Skip to main content

Table 1 The three research groups: CBC/DCa, CRP&CBC/DCb, and PCT&CBC/DCc

From: Bacteremia detection from complete blood count and differential leukocyte count with machine learning: complementary and competitive with C-reactive protein and procalcitonin tests

Groups

CBC/DC

CRP&CBC/DC

PCT&CBC/DC

Number of cases

350,775d

308,803

23,912

Training and validation set

290,425

253,009

17,033

 2014

54,729

44,846

2434

 2015

55,982

46,775

2898

 2016

61,438

53,959

2924

 2017

59,843

53,784

4246

 2018

58,433

53,645

4531

Testing set

60,350

55,794

6879

 2019

60,350

55,794

6879e

Age

48.7 ± 30.0

46.8 ± 30.7

53.3 ± 27.6

Male

54.6%

54.3%

58.3%

Female

45.4%

45.7%

41.7%

  1. aComplete blood count/differential leukocyte count
  2. bC-reactive protein and complete blood count/differential leukocyte count
  3. cProcalcitonin and complete blood count/differential leukocyte count
  4. eOf the 6879 cases, only 3070 inpatients’ data were used for prediction
  5. dNumbers in bold represent the total number of cases