Suicide attempts assessment in the medical wards: factors influencing admission to a psychiatric unit

Globally, suicides account for 50% of all violent deaths in men and 71% in women.¹ In New Zealand, approximately 550 people die by suicide annually, exceeding the combined deaths from road traffic accidents and homicides, with an over-representation of Māori and Pacific peoples.² There is a reasonable amount of local and international data on the demographic variables of those who complete suicide. Interestingly, mounting evidence suggests that suicide attempters and completers represent two distinct, partially overlapping populations.³ United States of America (USA) data showed that among 73,490 suicide deaths, 57,920 (79%) occurred on the first attempt,⁴ while according to coronial findings, among suicide deaths recorded in the Waikato Health New Zealand Region between 1 January and 31 December 2021, approximately two-thirds (21 out of 32 cases) of individuals had no documented history of previous suicide attempts (in an email from the Office of the Chief Coroner [Coronial.Information@justice.govt.nz] on 21 Nov 2024). Moreover, in the USA, fewer than 1 in 35 suicide attempts were successful.⁵ The ratio between increased suicide risk and actual attempts is even larger. In 2020–2022, it is estimated that one in six Australians (16.7% or 3.3 million people) aged 16–85 years had experienced suicidal thoughts or behaviours in their lifetime, while only 0.3% had attempted to take their own life.⁶ However, those who have attempted suicide are up to 30 times more likely to go on to complete suicide.⁷ In spite of this and despite its public health relevance, suicide remains challenging to predict. Research suggests that risk scales alone have limited utility in managing suicidal behaviour.⁸

In July 2023, the Consultation – Liaison Psychiatry team at Waikato Hospital implemented a semi-structured interview based on the Suicide Attempt Self-Injury Interview (SASII)⁹ to assess all patients admitted to a medical ward after a deliberate self-harm event. After 12 months, an audit was conducted. The objective was to analyse if there were significant associations between individual questionnaire items and the predictive likelihood of a psychiatric admission, to further improve the early detection of patients needing transfer to a psychiatric unit after medical clearance. The data were collected, analysed and interpreted using Microsoft Excel.

A total of 107 patients were assessed during the 12-month period and included in the final analysis. Female gender was prevalent (64%), and most patients were adults (75%), with an average age of 31.75 years (SD 16.9). Multiple response ethnicity showed a slight over-representation of Māori population (28% in the sample vs 25% in the 2023 census for the Waikato Region).

Self-harm events were predominantly due to overdose (85%), mostly involving pain medication or psychiatric medication. Over one-third of the patients required intensive care unit/high dependency unit (ICU/HDU) admission. Most episodes were impulsive (70%), with only 10 patients (9%) acknowledging a high degree of planning. While few patients (23%) self-harmed under the influence of substances, alcohol was the primary substance in these cases (84%). Interpersonal conflict was the main trigger for the events (50%), followed by depressive symptoms (21%). More than half of the patients self-harmed intending to end their lives, while others described different goals or were unable to provide a clear rationale. Although the primary reason for self-harming was to avoid suffering (44%), almost the same number of patients could not provide a clear rationale (45%). Over half of the patients alerted others before or after the attempt (53%), and almost half sought help immediately after self-harming (49%). Only 20% of the patients showed low regret when assessed, but most had a history of previous attempts (63%). Further details of the outcomes are described in Table 1. A total of 15 patients (14%) were required to be transferred to a psychiatric inpatient unit after being medically cleared. In a regression analysis (Table 2), only low regret had a significant association with the need for psychiatric admission.

The use of a semi-structured interview based on the SASII did not show to be particularly useful in our sample to identify patients who would require continued mental health follow-up in an inpatient unit. People who self-harm often have complex and difficult life circumstances and need assessment, but we need to move away from assessment models that prioritise risks at the expense of needs. Although further research should confirm these findings, it might be worth noting that low regret could be a warning sign for requesting further mental health assessment in secondary care settings.

View Table 1–2.