Complex Survey Techniques

Aims Of Project

The ready availability of public-use data from large national population-based complex surveys has immense potential for the assessment of (1) population frequency of cancer; (2) hospital length of stay and related costs for treatment; (3) cancer screening rates; (4) associations between risk factors (e.g. screening rates, diet) and cancer. The goal of this project is to demonstrate this potential using novel statistical methods with United States complex surveys, including the Behavioral Risk Factor Surveillance System (BRFSS) and Health Information National Trends Survey (HINTS) to describe screening rates, the National Health and Nutrition Examination Survey (NHANES) to explore behaviors (diet, smoking, etc) in current and future cancer patients, the Nationwide Inpatient Sample (NIS) and the Medical Expenditure Panel Survey (MEPS) to describe hospital length of stay and related costs for treating cancer, the National Home and Hospice Care Survey (NHHCS) to explore end-of-life care for cancer patients and the National Health Interview Survey (NHIS) to examine follow-up of cancer survivors. An additional goal is to make the newly developed statistical/epidemiological methodology widely accessible to non-statisticians.

For valid inference with complex surveys, both the design (stratification and clustering) and the subject-specific probability of selection should be incorporated into the analysis. Despite the many recent advances in methods for analyzing complex survey data, most software packages have severe limitations in many common settings, such as subgroup analyses or non-parametric approaches. Three Specific Aims are developed, which follow.

1. Statistical Approaches for Small Subgroup Analyses 2. Statistical Procedures for Large Databases 3. Statistical Methods for Skewed Data