A test of the substitution hypothesis: An analysis of urban and rural trends in solid/liquid poisoning suicides in Taiwan

Abstract

Taiwan experienced both a significant decrease and a significant increase in the suicide rate during the period 1986–1999, which provides a unique opportunity to examine the substitution hypothesis in suicide research: that is, whether a reduction in the use of a certain method of suicide would be offset by a parallel increase in the use of other methods. We also explored whether such method substitution, if it existed, differed across urban and rural settings. Data on age-, sex-, method-, and urban-/rural-specific suicide rates for the period 1986–1999 in Taiwan were obtained and the year of 1993 is found to be the inflection point. We analyzed using Poisson regression to estimate the average annual percentage change (AAPC) for periods of decline (1986–1993) and increase (1993–1999) in suicides. The rapid decline in suicide by solid/liquid poisoning (mostly using pesticides) during a period of accelerated economic development (1986–1999) in Taiwan was found to be associated with the increased use of alternative methods. An interaction model found a marked decrease in solid/liquid poisoning suicide in both urban and rural Taiwan over the period of decline (1986–1993). The extent of the decrease was greater in rural areas but was accompanied by a rise in the use of several other suicide methods. However, the net effect was still a marked reduction in the suicide rate. A general increase in suicide among all age groups, for all methods, in both rural and urban settings, and for both sexes was found during the period when the suicide rate increased (1993–1999). We conclude that restricting access to the means of pesticide suicides reduces not only the method-specific suicide rate but also the overall suicide rate; nonetheless, suicide method substitution is not an all-or-nothing phenomenon. Stratified analyses by geographical (i.e. urban/rural) area can help to disentangle the patterns in each subgroup, which will improve our understanding of the phenomenon and make suicide prevention efforts more focused and effective.