Skip to main content

Table 2 A summary of the shared task dataset’s risk factor tags

From: Adapting transformer-based language models for heart disease detection and risk factors extraction

Risk factor tags

Indicator

Time

Before DCT

During DCT

After DCT

(a) Tag: CAD Indicator

Mention

Event

Symptom

260

224

54

261

20

24

259

2

3

(b) Tag: Diabetes indicator

Mention

Glucose

A1C

518

16

89

524

9

21

518

0

0

(c) Tag: Hyperlipdemia indicator

High LDL

High chol.

Mention

23

5

340

10

1

340

0

0

340

(d) Tag: Hypertension indicator

High bp

Mention

41

523

322

521

0

519

(e) Tag: Obese indicator

DMI

Mention

3

133

15

147

2

133

(f) Tag: Medication type (type1)

Thienopyridine

Statin

Thiazolidinedione

Aspirin

Metformin

Insulin

Fibrate

Ezetimibe

Diuretic

Anti diabetes

ARB

Sulfonylureas

DPP4 inhibitors

ACE inhibitor

97

436

43

424

187

204

22

12

113

1

98

159

1

326

98

427

41

6

176

218

20

12

99

1

93

155

0

318

97

438

40

424

181

212

22

12

106

1

97

157

0

323

(g) Tag: Family_history indicator

Not present

Present

NA

NA

768

22

(g) Tag: Smoker status

Current

Ever

Never

Past

Unknown

NA

NA

58

9

184

149

371

  1. Annotation-level training and testing set sizes, as well as indicators for each heart risk factor