| Outline of Annual Research Achievements |
In the first year of this project, we have collected daily time-series data on suicide and weather factors, including 384 locations across 17 countries. Data collection for three more countries is now in progress.
With a subset of the data for 12 countries, we examined seasonal patterns of suicide and spatial heterogeneity of the suicide seasonality, using a conditional Poisson regression model and a meta-regression model. We constructed weekly time-series data for the analysis due to low counts of suicide in small areas. We observed consistent seasonal patterns of suicide with a peak in spring and a trough in winter. In addition, we identified some potential regional characteristic to explain the seasonality of suicide in which the peak to trough ratio was more significant for colder areas and older populations in general but weak for areas with higher unemployment rates. We also examined temporal changes in the suicide seasonality by comparing two periods for eight countries, although the results were inconclusive.
In the second year of this project, we will update our database further and examine the spatial heterogeneity of the temperature-suicide association. Also, we plan to refine the statistical model to scrutinize the temporal changes for both seasonality of suicide and the association with temperature.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
As of the end of FY2019, we have updated our existing multi-country multi-city data containing daily time-series of suicide and weather factors. Specifically, we extended a study period for two countries (Canada and Spain) and added newly collected five countries with 25 locations (Chile, Estonia, Italy, Mexico, and Romania). Therefore, the data now contains 384 locations across 17 countries in total from 4 to 41 years (please see the list* below).
To examine spatial heterogeneity of suicide seasonality, we extracted a subset of 12 countries from the data and merged them with some regional characteristics derived from another database. We observed consistent seasonal patterns of suicide as expected, with a peak in spring and a trough in winter for 10 out of the 12 countries. The rest two of them showed summer peaks of suicide rather than a spring peak. In the meta-regression analysis, we observed that the seasonality of suicide was larger for colder areas and older populations in general but weak for areas with higher unemployment rates.
*List of the 17 countries: Brazil, Canada, Chile, Estonia, Italy, Japan, South Korea, Mexico, Philippines, Romania, South Africa, Spain, Switzerland, Taiwan, UK, USA, and Vietnam
|
| Strategy for Future Research Activity |
In the second year of this project, we are planning to add three more countries (Australia, Finland, and Ireland) to the database that discussion with overseas collaborators is in progress. Also, we will examine the spatial heterogeneity for an association between temperature and suicide.
Although we have examined temporal changes in the suicide seasonality in FY2019, the results were inconclusive. We presume that the methodology comparing two periods might be limited to investigate the temporal trend of the suicide seasonality. We are going to refine the statistical models and scrutinize the temporal changes of the seasonality, which we believe would also help to fit the model for temporal changes in the temperature-suicide association.
|
