From March 2006 to July 2010, a total of 498 interviews were carried out with informants in 22 communities across Nunavut. A standardised methodology was used for all the interviews to ensure comparability between cases and controls. For the cases, first-degree relatives (mostly parents) were approached for consent and then interviewed. The living controls were contacted first due to ethical reasons, but were not interviewed. Rather, controls’ relatives and/or friends were interviewed in order to keep the same methodology across both groups.
Overall, 498 interviews were conducted to obtain information on 240 individuals (120 cases and 120 controls). The number of informants per case or control was determined by the interviewer, based on the quality of the interviews and the amount of information gathered. For cases, 279 interviews were conducted (an average of 2.3 interviews per case), whereas for controls 219 interviews were carried out (an average of 1.8 interviews per control). Some interviews were carried out with 2 informants present (typically mother and father of a case). The majority of interviews were carried out with mothers, siblings, partners and friends of cases and controls ().
Types of informants used for the psychological autopsy interviews for both the case and control groups and the frequency that each type of informant was interviewed
Instruments used in the study measured socio-demographic data, psychopathology, impulsiveness, aggressiveness, history of suicide attempts, family history of psychopathology, development and life trajectory. The selection of these instruments reflected the major risk factors for suicide completion (i.e. previous suicide attempts, psychopathology, familial antecedents, high levels of impulsivity and aggression) (31
). Our group was well experienced with these instruments from using them in similar studies in Quebec ().
Description of the instruments that were used in the study
The instruments were thoroughly reviewed prior to the fieldwork to ensure that their content was appropriate for the Inuit context. Some items of the Life Trajectory Scale were modified, and others added in order to encompass important aspects of life in Inuit culture, such as experiences with qallunaat (white people), pride in Inuit culture, experiences with residential schools, opportunities to hunt and fish, ability to speak English, contact with the government and thoughts for the future of Nunavut among others.
The interviews were followed by a complete review of relevant medical records. For the controls, medical charts were available locally at the community's health centre. The charts for suicide cases in the Qikiqtani region were stored in the Qikiqtani Regional Hospital, while the charts of suicide cases in the Kivalliq and Kitikmeot regions were stored in the health centres of each deceased individual's community. With the available qualitative and quantitative data collected during the interviews and the information from the medical charts, the interviewer was able to determine whether each individual met specific SCID criteria for Axis I and II psychiatric diagnoses. Additionally, with information collected during the interview, the interviewer completed the measures of impulsivity and aggressiveness, history of suicidal behaviour, family antecedents of psychiatric disorders and the individual's life trajectory. Finally, the interviewer wrote a clinical-biographical narrative for each case or control, in which details of the individual's life were summarised. The biographical narrative described the individual's upbringing, familial relationships, academic performance, romantic experiences, interpersonal relationships, occupational life and detailed information about any psychiatric symptoms. This narrative, a copy of the medical records, and the completed set of instruments were sent to the coordinating centre for further processing.
At the coordinating centre, the instruments were assessed to ensure completeness. Any discrepancy in information between instruments (or between an instrument and the content of the narrative summary) was identified and resolved by discussion with the interviewer. The narratives were then blinded to the group each individual was in so that cases and controls could not be obviously distinguished (i.e. the case was disguised, details on the circumstance of death were removed and verbs were all changed to the past tense). The standardised case narrative, a summary of the medical records, and the final Axis I and II diagnoses were then forwarded to a panel of research collaborators to validate the final Axis I and II diagnoses that were given by the interviewer.
The panel was composed of a clinical psychiatrist, a clinical psychologist with research experience and a senior research coordinator with extensive experience with the SCID I and II instruments. A detailed examination of the diagnostic criteria was carried out for each participant, in order to validate or change the diagnoses. This included determining whether the developmental and clinical history, as described by the narrative, supported the clinical diagnoses. DSM-IV-TR criteria were used as the basis for these assessments. An example of a change made by the panel involved replacing an original diagnosis of adjustment disorder with depressed mood by major depressive episode in a situation where the depressive symptoms started just after an important loss, but included consistent suicidal ideation and gestures, sadness, anhedonia, guilt and motor retardation. Particular attention was also paid to diagnoses of substance abuse (since the informants sometimes minimised levels of substance abuse, even though the medical records and narratives made the level of abuse clear). Finally, the panel carefully verified Axis II criteria to detect potential overlap with Axis I symptoms, such as items for paranoid personality disorder being endorsed in the presence of a schizophrenia diagnosis. Typically, panel sessions lasted 1.5–2 hours, and 7–10 cases were examined per session ().
Flowchart of the procedures for data collection and evaluation.