Paperonium

Mpox testing · Lancet Microbe · 2025

Mpox blood test: paper says it catches 98%; its own counts say 95%. Two figures swapped.

The summaryIts own counts98%Had mpox or itsvaccine (63):share caught95%Had mpox or itsvaccine (63):share caught60 of 6395%Had neither(435): sharecorrectly told “no”98%Had neither(435): sharecorrectly told “no”428 of 435
The summaryIts own counts98%Had mpox or itsvaccine (63):share caught95%Had mpox or itsvaccine (63):share caught60 of 6395%Had neither(435): sharecorrectlytold “no”98%Had neither(435): sharecorrectlytold “no”428 of 435

Share of each group the test gets right: paper’s figures versus its own counts

The paper says

MpoxPlex is a blood test for antibodies, proteins the body makes after mpox infection or vaccination. The paper’s summary says it catches 98% of samples from people who had either (98.39% in its table) and correctly says “no” to 95% from people who had neither.

Its own data show

Only 63 samples came from people who had either, and no count out of 63 makes 98.39% (62 of 63 is 98.41%). 98.39% is 428 of the other 435 samples, correctly told “no”. The 95% is 60 of 63 caught. The authors’ data, posted online, confirm both. Two tables repeat this swap in every row.

A test can be wrong two ways: missing someone who had mpox or its vaccine, or a false alarm for someone who had neither. The paper implies under 2% missed and about 5% false alarms. Its counts show the reverse: 3 of 63 missed (5%), 7 of 435 false alarms (under 2%). Checking one person, it does worse than reported.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Assessment of MpoxPlex, a high-throughput and multiplexed immunoassay: a diagnostic accuracy study · Lancet Microbe (2025) · doi:10.1016/j.lanmic.2024.100987