Trivellore Raghunathan Chair and Professor of Biostatistics School of Public Health Research Professor Institute for Social Research University of Michigan Presented at the National Conference on Health Statistics August 1618 2010 ID: 1035699
Download Presentation The PPT/PDF document "A Bayesian change point analysis of prev..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. A Bayesian change point analysis of prevalence rates of chronic diseases by gender and raceTrivellore RaghunathanChair and Professor of Biostatistics, School of Public HealthResearch Professor, Institute for Social ResearchUniversity of MichiganPresented at the National Conference on Health Statistics, August 16-18, 2010
2. SetupNeed for assessing the health of the nation cannot be understatedNCHS has been in the forefront of collecting data neededA look back on one such long seriesNational Health Interview Survey (as it is currently known)Began in 1957Limited Data available prior to 1962Relate the prevalence information to time line of our history
3. DataDiseasesDiabetes (1957-1959, 1962-2008)Asthma and Hypertension (1962-2008)PopulationsOverallMale/FemaleWhite/BlackTime lineImportant events (health and economic)The analysis is preliminary
4. Prevalence of DiabetesCivil Rights (64)Medicaid/AA (65)Medicare (66)Recession (74)Oil crisis (73-75)Hostage (79)Recession (80)Gulf warRecession (90)Vietnam WarRecession (97)9/11
5. Model
6. AnalysisFit the model using MCMC or MLECalculateChoose knots based on the time-lineFor this analysis choose 5 knots (1965, 1974, 1980, 1990, 2001)Modification (in progress) Treat knots as unknown parameters and estimate them from the data
7. Diabetes by Gender
8. Diabetes by Race
9. Prevalence rates of Asthma
10. Prevalence rates of Hypertension
11. Diabetes by administrationJohnsonNixon/FordCarterReagan/BushClintonBush
12. AsthmaJohnsonNixon/FordCarterReagan/BushClintonBush
13. HypertensionJohnsonNixon/FordCarterReagan/BushClintonBush
14. Issues and ConcernsMajor design changes, instruments, definitions etcReflecting such changes in the model developmentAny policy change takes time to have an impact. Incorporation of time lag in the modelOutliersRegardless of these concerns:Persistent differences between raceLessening of differences between men and women
15. DiscussionRelating the time series data to important epics/events can be useful to understand potential effects of policy/cultural changesInvestigation of health disparities across various subpopulations Subgroups can be determined based on economic, social, cultural aspects of the societyUseful exercise for any democratic society to look at the health of the nationThank you NCHS, for this great data source!!