A pair of researchers in the United Kingdom suggest that suicide assessments can be more reliable if clinicians focus on the need for additional assessment and the presence of risk factors associated with initial, continuing, re-occurring, and escalating suicidal behavior rather than rating risk using ambiguous terms (such as low, medium, and high) and failing to distinguish between types of suicide events for which the patient may be at risk.
Social workers and clinical psychologists from community mental health teams were given reports about hypothetical patients referred by general practitioners after an incident of self-harm. Each clinician assessed the patients for suicide risk (using the categories of high, medium, and low) as well as the probability that the patients would engage in self-harm within the next six months. They also provided a recommendation of the need for additional assessment. The cases were defined (and the assessments analyzed) on the basis of 10 “cues” included in the referral letters that professional guidelines agree influence the risk of self-harm. These cues were: 1) sex (male or female), 2) age, 3) substance abuse, 4) a precipitating event (trigger), 5) social support, 6) a major adverse life event, 7) current mental state (hopeful or hopeless), 8) a suicide plan, 9) how well the patient managed a previous episode of self-harm, and 10) the seriousness of that episode.
The clinicians’ assessments of suicide risk and need for further assessment were “consistent with practice guidelines.” Each clinician was generally consistent in his or her assessments, although more experienced providers were more consistent than younger clinicians. There was less agreement among assessments by the clinicians, especially in the assessments of patients whose risk was judged to be low. The research team concluded that these differences reflected differences in defining ‘low,’ ‘medium,’ and ‘high’ risk as well as how the clinicians interpreted the term “risk.” They theorized that some clinicians think of risk as a measure of probability (that is, how likely is a person to harm him- or herself in the future) while others interpret risk as a measure of probability and a measure of the severity of the potential self-harm. The group that includes severity in risk is more likely to prioritize additional assessment for patients judged to be at risk, since they anticipate that the risk will have severe consequences.
The authors suggest that better definitions of risk would reduce these inconsistencies as well as inconsistencies caused by differing interpretations of the impact of some risk factors. For example, despite the fact that research has demonstrated that major adverse life events raise suicide risk, clinicians assessed patients who had undergone such an event to be less at risk, and less in need of additional assessment, than others. The authors speculate that the clinicians believed that the patients who had engaged in self-harm during an adverse life event would be less at risk in the future as the effects of the event dissipated. They suggest that it is important to use appropriate criteria to assess the probability of initial, continuing, re-occurring, or escalating suicidal behavior (given that each is associated with different risk factors).
Combining results from two or more participants to simulate assessment teams did not significantly improve overall assessment results. Nor were the assessments produced by these teams more reliable than the most reliable individual assessors (that is, the assessors who were most consistent in applying criteria for assessing risk and prioritizing follow-up). However, any combination of participants tended to be more consistent than the least reliable individual clinicians. The authors suggest that this finding provides support for using teams to assess suicide risk.