Temporal pattern of suicide risk in young individuals with early psychosis

Abstract

Individuals with a first episode of psychotic illness are known to be at high risk of suicide, yet little is understood about the timing of risk in this critical period. The present study aimed to examine the temporal pattern of suicide risk in patients with early psychosis (EP) and to determine whether discrete periods of significantly elevated risk can be identified up to 24 months after commencing treatment. Suicidality ratings collected each month as part of patient routine assessment at the Early Psychosis Prevention and Intervention Centre (EPPIC) were retrieved from the service database for patients treated between December 2002 and December 2005 (N = 696). Time-series analysis was performed on suicide risk estimated from the aggregated data of 94 individuals who met the study inclusion criteria. Suicide risk was highest in the first month of treatment, decreasing rapidly over the next 6 months and declining slightly thereafter. A power function adequately described this curvilinear trend. Fluctuations around the trend were unpredictable, except for a mild tendency to reverse from month to month, and did not reach statistical significance. The findings suggest limited scope for preventative interventions driven by chronology alone. Intensive routine suicide screening across the course of treatment may facilitate identification and early management of EP patients at suicide risk.

Introduction

Suicide is a major cause of premature death in individuals with psychotic disorders, particularly schizophrenia (Brown, 1997). While it is estimated that approximately 5% of schizophrenic patients commit suicide during their lifetime (Palmer et al., 2005), suicide rates near illness onset are three times higher than in the chronic stages of the illness (Mortensen and Juel, 1993, Brown, 1997). Among patients with early psychosis (EP), the rates of other suicidal behaviours are even higher, with 10%–30% attempting suicide prior to the first treatment contact for psychotic symptoms (Nordentoft et al., 2002, Clarke et al., 2006, Melle et al., 2006, Bertelsen et al., 2007). The prevalence of suicide attempts at a 1- or 2-year follow-up falls between 3% and 22% (Nordentoft et al., 2002, Addington et al., 2004, Petersen et al., 2005, Bertelsen et al., 2007), or 11%–18%, respectively (Verdoux et al., 2001, Petersen et al., 2005, Bertelsen et al., 2007). In addition, a large proportion (20%–51%) of EP patients seriously consider suicide while in treatment (Petersen et al., 2005, Bertelsen et al., 2007). Given that suicidal ideation and a history of suicide attempt are significant predictors of future attempts (Tarrier et al., 2004, Shoval et al., 2006) and eventual suicide in psychotic illness (Roy, 1982, Stephens et al., 1999, Sinclair et al., 2004), these figures suggest that management of suicide risk is an important challenge for psychiatric services treating individuals with EP (Nordentoft et al., 2008).

The knowledge of when EP patients are most likely to pose a risk to themselves during treatment is clinical information that would be useful for prediction, intervention and prevention of suicide attempts. Close monitoring of inpatients and more intensive outpatient case management could be provided during these periods of elevated suicide potential even if patients do not appear to be at immediate risk. However, currently little is known about the timing of suicide risk during the early stages of psychotic illness.

Temporal variability in psychopathology is best examined using repeated measures collected longitudinally at regular intervals over the period of interest. Statistical procedures, such as time-series analysis, enable identification of meaningful patterns in the data by separating reliable changes from inherent variability (Morley, 1996). The only published study reporting on the temporal pattern of suicide risk in EP during the first 20 months of treatment (Power et al., 2003) found that the risk was highest in month 1 and 5, and, again, between month 10 and 16. However, the findings of this investigation need to be interpreted with caution due to several methodological limitations: the study used a cross-sectional design; speculations were made entirely on the basis of inspection of descriptive data, and no controls were introduced to account for attrition rates. If the existence of discrete periods of increased risk in this already high risk population can be confirmed in a longitudinal study using statistical methods recommended for this type of data, the findings would have important implications for clinical practice.

The present study examined the course of suicide risk in EP patients whose level of suicidality was repeatedly evaluated throughout treatment as part of routine assessment protocols. The aim of the study was twofold: firstly, to investigate the temporal pattern of suicide risk over the 24-month follow-up and, secondly, to determine whether there are discrete periods of significantly elevated suicide risk using time-series analysis.