Smoking trajectories among Koreans in Seoul and California: exemplifying a common error in age parameterization

Abstract

Immigration to a nation with a stronger anti-smoking environment has been hypothesized to make smoking less common. However, little is known about how environments influence risk of smoking across the lifecourse. Research suggested a linear decline in smoking over the lifecourse but these associations, in fact, might not be linear. This study assessed the possible nonlinear associations between age and smoking and examined how these associations differed by environment through comparing Koreans in Seoul, South Korea and Korean Americans in California, United States. Data were drawn from population based telephone surveys of Korean adults in Seoul (N=500) and California (N=2,830) from 2001-2002. Locally weighted scatterplot smoothing (lowess) was used to approximate the association between age and smoking with multivariable spline logistic regressions, including adjustment for confounds used to draw population inferences. Smoking differed across the lifecourse between Korean and Korean American men. The association between age and smoking peaked around 35 years among Korean and Korean American men. From 18 to 35 the probability of smoking was 57% higher (95%CI, 40 to 71) among Korean men versus 8% (95%CI, 3 to 19) higher among Korean American men. A similar difference in age after 35, from 40 to 57 years of age, was associated with a 2% (95%CI, 0 to 10) and 20% (95%CI, 16 to 25) lower probability of smoking among Korean and Korean American men. A nonlinear pattern was also observed among Korean American women. Social role transitions provide plausible explanations for the decline in smoking after 35. Investigators should be mindful of nonlinearities in age when attempting to understand tobacco use.

Figure 1. Mean Smoking Trajectories among Koreans in Seoul, South Korea and Korean Americans in California, United States

(a) and (b) show bivariable lowess plots (bandwidth=0.80) between age and current smoking status. The later presents estimates produced by simulating the probabilities using 1000 randomly drawn sets of estimates from the coefficient covariance matrix of the spline-logistic regression model with a single knot at 35 years of age with covariates held at their means. Korean Americans are indicated by a dashed line and Koreans are indicated by a solid line.

Figure 2. Linear regression yields invalid estimates, a case with men

The background replicates the mean trends among men shown in Figure 1A with emboldened trends shown from estimates of the same model where the spline terms for age were replaced by a single continuous age indicator. Korean Americans are indicated by a dashed line and Koreans are indicated by a solid line.