Application of the Time-Series Approach to Assess the Temporal Trend of Disparities in Health Outcomes

• Activity Number: 581
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Lin Tian*+
• Companies: CDC
• Keywords: Time series analysis ; Regression with ARMA errors ; Disparity ; Temporal trend ; Health outcomes

Abstract

The two-point-in-time-comparison method is a commonly used method for measuring changes in health outcome disparities over time. In this approach, the interim data between the 2 time points are ignored or incorporated that lead to loss of information and inaccurate conclusions. Tian et al. used the time-series approach to assess the temporal trend of racial disparity in Chlamydia prevalence. However, the details of the techniques were not provided. The purposes of this talk are to illustrate this statistical technique and to apply it into other health outcome disparity data analyses. A regression with ARMA errors model is illustrated to test trends in disparities in health outcomes when serial correlation was detected. First, a simple regression model is fit. If the residual series do not correspond to white noise, a 3-stage model-building strategy is used to obtain an adequate ARMA model. Finally a health surveillance data set is used to demonstrate the method using SAS ARIMA procedure.