Table 3 NLP-identified reasons for statin nonuse in patients with ASCVD, stratified by type of ASCVD and race/ethnicity.

From: Using deep learning-based natural language processing to identify reasons for statin nonuse in patients with atherosclerotic cardiovascular disease

Cohort (N = 3929)

Reason for nonuse

p value

Side effect

   

Muscle

Other

Nonspecific

Perceived lipid control

Patient preference

Total number

385

404

1011

274

321

Stratified by type of ASCVD (N, % by column)

 Coronary artery disease

224 (58.2)

233 (57.7)

472 (46.7)

123 (44.9)

175 (54.5)

<0.001

 Peripheral artery disease

35 (9.1)

38 (9.4)

129 (12.8)

29 (10.6)

47 (14.6)

 Cerebrovascular disease

87 (22.6)

85 (21.0)

310 (30.7)

95 (34.7)

74 (23.1)

 Polyvascular disease

39 (10.1)

48 (11.9)

100 (9.9)

29 (10.6)

47 (14.6)

Stratified by race/ethnicity (N, % by column)

 Non-Hispanic White

244 (63.4)

273 (67.6)

554 (54.8)

132 (48.2)

214 (66.7)

<0.001

 Non-Hispanic Black

13 (3.4)

25 (6.2)

62 (6.1)

19 (6.9)

13 (4.0)

 Hispanic

30 (7.8)

15 (3.7)

96 (9.5)

34 (12.4)

19 (5.9)

 Non-Hispanic Asian

40 (10.4)

46 (11.4)

96 (9.5)

34 (12.4)

19 (5.9)

  1. ASCVD atherosclerotic cardiovascular disease, NLP natural language processing.