Ecologic Analysis of Lung and Stomach Cancer in Ontario

Abstract

<p> Using maps, correlation and multiple regression, an ecologic analysis was performed to examine the geographic distribution of cancer incidence in ontario with respect to selected ethnic, socio-economic and environmental characteristics for the 10 year period, 1976-1985. Two of the most common causes of cancer deaths, stomach and lung, were studied for each sex separately. The unit of analysis consisted of census divisions. The information used for the cancer were standardized incidence rates from the Ontario cancer Registry. The data for the ecologic variables was obtained from the 1981 Census of Canada. Two of the ecologic variables, education and income (low and median) were used to account for the effects of smoking. </p> <p> Correlation co-efficients were significant for both sites of cancer for males and females for % urban and population density revealing the possibility of a positive relationship with cancer incidence and environmental characteristics. Ethnicity was strongly related to male and female stomach cancer. </p> <p> Significant regression models were obtained for each of the cancer sites using a stepwise procedure with backward elimination. For each of the "best fit" equations, median income and education were included to control for smoking effects. Population density was significant in all equations at the 0.05 level. The percentage urban was significant for all except female stomach cancer. Manufacturing had a negative significant relationship for all cancer sites (male and female). </p> <p> Also included in the study were descriptive statistics and cancer maps to determine the strongest cancer distributions in Ontario. For each site, northern Ontario contained the highest rates. In southern Ontario, urban areas such as Hamilton-Wentworth, and Toronto-York had high rates for all cancers (except Hamilton-Wentworth for male lung cancer). </p>