You are here

CRDCN session on health at CEA meeting

  • Ted McDonald: Spatial and Temporal Analysis of Cancer Incidence and Mortality Canada

    According to Statistics Canada, an average of 21 people are diagnosed with some type of cancer every hour. Although considerable research exists on how cancer incidence varies by socioeconomic status, geographic location and other contextual factors, little work has been done to investigate whether the importance of context varies with duration of exposure and age at exposure. In this paper we use the 1991 Canadian Census Cohort: mortality & cancer follow-up (1991 CCC) to model the effect of socioeconomic and demographic characteristics on cancer outcomes. The 1991 CCC includes the tax file summary which provides an accurate location of each individual every year tax is filed between 1990 and 2008, thus permitting us to identify and differentiate the importance of current and area-level socioeconomic and demographic factors affecting cancer incidence. For example, having information on place of residence over this time period enables us to determine whether observed differences in cancer incidence between rural and urban areas are due more to differences in diagnosis (reflected in the significance of current location of residence) or environmental/contextual factors (reflected in the significance of previous location(s) of residence). Event history analysis is utilized to estimate the determinants of the time until cancer is diagnosed (if at all). Event history analysis models are preferable to traditional regression approaches because they allow for the occurrence of time-varying characteristics (such as age, residence in an area, and duration of residence in that area) while accounting for within-individual correlations in unobserved determinants of cancer incidence.

  • Logan McLeod: Longitudinal Relationship between Participation in Physical Activity and Health

    We address the econometric issues encountered within a Grossman health production framework when empirically testing the effect of participation in leisure time physical activity on health outcomes. Lifestyle choices, more narrowly defined as health-related behaviours, are widely recognized in the epidemiological and economics literature as important non-medical determinants of health. One specific lifestyle choice associated with better health is regular physical activity. While the benefits of regular physical activity are well documented in the clinical and public health literature, they have not been as extensively studied as a determinant of health in the health economics literature. The primary empirical challenges when empirically estimating health production functions are unobservable individual heterogeneity and the endogeneity of health inputs. It is likely unobserved individual heterogeneity affects both the probability of participating in physical activity and the probability of engaging in other healthy behaviours. This, in turn, affects both the probability of participating in physical activity and the probability of engaging in other healthy behaviours. We investigate the longitudinal relationship between participation in physical activity and health outcomes using a random effects probit model. The paper uses data from eight cycles (1994/1995 - 2009/2010) of the Canadian National Population Health Survey (NPHS). The longitudinal sample is representative at the provincial level and contains 17,276 respondents. Preliminary results indicate lagged participation in physical activity reduces the reported incidence of high blood pressure, heart disease, ulcers and arthritis as well as being in fair or poor health.

    Related publication:

  • Emmanuelle Piérard: Child Health and Parental Smoking: Panel Data Evidence from the National Longitudinal Survey of Children and Youth

    Numerous panel data studies have estimated the impacts of maternal smoking on early child health. However, most of this research is focused on a single outcome. We exploit the 1994 - 2001 waves of the National Longitudinal Survey of Children and Youth, which allow us to evaluate panel data based effects of maternal and paternal smoking on a diverse array of child health outcomes as well as the use of health services. Our empirical estimates indicate that smoking by both parents is significantly correlated with a higher likelihood of respiratory illnesses such as asthma and bronchitis and ear, nose, and throat infections. In particular, our results are significant with respect to low income households. In summary, our findings point to the harmful effects of Environmental Tobacco Smoke (ETS).