Table 1 Descriptive statistics of the development and validation cohorts.

From: Machine learning did not beat logistic regression in time series prediction for severe asthma exacerbations

 

Development cohort

Validation cohort

Demographics

Patient, N

165

101

Total daily measurements, N

92,787

40,185

Observational period, median (25–75)

610 (580–640)

417 (376–473)

Age, median (25–75)

38 (28–47)

46.5 (34–56)

Sex (female), N (%)

92 (56%)

62 (61%)

Predictors

Peak expiratory flow, mean (std)

438 (98)

404 (104)

 Missing (%)

477 (0.5%)

1171 (2.9%)

Peak expiratory flow personal besta, mean (std)

467 (100)

437 (103)

Nocturnal awakening, mean % per patient

6.3%

4.7%

 Missing (%)

876 (0.9%)

1665 (4.1%)

Use of \(\upbeta \)2 reliever, mean % per patient

7.2%

8.9%

 Missing (%)

302 (0.3%)

1188 (3.0%)

Outcome

Exacerbations per patient, N (%)

 0 exacerbations

116 (70%)

63 (62%)

 1 exacerbation

25 (15%)

20 (20%)

 2 or more exacerbations

24 (15%)

18 (18%)

Total exacerbations, N (%)

154 (0.2%)

94 (0.2%)

  1. Statistics were calculated for each individual patient over their respective observational periods. Then these statistics were pooled across patients.
  2. aNo % missing is reported for maximum peak expiratory flow as this is a summary statistic calculated per patient over a run-in period of 4 weeks.
Back to article page