Table 2 Hazard ratios derived from the four predictive models fitted: model C.1 considers age as categorical variable; model C.2 considers age as linear continuous variable; model C.3 considers restricted cubic splines on age; model C.4 considers restricted cubic splines on age and adherence as time varying.

From: Using population-based data to evaluate the impact of adherence to endocrine therapy on survival in breast cancer through the web-application BreCanSurvPred

 

C.1 (age: categorical)

C.2 (age: linear)

C.3 (age: splines)

C.4 (age: splines/adherence: time-varying)

Hazard ratio

95% CI

Hazard ratio

95% CI

Hazard ratio

95% CI

Hazard ratio

95% CI

Age

 ≤ 49 years

Ref

 

 

 

 

50–59 years

1.44

0.91–2.88

 

 

 

60–74 years

2.08

1.37–3.18

 

 

 

75–84 years

6.14

4.08–9.25

 

 

 

HER2 status

Negative

Ref

 

Ref

 

Ref

 

Ref

 

Positive

0.95

0.69–1.30

0.97

0.71–1.34

0.97

0.71–1.33

0.98

0.71–1.34

Stage

I

Ref

 

Ref

 

Ref

 

Ref

 

II

2.25

1.54–3.28

2.43

1.67–3.52

2.24

1.51–3.30

2.22

1.53–3.21

III

5.41

3.67–7.92

5.71

3.91–8.34

5.11

3.46–7.51

4.98

3.38–7.32

Adherence rate

 ≤ 80%

Ref

 

Ref

 

Ref

 

 

 > 80%

0.58

0.42–0.82

0.54

0.39–0.76

0.57

0.41–0.79

 

Adherence (time varying)

Yes, follow-up ≥ 5 years

 

 

 

0.57

0.38–0.84

Yes, follow-up < 5 years

 

 

 

0.58

0.33–0.87

Ratio follow-up ≤ 5/ follow-up > 5***

 

 

 

1.02

0.51–2.02

AIC

2759.12

 

2753.25

 

2747.59

 

2759.54

 
  1. CI confidence interval, AIC Akaike Information Criterion, *: Beta coefficient for adherence was considered as time varying coefficient. **: hazard ratio for adherence for patients with follow-up > 5 years; *** Ratio of the hazard rates of the variable adherence = “Yes” for patients with follow-up ≤ 5 years versus > 5 years.