Skip to main content

Table 2 Data volume for the three clinical trials: number of patients treated by the department during the recruitment period, number of patients included, number of attributes documented for all these patients and the fraction of attributes with a documented value for each patient

From: Evaluating predictive modeling algorithms to assess patient eligibility for clinical trials from routine data

Trial

Patients treated

Patients included

Number of attributes [n]/Valued cells [%] (target variable, age, gender + no. of codes)

Patients without any

 

[n]

[n]

[%]

No aggregation

Category-level aggregation

Block-level aggregation

Diagnosis

Procedure

A

511

361

70.6

3,689/0.7

1,031/1.9

280/5.6

3

6

B

8,170

320

3.9

11,773/0.3

1,627/1.6

305/6.6

6

215

C

5,573

87

1.6

10,336/0.4

1,494/2.1

297/8.1

8

0

  1. Each different diagnosis and procedure code was treated as an independent patient attribute.