Skip to main content
. 2023 Sep;29(9):1789–1797. doi: 10.3201/eid2909.230052

Table. Estimated annual incidence of human brucellosis worldwide determined by using 3 statistical models*.

Region Estimated human cases
2.5% Quantile 25% Quantile Median 75% Quantile 97.5% Quantile
Total Mean (SD)
Weighted average interpolation
World 1,621,468
Asia 1,103,122
Africa 514,001
Americas 3,335
Europe
1,010






Bootstrap resampling
World 1,691,666 (975,292) 679,393 1,080,049 1,416,482 1,906,564 4,651,474
Asia 1,172,573 (959,859) 261,493 566,081 887,126 1,355,607 4,107,355
Africa 513,928 (171,607) 257,863 380,681 487,549 624,155 902,139
Americas 3,343 (214) 3,133 3,181 3,272 3,448 3,912
Europe

1,821 (424)
1,595
1,632
1,688
1,818
3,717
Hierarchical Bayes†
World 2,096,080 (1,754,315) 568,038 1,063,620 1,592,291 2,511,881 6,616,334
Asia 1,622,446 (1,680,985) 246,536 639,906 1,117,309 1,993,573 5,972,342
Africa 468,321 (291,337) 168,919 283,125 393,384 562,957 1,210,226
Americas 3,425 (362) 3,133 3,215 3,319 3,503 4,347
Europe 1,889 (446) 1,593 1,654 1,746 1,944 3,050

*Calculation of uncertainty intervals in the weighted average interpolation method was not performed due to the nature of the model. During bootstrap resampling, uncertainty intervals were calculated using one million resampled risk estimates based on observed reported case count values. †The hierarchical Bayes model intervals were calculated using 1 million posterior samples. Posterior distributions were estimated using a Markov chain Monte Carlo (MCMC) algorithm based on observed reported case count values. For the MCMC algorithm, 50,000 burn-in iterations were performed before the samples were retained.