An audit of risk assessments for suicide and attempted suicide in ED: a retrospective review of quality

Suicide, the act of intentionally killing oneself (as determined by coronial ruling), is rare but remains a significant cause of mortality in 34 of the Organisation for Economic Co-operation and Development (OECD) affiliated countries.1 In 2014, 504 people died by suicide in New Zealand (ie, 10.7 per 100,000 population).2 In contrast, the number of intentional self-harm hospitalisation events in 2013 (including short stays in the emergency department) was 7,267 (ie, 176.7 per 100,000 population).2 In 2009, New Zealand had the highest rate of suicide in the OECD.3

Self-harm includes broader classes of behaviours that includes non-suicidal self-injury (NSSI) where the intent to die is absent. Attempted suicide is a deliberate self-harm (DSH) event with the purpose of ending one’s life but does not lead to death.4–5

Assessments following suicide attempts can be challenging evaluative exercises. However, society expects that it will be completed with some degree of accuracy.6 The strongest predictive factor for future suicide is a history of previous suicide attempts, especially if this occurred recently.7 Stressful life events or psychological distress (eg, relationships and financial difficulties) are frequent precipitating factors.8–9 Multiple psychiatric disorders (depression, anxiety, substance abuse disorders, personality disorders, etc.) have been associated with an increase in attempted suicide risk.10 In 2006, depression and anxiety were ranked as the second leading cause of health loss in children, adolescent and middle-aged people.11 As the population ages, the presence of social isolation, lack of connectedness and chronic medical conditions also contribute as risk factors for suicide.12–13

When patients present to the hospital’s emergency department (ED) following a suicide attempt, the psychiatric assessment typically occurs once clinicians have confirmed medical stability, and the patient is sufficiently alert to participate in an interview. Mental health clinicians conduct a comprehensive, individualised psychiatric assessment to help identify the high-risk group for repeat suicide attempts.7 Recent studies showed that between 1.6–2% of hospital-treated people, following an attempted suicide episode, will die by suicide in the next 12 months.14–15

The focus of this audit was to determine the quality of the psychiatric risk assessments conducted by mental health clinicians. This would help indicate which components of the psychiatric risk assessment, recommended in the Ministry of Health’s (MOH) Clinical Practice Guideline (CPGs), are routinely and satisfactorily completed, and which are neglected or do not adhere to CPG recommendations.

Methodology

This was a descriptive study. A retrospective review of the electronic medical records was performed for patients who had presented to Waikato Hospital ED with either active suicidal thoughts or a suicide attempt over a 12-month period (1 July 2015 to 30 June 2016). An additional inclusion criterion was a subsequent assessment by a mental health clinician.

Using the outlined criteria and relevant coding for the various DSH methodologies, over 900 medical records were obtained. The original pool of medical records was randomised (by way of using a random number generator formula process in Excel), and the first 376 files were reviewed.

The Royal Australian and New Zealand College of Psychiatrists (RANZCP) has published CPGs for the management of DSH; the most recent in 2016.14 The authors, however, selected the 2003 MOH CPG, “Assessment and Management of People at Risk of Suicide” as the ‘gold standard’ for risk assessment, as this was the standard against which ED and Mental Health Services in New Zealand are audited.16 Although the MOH has more recently published “Preventing Suicide; Guidance for emergency departments”, it does not directly address the psychiatric risk assessment.17 The authors were not aware of any pre-existent key performance indicators in the Waikato region for psychiatric risk assessments for community DSH. While clinical practice guidelines exist, international reporting suggests that adherence to them is consistently poor.18–19

A 10-item checklist of measurable, ED-based activities conducted by clinicians was developed from the MOH 2003 CPG (Table 1). Standards such as staff training in suicide, and the recommendation for staff to receive regular supervision identified in the CPG, were excluded. An additional 16-item risk assessment checklist was developed to address item 5, “Mental Health Staff conducted a comprehensive suicide assessment” (Table 2). This incorporated the central tenets of a standard psychiatric risk assessment as outlined in a standard psychiatry text.

In addition to collating de-identified sociodemographic data, three researchers systematically reviewed patient records and benchmarked MH&AS assessments against the two checklists. Statistical analysis was performed using SPSS and Excel. Ethics was sought from the New Zealand MOH’s Health and Disability Ethics Committee (HDEC), but was deemed unnecessary for the audit process or publication of the results.

Results

During the 12-month audited period, 233 females (112.6 per 100,000—figures taken from the New Zealand Census population data 2013)20 and 143 males (72.6 per 100,000) presented to ED (ratio 1.6:1). Patient age ranged from 12 to 84 years (M=30.8, SD 15.3). The highest rate of attempted suicide presentation occurred among 15–24-year-olds, with 179 individuals presented to ED (324.7 per 100,000). Māori had the highest rate of attempted suicide presentation (n=112; 126.7 per 100,000) followed by New Zealand European (n=238; 76.2 per 100,000). People from urban settings were twice as likely to present with attempted suicide as those from rural settings (rates 139.8 per 100,000, and 67.9 per 100,000 respectively).

Table 1: Ten-item checklist assessed by MH&AS clinicians at ED.

Table 2: Sixteen-item risk assessment checklist.

Comprehensive suicide assessment:

‍

Table 3: Level of achievement of checklist items.

‍

Table 4: Level of achievement of Clinical Practice Guidelines.

‍

*68.4% of patients were either admitted or were not on medications requiring close control prescribing.
Note: All n listed above include patients with missing data (ie, not recorded on the assessment checklist).  

Results for adherence by MH&AS clinicians to the psychiatric assessment standards are outlined in Tables 3 and 4. The levels of adherence (recorded as either ‘achieved’ or ‘not achieved’) indicate the percentage of clinicians that adhered to the checklist items. ED clinicians cannot refer patients who are sedated or intoxicated (Item 4, Table 1) for the psychiatric assessment, therefore this item was excluded.

Discussion

The face-to-face interview has traditionally remained an important assessment to help prevent repeat attempted suicide.21 The highest risk factors identified during the clinical assessment are suicide intent and the degree of planning involved in the attempt, including availability and ease of access to suicide means.

While the clinical assessment helps identify high risk factors and provide appropriate, targeted interventions, based on the synthesis of interview and collateral information, systematic literature reviews have failed to identify the positive predictive validity of the psychiatric risk assessment.14 The traditional use of categorical labels, ie, high, moderate and low risks, have to date held poor predictive validity and low inter-rater reliability.22 The psychiatric assessment has also frequently been augmented by other diagnostic tools and rating scales designed to predict short-term suicidal behaviour following a suicide attempt. Likewise, the predictive validity of these various tools in predicting short-term suicidal risk remains poor.23–25 In his study, Horowitz calls for more meticulous exploration of risk; the psychiatric assessment offers this opportunity.25

This study relied solely on recorded information and this limits discussion on the findings and recommendations. The authors acknowledge that discrepancies are likely to exist between actual clinical practices and what was documented in the electronic medical records. However, retrospective formal analysis, as part of a critical review process following a sentinel event (such as suicide) can be aided by contemporaneous documentation, as clinicians’ memories can be fallible.26

Regarding the recorded risk assessment, the clinicians’ practices appeared to have focused on the historical facts of the event: recording details of the attempted suicide episode and determining the presence of other psychiatric conditions. Those areas where clinicians adhered to CPG recommendations appeared to coincide with electronic note categories. The format and requirements of the electronic note record has, and could play a further quality improvement role in CPG adherence.

Clinicians tended not to record extrapolations or judgements about future suicidal behaviours and outcomes. For example, attitude to current and immediate future safety was often missing, as was information about access to further potential DSH methods. Clinicians recorded their interactions with family members (ie, identifying specific concerns or checking their ability to assist in keeping the patient safe) in less than 50% of the cases reviewed. The poorest result observed was the low percentage (recorded in less than 10% of clinical records) of documented invitation for provision of culturally appropriate services for Māori patients as part of their assessment.

More than 75% of patients who were discharged from the ED were referred for outpatient review to occur within a week of the incident with community services. This study did not assess whether this goal for re-assessment was achieved. Standard practice ensured that all general practitioners (GPs) received an automatically generated letter about the risk assessment.

Based on the above findings, we recommend that the psychiatric risk assessment and adherence to an identified CPG requires further additional work. An Australian study into completed suicide found that of the suicides that could have been prevented, 59% had incomplete or poor assessment of suicide risk; suicide risk was not adequately determined and given due consideration.27 The authors would suggest the development of educational workshops to include the following topics:

a) Psychiatric risk assessment for clinicians working in acute mental healthcare settings. Clinicians should be aware of the recommended standards when it comes to suicide risk assessment.

b) Clinicians must comply with medicolegal requirements for documentation in clinical records. Aligning electronic clinical notes with CPG recommendations could improve adherence.

c) The importance of the individualised suicide risk assessment. A diagnostic statement or risk category is insufficient and staff should be encouraged to provide a psychiatric formulation. This should include the judgement of immediate risk for future harm by weighing up the interactions of the multiple identified (dynamic and static) risk factors (for suicide) against the protective and resilience factors. Isolated and even combinations of risk factors indicate little about immediate risk—hence the need for a comprehensive formulation.28

d) Addressing privacy and confidentiality aspects of the clinical assessment especially pertaining to risk. The risk exists that clinicians may be opting for non-disclosure and non-involvement with family and caregivers as the norm rather than exploring how much, and when to safely share information with relatives. It is important for services to balance the need to engage family and caregivers in the assessment and care of patients with suicidal behaviours while ensuring appropriate patient confidentiality.

e) Cultural competency training. Organisations need to ensure that their clinicians have had adequate training in addressing cultural needs of their patients. Many large hospitals in New Zealand employ cultural workers, therefore it is essential that clinicians know how to access these staff members.

The authors of this study concluded that adherence to CPG was moderate to poor. (It remained uncertain whether clinicians performed the CPG activities but had failed to record these in the patients’ clinical records). Despite uncertainties existing about the predictive validity of the comprehensive psychiatric assessment, a comprehensive and individualised assessment remains the cornerstone of clinical assessment and guides the treatment and intervention plans.

‍