Transfer Function Model for COVID-19 Deaths in USA Using Case Counts as Input Series.

Area of Science:

• Epidemiology
• Biostatistics
• Time Series Analysis

Background:

• Accurate forecasting of COVID-19 (Coronavirus Disease 2019) mortality is crucial for public health preparedness.
• Existing time series models may not fully capture the dynamic relationship between COVID-19 cases and deaths.
• The Center for Disease Control (CDC) provides vital data for epidemiological modeling.

Purpose of the Study:

• To develop and evaluate a transfer function time series model for predicting COVID-19 deaths.
• To utilize reported COVID-19 case positivity counts as the primary input series for the forecast model.
• To compare the predictive performance of the proposed transfer function model against established statistical methods.

Main Methods:

• A transfer function time series model was developed using COVID-19 case and death data from the USA (July 24 - December 31, 2021).
• The model's forecast errors were compared against those generated by an autoregressive integrated moving average (ARIMA) model.
• Separate ARIMA models were also fitted for COVID-19 cases and deaths to provide additional benchmarks.

Main Results:

• The transfer function model demonstrated superior forecast accuracy compared to the autoregressive integrated moving average model.
• Analysis of forecast errors indicated a significant improvement in predictive performance with the transfer function approach.
• The study provides a robust methodology for leveraging case data to forecast mortality trends.

Conclusions:

• The transfer function time series model offers a more effective approach for forecasting COVID-19 deaths than standard ARIMA models.
• This methodology can enhance the accuracy of public health predictions and resource allocation during pandemics.
• Utilizing case positivity as an input series provides valuable insights into mortality trends.