Skip to main content
. 2018 Nov 27;6(4):e49. doi: 10.2196/medinform.9907

Table 3.

Patient characteristics and prevalence of the dependent variable in the model development and validation cohorts.

Variable Model development cohort, (n=290,434) Validation cohort 1 (n=289,426) Validation cohort 2 (n=1815)
Statistic P value Statistic P value
Prediction date January 1, 2014 February 1, 2014 N/Aa February 1, 2014 N/A
Age (years), mean (SD) 81.3 (11.5) 81.2 (11.5) .20 79.8 (11.5) <.001
Female patients, n (%) 234,817 (80.85) 233,692 (80.74) .66 1461 (80.50) .91
PERSb self-reported medical conditions, n (%)

0 59,769 (20.58) 61,685 (21.31) <.001 125 (6.89) <.001

1-2 72,067 (24.81) 71,094 (24.56) .02 656 (36.14) <.001

3-4 77,739 (26.77) 76,384 (26.39) .27 519 (28.60) .31

≥5 80,859 (27.84) 80,263 (27.73) .25 515 (28.37) .84
Patients with 30-day emergency hospital transport, n (%) 6686 (2.30) 6411 (2.22) .08 40 (2.20) .99
Patients with ≥ 1 emergency encounters at 3 time points after the prediction date, n (%)

30 days N/A N/A N/A 64 (3.53) N/A

180 days N/A N/A N/A 321 (17.69) N/A

365 days N/A N/A N/A 509 (28.04) N/A

aN/A: not available.

bPERS: personal emergency response system.