Predictors of repeated adolescent self-harm: The display of multiple variables and mediators in a 2-year study.

BACKGROUND: Data on self-harm (SH) repetition in non-Western adolescents are limited; this study is to survey the predictors. METHODS: A total of 5879 adolescents (mean age 16.02 years) in Northern Taiwan were recruited. The participants filled in online questionnaires about their sociodemographic data, suicidality, depressive symptoms, self-esteem, social support, family discord, impulsivity, and alcohol and tobacco use at baseline (T1) and at the 1 year follow-up (T2). We used logistic regression analysis to examine the predictors of SH continuation. Generalized structural equation modeling (GSEM) was then estimated to analyze the treatable variables for both years and to investigate their relationships and mediating effects. RESULTS: A total of 125 students were identified as being in the SH continuation group; while 470 students were identified as being in the SH stop group. The SH continuation rate was 21%; no significant gender difference was found. Logistic regression analysis showed that the predictors of SH continuation were low school ranking, poor quality of listening from relatives, use of the cutting method for SH, and a suicide plan in the past year at T1, and more depressed mood, use of the cutting method for SH, more suicide ideation and plans at T2. Similar predictors were found by GSEM; self-esteem at T1 and depressed mood at T2 were found to be mediators in the pathways. CONCLUSION: The continuation rate of SH was similar to that reported in Western countries. These predictors should be included in the treatment plan to prevent SH continuation.