Table 1 Patient demographics for prediction of 1-year expenditure.

From: Prediction of future healthcare expenses of patients from chest radiographs using deep learning: a pilot study

Characteristics

Count (%)

# below ave cost (%)

# above ave cost (%)

Gender (p = 0.112)*

Female

9616 (49.3)

5091 (49.8)

4525 (48.7)

Male

9908 (50.7)

5133 (50.2)

4775 (51.3)

Age (p < 0.0001)*

18–37

4055 (20.8)

2949 (28.9)

1106 (11.9)

38–57

5423 (37.8)

3081 (30.1)

2342 (25.2)

58–76

6212 (31.8)

2634 (25.8)

3578 (38.5)

77–96

3639 (18.6)

1485 (14.5)

2154 (23.1)

97–116

195 (1.0)

75 (0.7)

120 (1.3)

Region (p < 0.0001)*

Bay area

17,153 (87.9)

9093 (88.9)

8060 (86.7)

Not bay area

2367 (12.1)

1129 (11.1)

1238 (13.3)

Not assigned

4 (0.0)

2 (0.0)

2 (0.0)

Race (p < 0.0001)*

American Indian or Alaska Native

67 (0.3)

30 (0.3)

37 (0.4)

Asian

4196 (21.5)

2072 (20.3)

2124 (22.8)

Black or African American

2326 (11.9)

1219 (11.9)

1108 (11.9)

Native Hawaiian or other Pacific Islander

454 (2.3)

285 (2.8)

169 (1.8)

Other

2923 (15.0)

1569 (15.3)

1354 (14.6)

Unknown/declined

625 (3.2)

411 (4.0)

214 (2.3)

White or Caucasian

8933 (45.8)

4638 (45.4)

4295 (46.2)

  1. *p-value is comparing the relationship between cost and the corresponding variables using chi-squared test for independence for categorical variables, two-tails at significance level 0.05.