5032.0: Wednesday, November 15, 2000 - Table 3

Abstract #10540

Developing and evaluating the NYS WIC Program caseload and food expenditure forecasting system using time series analysis

Gene Shackman, PhD1, Wayne DeNyse, MS1, and Abdel Hamdan, PhD2. (1) Department of Health, Division of Nutrition, New York State, 150 Broadway, River Review, 6th floor west, Albany, NY 12223, 518-402-7304, gxs03@health.state.ny.us, (2) Division of Nutrition, Evaluation and Analysis Unit, NYS Department of Health, 150 Broadway, FL6 west, Albany, NY 12203

The New York State WIC program, with a budget of approximately $360 million, provides services each month, on average, to 480,000 low-income women, infants and children through a statewide network of 100 local WIC agencies with oversight of over 500 WIC sites. The WIC program is a USDA sponsored discretionary grant program, rather than an entitlement program, so the number of eligible participants that can be served depend, for the most part, on appropriations from the US Congress. In order to ensure that the maximum number of eligible participants is served, accurate forecasting of expenditure and participation is critical.

We present a number of time series models that we use to predict food expenditure and participation for up to four months ahead. The models used included the Autoregressive Moving Average , transfer function, and intervention analysis models. The results are evaluated using the mean absolute percentage error (MAPE), the consistency of performance, the tendency to over or underestimate, and the range of errors. Our results showed that our forecasts were over 99 percent accurate, that the errors were consistently small and that results were generally within a specified range.

We also present methods we use to forecast participation and food expenditure for one year in advance, testing quantitative naive methods such as absolute change, absolute percentage change, and three year moving average of absolute percentage change. Using the method that gave the smallest MAPE, most consistent and most unbiased, enabled us to forecast yearly participation and expenditure with 99 percent accuracy.

Learning Objectives: At the end of the session, the participants will have a clear understanding of the following: 1. What is WIC program and it's objective? 2. What is time series analysis? 3. When do we use time series analysis? 4. How do we apply it to a public health program. 5. How do we measure the accuracy of a forecasting model. 6. How accurate are the forecasts produced the time series models

Keywords: WIC, Intervention

Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

The 128th Annual Meeting of APHA