Skip to main content

Table 2 Sensitivity (S) and Positive Predictive Values (PPV). (by Algorithm and Region)

From: Is hospital discharge administrative data an appropriate source of information for cancer registries purposes? Some insights from four Spanish registries

 

Algorithm 1

Dx1

Algorithm 2

Dx1 to Dx5

Algorithm 3

Dx1&Pr1-Pr5

Algorithm 4

Dx1-Dx5&Pr1-Pr5

Granada

    

Cases identified in Registry

231

231

231

231

Cases identified in HDAD

55

113

33

33

Sensitivity (CI95%)

12.9 (8.9-18.0)

14.5 (10.3-19.6)

8.7 (5.4-13.1)

8.7(5.4-13.1)

PPV (CI95%)

54.5 (40.6-68.0)

39.8 (30.7-49.5)

60.6 (42.1-77.1)

60.6 (42.1- 77.1)

Basque Country

    

Cases identified in Registry

1150

1150

1150

1150

Cases identified in HDAD

372

632

245

250

Sensitivity (CI95%)

17.8 (15.6-20.1)

24.3 (21.8-26.8)

13.2 (11.3-15.3)

13.5 (11.5-15.6)

PPV (CI95%)

55.1 (49.9-60.2)

44.1 (40.2-48.1)

62.0 (55.6-68.1)

62.0 (55.6-68.0)

Murcia

    

Cases identified in Registry

411

411

411

411

Cases identified in HDAD

118

218

56

59

Sensitivity (CI95%)

14.4(11.1-18.1)

23.6 (19.5-28.0)

7.3 (4.9-10.2)

8.0 (5.5-11.0)

PPV (CI95%)

50.0 (40.6-59.3)

44.5 (37.7-51.3)

53.6 (39.7-67.0)

55.9 (42.4-68.8)

Zaragoza

    

Cases identified in Registry

494

494

494

494

Cases identified in HDAD

217

378

169

171

Sensitivity (CI95%)

30.4 (26.3-34.4)

45.7 (41.4-50.1)

25.3 (21.5-29.3)

25.7 (22.0-29.8)

PPV (CI95%)

69.1 (64.1-74.2)

59.8 (55.4-64.8)

74.0 (66.6-80.4)

74.3 (67.0-80.6)

  1. S: sensitivity, PPV: Positive Predictive Value; CI95%: Confidence Interval assuming a type I error of 5%