Skip to main content

Table 1 Data characteristics

From: Predicting decompression surgery by applying multimodal deep learning to patients’ structured and unstructured health data

Characteristics

Early surgery

Late surgery

N

8387

8620

 Negative

8189 (97.6%)

8189 (95.0%)

 Positive

198 (2.4%)

431 (5.0%)

Average days between LIRE enrollment and decompression surgery

34.3

168

Age

57

57.2

Gender

  

 Female

4713 (56.2%)

4845 (56.2%)

Race

  

 White

5317 (63.4%)

5502 (63.8%)

 Black

991 (11.8%)

1007 (11.7%)

 Unknown

990 (11.8%)

1000 (11.6%)

 Asian

928 (11.1%)

948 (11.0%)

 Pacific Islander

50 (0.6%)

51 (0.6%)

 Other

27 (0.3%)

26 (0.3%)

 Multiracial

17 (0.2%)

19 (0.2%)

Ethnicity

  

 Not available

5945 (70.9%)

6129 (71.1%)

 Not Hispanic

1233 (14.7%)

1263 (14.7%)

 Hispanic

1209 (14.4%)

1228 (14.2%)

Image type

  

 MRI

5810 (69.3%)

5980 (69.4%)

 X-ray

2517 (30.0%)

2576 (29.9%)

 CT

60 (0.7%)

64 (0.7%)

System

  

 Kaiser Permanente

7071 (84.3%)

7274 (84.4%)

 Henry Ford

654 (7.8%)

657 (7.6%)

 Group Health

486 (5.8%)

517 (6.0%)

 Mayo Clinic

176 (2.1%)

172 (2.0%)