Participants and procedures
A two-stage stratified cluster sampling method was used in order to select research sites and 12 counties were randomly selected to recruit cases in Shandong, Hunan and Guangxi Province, which are top 10, 11–20, and 21–31 provinces ranking of GDP per capita, respectively. A case-control psychological autopsy study [14, 15] was conducted to collect targets’ information from June 2014 to September 2015.
All village doctors and local public health professionals involved in death certification were given brief training and required to report all elderly suicide death to local Centers for Disease Control and Prevention (CDCs) before the study begins. The investigators were composed of the teachers and the graduate students from Shandong University, Central South University and Guangxi Medical University. All investigators were trained intensively for 10 days in a standardized way on the instruments and the skills of the interview. The manner of death was determined by the investigators.
In each county, suicide cases aged 60 and above were collected consecutively based on the death certification system. Age (± 3years)-, sex-, and location-matched living people were selected as the community control group. For every target, two informants were selected. Generally, the first informant was one next-to-kin who lived with the targets; and the second informant was always a friend, a neighbor, or a remote relative. It means there were 5 interviews in one pair of case-control. We did 1210 interviews in total for the 242 completed suicides and living controls. The survey was conducted in face-to-face interviews. The average interview time was 90 min.
The study was approved by the IRB of the Shandong University, Central South University, and Guangxi Medical University. The aim and procedure of the research were explained to all participants. Written informed consent must be obtained before interviews were conducted.
Measurements
Demographic characteristics
Demographic data were collected including age, sex, marital status, family income, physical disease, mental disorder, and living arrangements. In the study, people that were married and living with a spouse, or cohabitating were classified as “stable”, while people with other marital status were classified as “unstable”.
Life Event Scale for the Elderly (LESE)
In this study, life events were measured by the LESE, which was developed specifically for older Chinese adults [16]. It is a validated and reliable scale used among the elderly generally, and in particular, elderly people who are suicidal[16, 17]. We divided 46 life events into three separate categories: (1) Health/Hospital, 16 items; (2) Family/Home, 18 items; (3) Friends/Relationships, 12 items. Each of the 46 life event items was recorded for five questions: (1) The date it happened, measured by never occur, in one month or one year or long term; (2) Whether it was positive or negative for the target person; (3) The impact to the target person’s mental health, measured by a five-point scale from no impact to very severe impact; (4) The duration of the event, measured by a four-point scale from three months, six months, less to one year or longer; (5) The number of times it happened.
Statistical analyses
The information which was provided by two informants was combined as proxy data for the suicide victims and living controls, and this proxy data was then used. For demographic characteristics, the information provided by the first informant was relied on. Answers which were associated with an increased risk of suicide were used when two informants reported differently for each item of LESE. The rationale for this practice is that a targeting behavior may exist as long as one of the two informants has observed it.
The t-test and chi-square test were used to compare different demographic characteristics between the people who experienced fewer life events and more life events both among suicide victims and living controls, respectively. χ2 was used to analyze the incidence of 46 life events between suicides and controls. χ2 was also used to compare the frequencies of top 10 life events between men and women among suicide victims. Logistic regression analysis was used to explore the independent factors for suicides.
All analyses were performed using SPSS 24.0 for Windows (SPSS Inc, Chicago, IL). In the study, statistical significance was determined prior to Ρ<0.05.