Suicide seasonality: complex demodulation as a novel approach in epidemiologic analysis

Area of Science:

• Epidemiology
• Time Series Analysis

Background:

• Suicide seasonality is widely recognized but shows varying trends globally.
• Previous research suggested a link between seasonality strength and suicide prevalence.
• Methodological challenges exist in accurately analyzing changes in suicide seasonality.

Purpose of the Study:

• To investigate methodological difficulties in examining changes in suicide seasonality.
• To analyze the hypothesis of decreasing suicide seasonality using a continuous modeling approach.
• To assess the association between the strength of seasonality and absolute suicide numbers.

Main Methods:

• Analysis of suicide data from Austria (1970-2008, N=67,741).
• Application of complex demodulation, a local harmonic analysis method, for continuous seasonality modeling.
• Utilized regression models to evaluate time trends and the association between seasonality amplitude and suicide counts.

Main Results:

• The strength of suicide seasonality was found to be significantly associated with absolute suicide numbers.
• When this association was accounted for, the strength of seasonality remained stable throughout the study period.
• Continuous modeling avoided issues associated with piecewise time series segmentation.

Conclusions:

• Continuous modeling of suicide seasonality, specifically with complex demodulation, prevents spurious findings.
• Disregarding the association between seasonality strength and suicide prevalence can lead to inaccurate conclusions.
• The study highlights the importance of advanced analytical methods for reliable suicide seasonality research.