Cumulative Geographic Residual Test Example: Taiwan Petrochemical Study Andrea Cook
Outline 1.Motivation Petrochemical exposure in relation to childhood brain and leukemia cancers Petrochemical exposure in relation to childhood brain and leukemia cancers 2.Cumulative Geographic Residuals Unconditional Unconditional Conditional Conditional 3.Application Childhood Leukemia Childhood Leukemia Childhood Brain Cancer Childhood Brain Cancer
Taiwan Petrochemical Study Matched Case-Control Study 3 controls per case 3 controls per case Matched on Age and Gender Matched on Age and Gender Resided in one of 26 of the overall 38 administrative districts of Kaohsiung County, Taiwan Resided in one of 26 of the overall 38 administrative districts of Kaohsiung County, Taiwan Controls selected using national identity numbers (not dependent on location). Controls selected using national identity numbers (not dependent on location).
Study Population Due to dropout approximately 50% 3 to 1 matching, 40% 2 to 1 matching, and 10% 1 to 1 matching. Leukemia Brain Cancer Cases Controls
Map of Kaohsiung
Cumulative Residuals Unconditional (Independence) Unconditional (Independence) Model definition using logistic regression Model definition using logistic regression Extension to Cluster Detection Extension to Cluster Detection Conditional (Matched Design) Conditional (Matched Design) Model definition using conditional logistic regression Model definition using conditional logistic regression Extension to Cluster Detection Extension to Cluster Detection
Logistic Model Assume the logistic model where, and the link function,
Residual Formulation Then define a residual as, Assuming the model is correctly specified would imply there is no pattern in residuals. => Use Residuals to test for misspecification. Cumulative Residuals for Model Checking; Lin, Wei, Ying 2002
Hypothesis Test Hypothesis of interest, Geographic Location, (r i, t i ) Independent of Outcome, Y i |X i Cumulative Geographic Residual Moving Block Process is Patternless
Unconditional Cluster Detection Define the Cumulative Geographic Residual Moving Block Process as,
Asymptotic Distribution However, the distribution of, is hard to define analytically, but we have found another distribution that is asymptotically equivalent, which consists of a fixed component of data and random variables
Significance Test Testing the NULL Simulate N realizations of Simulate N realizations of by repeatedly simulating, while fixing the data at their observed values. Calculate P-value Calculate P-value
Conditional Logistic Model Type of Matching: 1 case to M s controls Data Structure: Assume that conditional on, an unobserved stratum-specific intercept, and given the logit link, implies, The conditional likelihood, conditioning on is,
Conditional Residual Then define a residual as, => Use these correlated Residuals to test for patterns based on location.
Conditional Cumulative Residual However, the distribution of, is hard to define analytically, but we have found another distribution that is asymptotically equivalent, which consists of a fixed component of data and random variables
Significance Test Testing the NULL Simulate N realizations of by repeatedly simulating, while fixing the data at their observed values. Calculate P-value
Application Study: Study: Kaohsiung, Taiwan Matched Case-Control Study Method: Method: Conditional Cumulative Geographic Residual Test (Normal and Mixed Discrete)
Results Odds Ratio (p-values) Marginally Significant Clustering for both outcomes without adjusting for smoking history.
Childhood Leukemia
Childhood Brain Cancer
Discussion Cumulative Geographic Residuals Unconditional and Conditional Methods for Binary Outcomes Unconditional and Conditional Methods for Binary Outcomes Can find multiple significant hotspots holding type I error at appropriate levels. Can find multiple significant hotspots holding type I error at appropriate levels. Not computer intensive compared to other cluster detection methods Not computer intensive compared to other cluster detection methods Taiwan Study Found a possible relationship between Childhood Leukemia and Petrochemical Exposure, but not with the outcome Childhood Brain Cancer. Found a possible relationship between Childhood Leukemia and Petrochemical Exposure, but not with the outcome Childhood Brain Cancer.