OBJECTIVES: To identify geographic variations in colorectal cancer by stage at diagnosis in California using a descriptive analysis coupled with a spatial analysis and to discuss methodological considerations concerning the spatial statistical method. METHODS: We analyzed 59,076 colorectal cancer cases diagnosed in California from 1996 to 2000 by logistic regression and by a spatial scan statistic to identify areas with a higher and lower relative risk of late-stage colorectal cancer. RESULTS: In California, 57% of overall cases of colorectal cancer were diagnosed at a late stage. Californians diagnosed with late-stage colorectal cancer were more likely to be Hispanic and living in areas of lower socioeconomic status. The spatial scan identified two areas where the observed number of late-stage cancer was different than the number expected from the distribution in the rest of the state. CONCLUSIONS: Spatial scan analyses can complement descriptive statistics, but results must be interpreted with consideration of factors that affect the ability to detect meaningful differences such as the number of events observed, accuracy in geocoding rural versus urban addresses, and the difficulty of adjusting for covariates.
OBJECTIVES: To identify geographic variations in colorectal cancer by stage at diagnosis in California using a descriptive analysis coupled with a spatial analysis and to discuss methodological considerations concerning the spatial statistical method. METHODS: We analyzed 59,076 colorectal cancer cases diagnosed in California from 1996 to 2000 by logistic regression and by a spatial scan statistic to identify areas with a higher and lower relative risk of late-stage colorectal cancer. RESULTS: In California, 57% of overall cases of colorectal cancer were diagnosed at a late stage. Californians diagnosed with late-stage colorectal cancer were more likely to be Hispanic and living in areas of lower socioeconomic status. The spatial scan identified two areas where the observed number of late-stage cancer was different than the number expected from the distribution in the rest of the state. CONCLUSIONS: Spatial scan analyses can complement descriptive statistics, but results must be interpreted with consideration of factors that affect the ability to detect meaningful differences such as the number of events observed, accuracy in geocoding rural versus urban addresses, and the difficulty of adjusting for covariates.
Authors: Alyna T Chien; Kristen Wroblewski; Cheryl Damberg; Thomas R Williams; Dolores Yanagihara; Yelena Yakunina; Lawrence P Casalino Journal: J Gen Intern Med Date: 2011-12-13 Impact factor: 5.128
Authors: Jaymie R Meliker; Geoffrey M Jacquez; Pierre Goovaerts; Glenn Copeland; May Yassine Journal: Cancer Causes Control Date: 2009-02-15 Impact factor: 2.506