Skip to main content

Table 2 Data sets used in this study

From: Studying the potential impact of automated document classification on scheduling a systematic review update

TOPIC

TRAINING SET

TESTING SET

 

INCLUDED

EXCLUDED

TOTAL

INCLUDED

EXCLUDED

TOTAL

STUDIED

ADHD

285

2542

2827

43

245

288

Yes

AEDs

96

1436

1532

39

476

515

Yes

Antiemetics

133

2364

2497

29

334

363

Yes

Antiplatelets

20

0

20

5

0

5

No

AtypicalAntipsychotics

659

3829

4488

67

785

852

Yes

MSDrugs

138

1774

1912

38

601

639

Yes

NasalCorticosteroids

13

82

95

15

62

77

Yes

NSAIDs

79

106

185

1

3

4

No

OveractiveBladder

103

823

926

25

229

254

Yes

ProtonPumpInhibitors

204

2333

2537

50

375

425

Yes

Sedatives

145

1657

1802

26

206

232

Yes

  1. Eleven systematic review topics had both a prior report and an update completed within our data collection window. Included articles are those included in the final systematic review report, while excluded articles are those not included in the report. Drug Effectiveness Review Project (DERP) review inclusion judgments for articles with MEDLINE entry dates prior to the End of the Report Cycle (for the prior report) were used as training data for that topic. DERP review inclusion judgments for articles indexed in MEDLINE during the Pre-Update Period (after the End of the Report Cycle date and prior to the Report Search Begins for the next update) as the testing data for that topic.