Skip to main content

Table 4 Validity and completeness of outcomes in LHS administrative databases

From: A feasibility study to assess the validity of administrative data sources and self-reported information of breast cancer survivors

Outcome

No. of patients

Reference Standard

Incidence %

Proportion correct %

Kappa (95 % CI)

PPV (95 % CI)

NPV (95 % CI)

Completeness % (proportion)

Cardiovascular disease

119

Medical record

20.2

98.3

0.95 (0.88-1)

95.8 (79.8-99.3)

99 (94.3-99.8)

100 (119/119)

 Ischemic heart disease

119

Medical record

15.1

98.3

0.93 (0.84-1)

94.4 (74.2-99)

99 (94.6-99.8)

100 (119/119)

 Congestive heart failure

119

Medical record

2.5

100

1

100 (43.9-100)

100 (96.8-100)

100 (119/119)

 Conduction disorders and cardiac dysrhythmias

119

Medical record

8.4

99.2

0.94 (0.83-1)

100 (70.1-100)

99.1 (95–99.8)

100 (119/119)

 Stroke

119

Medical record

7.6

97.5

0.79 (0.55-1)

100 (61–100)

97.4 (92.5-99.1)

100 (119/119)

Fractures

119

Medical record

6.7

98.3

0.85 (0.64-1)

100 (61–100)

98.2 (93.8-99.5)

100 (119/119)

Osteoporosis

119

Medical record

25.2

98.3

0.96 (0.90-1)

93.8 (79.9-98.3)

100 (95.8-100)

100 (119/119)

Diabetes

119

Medical record

15.1

99.2

0.97 (0.91-1)

94.8 (75.4-99.1)

100 (96.3-100)

100 (119/119)

Chronic pain

67

SF-36

68.7a

94

0.85 (0.72-0.99)

92 (81.2-96.9)

100 (81.6-100)

100 (67/67)

  1. Abbreviations: PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; SF-36, 36-item short form
  2. a For chronic pain, the effect measure is prevalence