A Modest Tobacco Tax and Child-Care Attendance Mitigate Nicotine Exposure in Young Children

• In Event: The New Kids on the Methods Block: Using New Methods in Existing Data to Get New Answers

Abstract

The detrimental effects of tobacco smoke on young children’s health and development are well-known. Findings from health economics suggest that tobacco taxes provide an effective way to reduce adult smoking. It is unclear, however, whether these effects trickle down to impact young children’s nicotine exposure. Further, there is reason to suspect that the magnitude of any tax impact may differ as a function of the time children spend in smoking households, for instance whether they attend child care.
In the present study, we use data from the Family Life Project (n = 1,218)—a prospective longitudinal study of predominantly low-income families in rural contexts—to test the conditional impact of a modest ($.20 pack), North Carolina cigarette-tax increase on children’s levels of salivary cotinine (i.e., nicotine biomarker) across early childhood, along with the potential protective role that child care may play in children’s tobacco exposure.
Specifically, we leverage triple-differences (DDD) models to test the extent to which post-tax changes in children’s cotinine levels were comparatively more pronounced for: (a) those targeted by the tax (i.e., children in a smoking households, compared to non-smoking), (b) those in NC, compared to broader secular cotinine trends evident for the control group (PA); and (c) within NC, for children receiving larger doses of tobacco exposure at home (as inferred by external child care hours). To adjust for endogeniety, we used inverse-probability-of-treatment weights (IPTWs) to balance ~50 potential pre-treatment confounds across state and weekly child-care hours. All models were fitted using weighted mixed linear models.
The results indicated that the tax change led to substantial declines in young children’s cotinine levels. However, as hypothesized, the magnitudes of these effects were conditional on whether the child resided in smoking household and the child’s ‘dose’ of home tobacco exposure. We display the weighted DDD results for children in smoking households in Figure 1. The tax-increase was associated with noteworthy declines in children’s cotinine levels across the subsequent 2.5 years. As expected, the most extensive declines were evident for: (1) those exposed to the tax (NC), who were also (2) exposed to more tobacco smoke by virtue of spending more time at home (i.e., no outside child care). Comparable, control-group children (PA) showed only modest secular declines over this span. Approximately, 2.5 years after the tax change, the differences in these trajectories corresponded to a conditional average treatment effect (CACE) of 3.63 ng/mL. To put this in context, in the US, 3 ng/mL is a common threshold for a light-smoking adult.
Consistent with the idea that child care plays a protective role in children’s tobacco exposure, children attending greater hours of child care showed statistically lower cotinine levels prior to the NC change and were considerably less affected by the tax change. Within-child fixed-effect estimates of child-care exposure showed similar protective relations, irrespective of tax-change timing or state.
Collectively, these results suggest that policies concerning tobacco taxes and/or early childhood education may serve as leverage points for minimizing children’s exposure to tobacco smoke in early childhood.

Authors

• Daniel Berry, University of Minnesota - Twin Cities
  Presenting Author

• Lisa M. Gatzke-Kopp, Pennsylvania State University, University Park
  Non-Presenting Author

• Michael T. Willoughby, RTI International
  Non-Presenting Author

• Clancy Blair, New York University
  Non-Presenting Author