Predicting the United States Syphilis Rates in Years 1957-2009 Based on ARIMA, ARCH, and GARCH Time Series Models

Over the years, the United States (U.S.) has experienced rising incident rates among sexually transmitted diseases. For example, syphilis rates have increased from 11.52 per 100,000 cases in 2000 to 15.40 per 100,000 cases in 2008 (CDC Wonder, 2008). Research indicates that the increasing number of syphilis cases may be due to economic stressors and behavioral choices. This growing concern illustrates the need for effective interventions that require changes in economic and behavioral conditions. The purpose of this study is to predict the number of syphilis cases among the U.S. population at risk through the use of ARIMA and ARCH/GARCH models. Risk factors such as state median income, state unemployment rate, state poverty rate, and other variables were used in this study. The U.S. syphilis data from 1957-2009 will be collected from the Centers for Disease Control and Prevention and analyzed using STATA to determine which variables contribute to the syphilis rate. When looking at syphilis rates during this time frame, the identified risks within the US changed; therefore, there were periods of high volatility and periods of stability. Ways to address those changes in variance over time is to examine independent variables that measure the volatility. The inconsistency of these peaks and valleys present difficulties when predicting future data; therefore, the use of ARCH/GARCH models, have become the essential tools for measuring volatility and deciphering related questions in time series analysis.