Assessing Seasonality Variation with Harmonic Regression: Accommodations for Sharp Peaks.

Area of Science:

• Epidemiology
• Biostatistics
• Time Series Analysis

Background:

• Harmonic regression models are standard for analyzing seasonal disease patterns.
• Existing models struggle with diseases exhibiting sharp, asymmetric seasonal peaks.
• Complex disease incidence fluctuations require more sophisticated modeling techniques.

Purpose of the Study:

• To propose an improved harmonic regression approach for modeling sharp seasonal disease peaks.
• To enhance the fit of seasonal patterns in epidemiological data with complex fluctuations.
• To provide a more accurate method for assessing disease seasonality.

Main Methods:

• A two-step harmonic regression strategy was developed.
• Step 1: Basic model to estimate the seasonal peak.
• Step 2: Extended model with sine and cosine transforms to mimic quadratic terms for fitting sharp peaks.

Main Results:

• The proposed two-step method demonstrated improved model fit for data with sharp seasonal peaks.
• Simulated and actual data confirmed the effectiveness of the enhanced approach.
• The new functions successfully captured specific behaviors of complex seasonal patterns.

Conclusions:

• The two-step harmonic regression approach offers a significant improvement for modeling sharp seasonal disease peaks.
• This method is recommended for assessing seasonality across various diseases with complex seasonal profiles.
• The enhanced model provides a more robust tool for epidemiological and biostatistical analyses.