Table 1 Study data
From: Predicting phenotypes of asthma and eczema with machine learning
variable | median (IQR) | #missing (%) | ||||
|---|---|---|---|---|---|---|
age (years) | 42.6 (39.7-45.7) | 0 (0%) | ||||
year of birth | 1964 (1961-1967) | 0 (0%) | ||||
body mass index (BMI) | 26 (23.6-29.1) | 2 (0.4%) | ||||
whole body impedance | 613.5 (550-685) | 34 (6.1%) | ||||
fat % | 29.5 (23.7-36) | 34 (6.1%) | ||||
exhaled nitric oxide (eNO), ppb (loge scale) | 2.8 (2.4-3.3) | 94 (17.0%) | ||||
specific airway resistance (sRaw), kPa/s (loge scale) | -0.1 (-0.3-0.1) | 11 (2.0%) | ||||
peak expiratory flow (PEF) % predicted | 113.1 (102.1-124.6) | 9 (1.6%) | ||||
forced vital capacity (FVC) % predicted | 114.5 (105.6-123.1) | 11 (2.0%) | ||||
forced expiratory volume in 1 second (FEV^ % predicted | 106 (98.6-115.5) | 10 (1.8%) | ||||
forced expiratory flow (FEF25-75) % predicted | 80 (66-96.3) | 11 (2.0%) | ||||
total lung capacity (TLC) | 108.5 (101-116.9) | 11 (2.0%) | ||||
residual volume (RV) | 113 (98.6-127.8) | 11 (2.0%) | ||||
FEVj/FVC ratio | 0.8 (0.8-0.8) | 11 (2.0%) | ||||
provocative concentration of methacholine needed to produce a 20% fall in FEVj (PC20), of those completing the test | 5.3 (1.2-9.0) | 43 (7.8%) | ||||
methacholine dose-response slope (MDRS), transformed as 100/(MdRS+10) | 5.7 (4.2-7.5) | 43 (7.8%) | ||||
variable | N (%) | #missing (%) | ||||
Gender | male | 234 (42.2%) | 0 (0%) | |||
smoking status | never | 341 (61.6%) | 0 (0%) | |||
| Â | ex-smoker | 144 (26%) | 0 (0%) | |||
| Â | current | 69 (12.5%) | 0 (0%) | |||
cat/dog ownership | 186 (33.6%) | 1 (0.2%) | ||||
allergen sensitisation by skin prick test (SPT) | dust mite (mean wheal diameter >3 mm) | 162 (29.3%) | 1 (0.2%) | |||
| Â | cat (mean wheal diameter >3 mm) | 106 (19.2%) | 1 (0.2%) | |||
| Â | dog (mean wheal diameter >3 mm) | 48 (8.6%) | 1 (0.2%) | |||
| Â | tree (mean wheal diameter >3 mm) | 76 (13.8%) | 1 (0.2%) | |||
| Â | grass (mean wheal diameter >3 mm) | 129 (23.4%) | 1 (0.2%) | |||
| Â | mould (mean wheal diameter >3 mm) | 16 (2.9%) | 1 (0.2%) | |||
| Â | peanut (mean wheal diameter >3 mm) | 9 (1.7%) | 1 (0.2%) | |||
bird ownership | 13 (2.4%) | 1 (0.2%) | ||||
medications in the past three months | short-acting beta agonists (SABA) | 34 (6.1%) | 1 (0.2%) | |||
| Â | inhaled corticosteroids (ICS) or ICS/long-acting beta agonists (LABA) | 37 (6.7%) | 1 (0.2%) | |||
illness or problem caused by eating a particular food or foods, ever | 97 (17.5%) | 1 (0.2%) | ||||
accident at home, work or elsewhere exposing to high levels of vapours, gas or dust | 22 (4%) | 2 (0.4%) | ||||
carpets in the house | 292 (52.8%) | 1 (0.2%) | ||||
gas stove in the house | 432 (78.1%) | 1 (0.2%) | ||||
electric stove in the house | 231 (41.8%) | 1 (0.2%) | ||||
job causing wheezing problems | 33 (6%) | 3 (0.5%) | ||||
proportion of subjects not completing PC20 | 423 (82.8%) | 43 (7.8%) | ||||
proportion of subjects with current asthma (CA) | 93 (16.7%) | 0 (0%) | ||||
proportion of subjects with level-2 asthma (A2) | 70 (12.7%) | 0 (0%) | ||||
proportion of subjects with level-3 asthma (A3) | 24 (4.3%) | 0 (0%) | ||||
proportion of subjects with current wheeze (CW) | 68 (12.3%) | 0 (0%) | ||||
proportion of subjects with self-diagnosed eczema (SDE) | 146 (26.3%) | 0 (0%) | ||||
proportion of subjects with doctor's diagnosed eczema (DDE) | 120 (21.7%) | 0 (0%) | ||||
cross-tabulation of clinical outcomes (% of agreement) | ||||||
| Â | SDE | DDE | CA | A2 | CW | A3 |
SDE | Â | 95.47% | 74.82% | 74.82% | 73.37% | 74.09% |
DDE | 95.47% | Â | 77.17% | 77.90% | 76.09% | 78.26% |
CA | 74.82% | 77.17% | Â | 95.29% | 89.49% | 87.68% |
A2 | 74.82% | 77.90% | 95.29% | Â | 92.39% | 91.67% |
CW | 73.37% | 76.09% | 89.49% | 92.39% | Â | 90.94% |
A3 | 74.09% | 78.26% | 87.68% | 91.67% | 90.94% | Â |