Usefulness of the Short Personality and Life Event Scale (S-PLE) for detection of suicide attempters

Dear Editor,

Suicide is a worldwide public health problem, and it is the primary cause of non-natural death in our country.¹ The genesis of suicidal behaviour is influenced by multiple biological and social factors that often, as is the case with the economic recession, may have effects which are hard to determine.² To date previous attempted suicide (AS) has been considered to be the best predictor of the risk of suicide.³ We therefore consider that all efforts to predict and prevent AS to be an indirect way of contributing to a reduction of suicides.

One of the most interesting efforts aimed at identifying individuals at risk of suicidal behaviour was undertaken by Blasco-Fontecilla et al.⁴ They developed the Personality and Life Event scale (PLE) which includes 27 of the most discriminatory items from a series of questionnaires that are commonly used to evaluate suicidal behaviour (personality disorders scale, impulsiveness, aggressiveness, stressful life events and sociodemographic data) with excellent results in terms of sensitivity (80.8%) and specificity (89.6%). The authors developed this further with the brief version of the said scale (S-PLE),⁵ which with only 6 items makes it possible to indirectly and non-intrusively evaluate the risk of suicidal behaviour in situations where time is lacking.

The authors themselves propose the following cut-off points to maximise the precision of the classification: (i) healthy individuals, scores lower than 1.70; (ii) individuals with a possible mental disorder, scores from 1.70 to 2.46, and (iii) individuals at risk of suicide, scores higher than 2.46. Nevertheless, the authors accept that the S-PLE performs less well in differentiating individuals with a mental disorder and no history of AS and patients with a history of AS, although the area under the curve (AUC) of the receptor operative characteristic (ROC) remains acceptable (0.756).

Our group tried to replicate these earlier results in an independent sample of 197 patients [35.5% men; average age (SD) = 54.15 (10.54) years old], diagnosed with mood disorder [unipolar depression (74.6%); bipolar depression (8.1%) and dysthymia (17.3%)] with clinical severity of depression at the time of evaluation measured using the Hamilton Depression Rating Scale (HDRS)⁶ of 18.56 (5.95), which is equivalent to moderate to severe depression. 38.6% (n = 76) of the patients had a history of AS. The patients with a history of AS were significantly younger [51.79 (10.70) vs 55.64 (10.21); Student t-test = 2.526; p = .012] and they scored significantly higher in the S-PLE [2.10 (0.48) vs 1.79 (0.46); Student t-test = −4.424; p = .000], while they were similar in terms of sex and their average score on the HDRS.

When the precision of the scale was evaluated by means of ROC analysis, an AUC of 0.675 was obtained, which is lower than the AUC figure reported by the authors. When the cut-off point they proposed to differentiate individuals with a mental disorder without a history of AS from those with a history of AS/(scores higher than 2.46), we found a sensitivity of 17.10% and specificity of 95.90%, i.e., there are a high number of false negatives. If it is wished to use this scale as a screening tool, we believe it would be prudent to suggest a modification of the cut-off point to reduce the large number of false negatives, at the least when the aim is to detect the risk of suicide in patients with mental disorders. After evaluating our results we suggest that a more suitable cut-off point for the evaluation of the risk of suicide would be 1.70. With this cut-off point, at least in our sample, a major increase in sensitivity is achieved (85.5%), although specificity falls by 32.2%. Nevertheless, given the severity of the consequences of the event we wish to predict and the possibility of using the S-PLE in screening for the risk of suicide in situations where time is lacking, such as hospital emergency or primary care departments, we consider it to be clear that good sensitivity has to take priority over specificity.

We believe that the S-PLE may be a useful clinical instrument for detecting the risk of AS, although it is necessary to set a more precise cut-off point for the tool and determine its predictive capacity in prospective studies such as those currently underway in our research group.

Acknowledgements

This study was financed by the Spanish Government fund “Fondo de Investigación Sanitaria” FEDER (PI14/02029) and the Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM).