting such a death as involving COVID-19 is not automatically misclassification.
Probable or presumed COVID was allowed
When testing was unavailable or inconclusive, a certifier could report probable or presumed COVID-19 when the clinical circumstances supported that judgment.
That increases uncertainty in individual cases.
It does not mean every untested death was labeled COVID-19.
Certificate quality is imperfect
A 2026 U.S. study found that only around half of COVID-related death certificates in 2020–2021 contained causal sequences classified as acceptable by the study’s criteria, with quality declining later.
That is a serious documentation-quality issue.
A poorly written causal chain, however, is not automatically evidence that COVID-19 was incidental.
What the Evidence Shows
CDC’s formal guidance directly contradicts the common claim that every positive test was counted as a COVID death.
A 2024 Minnesota comparison supplied counts in both directions. Against the state’s standardized mortality definition, 13,108 deaths were identified by both methods; another 483 met the case-based COVID-19 death definition but did not mention COVID-19 on the death certificate, whereas 413 certificates mentioning COVID-19 were ruled out by the state definition. This corresponds to about 96% sensitivity for certificate mentions among the 13,591 COVID-19 deaths confirmed by that definition. It is not a population-wide 96% accuracy estimate, and a state-specific validation is not a direct estimate of national net error.
The study’s comparison set contained 14,004 deaths, not every death in Minnesota. The separate evidence on nationwide excess deaths is therefore important rather than interchangeable with this validation sample.
Excess-mortality analyses provide an independent consistency check.
A 2024 U.S. county-level analysis estimated large numbers of excess natural-cause deaths that were not officially reported as COVID deaths.
Excess deaths are not synonymous with direct COVID deaths; disrupted healthcare, overdose, other infections and many indirect mechanisms matter.
But a large pool of unassigned excess mortality makes a simple nationwide story of massive net overcounting difficult to sustain.
The Strongest Evidence for the Claim
The strongest case is not “CDC counted every positive test.” That is incorrect.
The strongest serious case is:
- cause-of-death certification contains meaningful error;
- probable/presumed COVID could be used without laboratory confirmation;
- some certificates contain weak or rejected causal sequences;
- individual incidental infections could therefore have been wrongly attributed.
That supports skepticism about perfect accuracy.
It does not establish substantial net overcount.
The Strongest Evidence Against It
The strongest evidence against a substantial overcount includes:
- formal CDC/NCHS rules requiring causal or contributing involvement;
- direct state-level comparisons showing high agreement and more undercount than overcount;
- excess-mortality estimates exceeding official COVID mortality in many places and periods.
Together, these evidence streams point away from substantial systematic upward bias.
Common Viral Arguments
“Anyone who died within 28 days of a positive test was counted as a U.S. COVID death.”
That is not the NCHS death-certificate definition.
Other surveillance systems or countries used different operational metrics at times. They should not be substituted for U.S. certified mortality counts.
“Only 6% died of COVID alone.”
Death certificates commonly contain multiple conditions. Pneumonia, respiratory failure, diabetes or heart disease can coexist in a causal chain.
Multiple causes do not mean COVID-19 was incidental.
“Probable COVID means doctors could count anyone.”
Certifiers could use probable/presumed when clinical evidence supported it. That introduces judgment, not unlimited classification.
“Excess deaths prove every official COVID death was correct.”
No. Excess mortality is a population consistency check, not validation of every certificate.
The Reasoning Error
A common error is a category error between:
- infection present;
- disease contributing;
- underlying cause;
- and unrelated incidental infection.
Another is base-rate neglect when isolated certification mistakes are treated as evidence about the accuracy of a national total.
What Would Change Our Conclusion?
Large representative medical-record or autopsy audits showing that a substantial fraction of U.S. U07.1 deaths involved incidental infection would change the conclusion.
Evidence that over-attribution consistently exceeded under-attribution by a large margin would also matter.
Conversely, broader chart-review studies reproducing Minnesota’s findings would strengthen the current direction.
Evidence Status
Status: Contradicted
Evidence quality: Moderate
Certification was imperfect, but the evidence reviewed does not support the strong claim that the official U.S. death toll was substantially inflated because incidental infections were routinely counted as causal COVID deaths.
Evidence & Sources
Central sources:
- CDC/NCHS COVID-19 mortality FAQ and certification guidance.
- CDC MMWR analysis of death-certificate ICD-10 coding.
- Minnesota COVID-19 death-determination comparison, Emerging Infectious Diseases, 2024.
- U.S. county excess natural-cause mortality analysis, PNAS, 2024.
- U.S. death-certificate quality study, American Journal of Epidemiology, 2026.
Corrections & Updates
October 10, 2026 — evidence clarification. Specified Minnesota COVID death-certificate validation counts (13,108 concordant, 483 missed and 413 over-attributed by study definition); local error does not determine national net overcount. Contradicted / Moderate unchanged.
Last Reviewed
October 10, 2026.