68-3

73/130

start finding value in one’s life and to participate in one’s own treatment17).The items that indicate the signs of imminent suicide risk in Table 2 represent the perspectives of the staff rather than those of the patients. The five major categories our analysis yielded reveal that suicide risk information can be obtained from ‘signs emanating from the patient’, ‘signs gleaned through engagement’, ‘signs from response to treat-ment’ (signs that occur as a result of medical inter-vention), ‘signs associated with reports from the family’, and ‘signs inferred from multiple sources of information’ (denoting the need to base decisions on multiple kinds of information). In psychiatric wards, not only the principal doctor and nurse in charge, but also staff from different professions are stationed to provide care collectively. The perspec-tives extracted in our analysis can be particularly useful to staff other than those who are immedi-ately in charge of patients. To such staff, the perspectives provide observational insights on the patients and can serve as criteria in deciding on the kinds of information to include (and excluding) while reporting to medical teams.In-patient suicide is an example of a sentinel event─a sudden and unanticipated event or complication in a healthcare setting that results in death or serious injury. We used the fishbone diagram, which is an example of a sequential acci-dent model. Other models for analysing accidents extend their focus to epidemiological and systemic accident models18). There is a drive for suicide in patient, so such issues with the handling of infor-mation by staff should be seen less as a factor that causes a suicide event and more as a factor pertaining to the defence barriers in an epidemio-logical accident model. Such models envisage effec-tive communication among healthcare workers as a defence against accidents. Defences must be as solid as possible to prevent accidents. We extracted five issues pertaining to how hospital staff handle information: ‘omission in diagnostic records during admission’, ‘omission in conference records’, ‘commu-nication lapse during transfer’, ‘need for integrated information’, and ‘systemic issues’. Information may be omitted during admission, conferences, or ward transfers because the staff are busy or have a defensive mental state, or the hospital is under-staffed. To address these issues, it may be useful to build a system that makes omissions impossible, such as by standardising the reporting format. The staff may be particularly tempted to omit or abridge information in the case of readmission. They may do so out of complacency, believing that the facts on the readmitted patient are already well known. The fact that a patient is being readmitted because of the lack of improvement in their condi-tion may create a sense of helplessness among the staff, which may, in turn, dull their alertness to the suicide risk. Thus, hospitals must impress upon the staff the importance of recording suicide risk infor-mation during readmission. We found that commu-nication lapses occur during ward transfers because it is [hard to relay complex information such as the principal doctor’s decision-making process]. It is difficult to explicate the principal doctor’s often inscrutable thinking to the staff at the destination ward. For example, the doctor may give the staff seemingly contradictory instructions to transfer a patient from a seclusion room in an open ward to a general area in a closed ward, i.e. to increase or decrease the level of protection. It is therefore essential to develop strategies to convey this infor-mation. For example, when the staff in the destina-tion ward prepares a nursing plan for the trans-feree, the principal doctor and staff from the original ward can participate in the process. The ‘need for integrated information’ implies the need for collating information from different staff members and evaluating multiple sources of infor-mation in an integrated manner. This task poses a major challenge in that it is subject to human factors such as how busy one is and what their career and abilities are. Possible strategies toward addressing ‘systemic issues’ include introducing an electronic medical record and displaying informa-tion on past suicidal behaviour patterns in places where the staff can easily see it.This study has four limitations. First, the results have limited transferability given the small size and peculiarity of the sample. Second, our analysis was unable to cover factors that were not included in the records. Third, as the cases were retrospec-tively analysed by parties who were not involved, the analysis could not cover the perspectives of psycho-dynamics and group dynamics between each case and staff and the participant observation of the staff. Fourth, another aspect that was not 269