Skip to main content
. 2019 Jul 2;26(8-9):787–795. doi: 10.1093/jamia/ocz093

Table 2.

Data set statistics

Construct Training set a
Validation set b
Test set c
Sentence # (%) Patient # (%) Sentence # (%) Patient # (%) Sentence # (%) Patient # (%)
BC 40 (0.05) 12 (14.12) 46 (0.15) 4 (8.0) 8 (0.02) 3 (6.0)
DE 222 (0.3) 15 (17.65) 85 (0.27) 9 (18.0) 127 (0.28) 10 (20.0)
FL 379 (0.51) 37 (43.53) 79 (0.25) 21 (42.0) 189 (0.42) 23 (46.0)
ML 84 (0.11) 9 (10.59) 6 (0.02) 4 (8.0) 33 (0.07) 6 (12.0)
PU 512 (0.69) 53 (62.35) 348 (1.12) 30 (60.0) 425 (0.94) 30 (60.0)
SS 222 (0.3) 16 (18.82) 21 (0.07) 4 (8.0) 92 (0.2) 7 (14.0)
UC 92 (0.12) 16 (18.82) 38 (0.12) 6 (12.0) 119 (0.26) 13 (26.0)
VI 590 (0.79) 56 (65.88) 355 (1.14) 26 (52.0) 383 (0.85) 34 (68.0)
WD 99 (0.13) 21 (24.71) 87 (0.28) 14 (28.0) 237 (0.52) 19 (38.0)
WL 42 (0.06) 8 (9.41) 33 (0.11) 5 (10.0) 161 (0.36) 12 (24.0)
No construct 72 391 (96.94) 30 028 (96.47) 43 374 (96.07)

Abbreviations: BC, absence of fecal control; DE, dementia; FL, falls; ML, malnutrition; PU, pressure ulcers; SS, lack of social support; UC, severe urinary control issues; VI, visual impairment; WD, walking difficulty; WL, weight loss.

a

85 patients and 74 673 sentences.

b

50 patients and 31 126 sentences.

c

50 patients and 45 148 sentences.