Developing an early warning system of suicide using Google Trends and media reporting

Area of Science:

• Public Health
• Epidemiology
• Data Science

Background:

• Conventional suicide surveillance systems have significant time lags (6 months to 2 years).
• There is a need for rapid detection of potential suicide outbreaks.

Purpose of the Study:

• To develop an early warning system for suicide outbreaks in Hong Kong.
• To utilize Google Trends and media reporting data for real-time suicide prediction.

Main Methods:

• Collected 3,534 suicide records (2011-2015) from official Hong Kong sources.
• Employed Poisson regression models with Google Trends and media data to predict weekly suicide counts.
• Utilized cumulative sum (CUSUM) control charts to identify suicide outbreaks.

Main Results:

• The model achieved low prediction errors (15.6%-24.16%) for suicide numbers.
• Suicide intensity was well-captured for young males and females.
• The CUSUM method demonstrated high precision (100%) for young females and (54%) for young males.

Conclusions:

• Google Trends and media reporting are valuable for predicting suicide outbreaks in Hong Kong.
• The developed system can support timely and targeted public health interventions.
• Further research could incorporate social media, socioeconomic, and climate data.