Research Article

Estimating and Interpreting Effects from Nonlinear Exposure-Response Curves in Occupational Cohorts Using Truncated Power Basis Expansions and Penalized Splines

Table 2

Estimated hazard ratios and 95% pointwise confidence intervals from separate Cox proportional hazard models using the carpal tunnel syndrome and strain index exposure data.

Exposure value
x
LinearLogarithmicLinear spline with knot at 13.5Penalized spline function with

0.81.01
(0.99, 1.03)
1.21
(0.92, 1.59)
1.10
(1.01, 1.20)
1.04
(0.97, 1.11)
6.01.10
(0.94, 1.29)
1.88
(0.77, 4.62)
2.09
(1.10, 4.00)
1.35
(0.81, 2.27)
9.01.15
(0.91, 1.46)
2.11
(0.73, 6.11)
3.03
(1.15, 7.99)
1.57
(0.78, 3.16)
13.51.24
(0.86, 1.77)
2.38
(0.69, 8.19)
5.27
(1.23, 22.57)
1.89
(0.78, 4.60)
18.01.33
(0.82, 2.14)
2.60
(0.67, 10.13)
4.85
(1.23, 19.04)
2.12
(0.79, 5.75)
20.31.38
(0.80, 2.36)
2.70
(0.66, 11.08)
4.65
(1.23, 17.65)
2.18
(0.78, 6.11)
54.02.33
(0.55, 9.84)
3.68
(0.58, 23.35)
2.51
(0.44, 14.42)
2.32
(0.38, 14.15)