Paperonium

Malaria · Lancet Infect Dis · 2014

The summary says 58 of 113 sites (53%) had chloroquine-resistant malaria. 58 of 113 is 51%.

53%The summary: too high51%Its own count, divided: 58 of 1130%255075100
53%The summarytoo high51%Its owncount, divided58 of 1130%255075100

Of the 113 sites, the share where P. vivax resisted chloroquine

The paper says

A review of 129 malaria trials says in its summary (the short paragraph at the top) that one malaria parasite, P. vivax, was resistant to the drug chloroquine at “58 (53%) of 113” sites: places where trials ran and the results could be judged.

Its own data show

Do the division: 58 ÷ 113 is 51%, not 53%. The results section also counts differently: of 177 sites, it sets aside 65 that could not be judged cleanly, and finds 57 of the remaining 112 resistant, also 51%. Nowhere else does the paper give 58 of 113.

A percentage is just a count divided by its total, so anyone with a calculator can check one. The gap is small, and the main finding stands: resistance at about half the sites. But the summary is the part most likely to be quoted, and its 53% follows neither from its own count nor from the paper’s results.

“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 →

Global extent of chloroquine-resistant Plasmodium vivax: a systematic review and meta-analysis · Lancet Infect Dis (2014) · doi:10.1016/S1473-3099(14)70855-2