中華心理衛生協會

Purpose: This macro-epidemiologic study aimed to develop a statistical regression model for identifying important predictors of the crude annual suicide mortality rate in Taiwan. Methods: A time series analysis with seasonality was conducted to examine the associations between 33 socioeconomic variables and the crude annual suicide mortality rates using the male and female data of 4 age groups in 7 regions of Taiwan between 1981 and 2007. Results: A well-fitted multiple linear regression time series model (coefficient of determination, R2 = 0.8661) revealed that conditioning on the effects of the past crude annual suicide mortality rates and temporal oscillations, (1) the higher the male unemployment rate of the t - 3 year, the higher the crude annual suicide mortality rate of the current year t; (2) if the gross national product (GNP) per capita of the t - 1 year minus that of the t - 2 year was larger than NT $10,000, then the crude annual suicide mortality rate of the current year t would drop; and (3) if the consumer price index of the t - 2 years was greater than or equal to 80, then the crude annual suicide mortality rate of the current year t would go up. Based on this statistical regression model, the crude annual suicide mortality rates of 2008 and 2009 were predicted and found to be reasonably close to the actual values of the official death statistics released after the study period. Conclusions: The fluctuation of economic conditions had a much stronger and more direct impact on the suicide mortality than that of social conditions in Taiwan. In particular, 3 economic indices were identified as having 1- to 3-year short-term lagged effects on the crude annual suicide mortality rates from 1981 to 2007. Given the relatively poor economy from 2007 to 2009 in Taiwan, a remarkable decrease in the crude annual suicide mortality rate is not expected in the coming years.