The identifiability for depressive symptoms of different methods in suicide prevention research / 中国神经精神疾病杂志

ABSTRACT

Objective To compare the identifiability for depressive symptoms using different instruments while interviewing with different respondents in suicide prevention research in China. Methods One hundred and fifty-one suicide death cases (suicide group) and one hundred and twenty suicide attempt cases (attempt group) were recruited. For each identified cases, one family member proxy respondent, and another associate proxy respondent (friend or neighbor) and suicide attempter (only for attempt group) were interviewed separately by qualified psychiatrists. The Di-agnostic Screening Instrument for Depression (DSID) and the Structured Clinical Interview for DSM-Ⅳ Axis Ⅰ Disorders (SCID-Ⅰ) were administered to each respondent to identify the depressive symptoms based on diagnostic criteria for major depressive episode in DSM-Ⅳ. Data collected from family members and associate respondents were merged as proxy data. The concordances of the DSID and SCID-Ⅰfor identifying depressive symptoms, meeting for criteria of Major Depressive Episode (MDE) and Mild and Major Depressive Episode (MMDE), were calculated based on different respondents' data. The prevalence of depressive symptoms, MDE and MMDE, were compared among merged proxy data, family member respondent's data, and associate respondent's data in suicide group and attempt group, and between self-respondent's data and merged proxy data in suicide attempt group. Results In suicide group, based on merged proxy data, the prevalence of MDE was 41.1%(62 cases) for DSID and 41.7%(63 cases) for SCID-Ⅰ, and the Kappa coeffi-cient was 0.77. Based on suicide attempters' self-raported data, the prevalence of MDE was 23.7% (27 cases) and 22.0% (24 cases) for DSID and SCID-Ⅰ respectively, with a Kappa of 0.74. Based on merged proxy report in attempt group, 16 (13.3%) and 15 (12.5%) cases were met for criteria of MDE (Kappa=0.89), using the 2 instruments. In both of the suicide and attempt groups, the merged proxy data got higher prevalence of depressive symptoms, MDE and MMDE than that only based on family respondent's data or associate's respondent's data using both of the 2 instruments (all P<0.05). Compared with merged proxy data, attempters' self-reported data got higher prevalence of MMD and MMDE using both of the 2 instruments (all P<0.05). Conclusions Based on same respondent's data, SCID-Ⅰ performs as well as DSID in identifying depressive symptoms. Collecting data from 2 respondents would get higher prevalence of MDE or MMDE than only from one family member or one associate. In attempt group, the prevalence of MDE or MMDE based on merged proxy data were lower than that based on attempters' self-reported data.