Suicide attempts at the time of first admission and during early course schizophrenia: A population based study

Abstract

This article examined suicide attempt rates at first psychiatric hospitalization and risk factors for subsequent suicide attempts over the early course of schizophrenia in national population-based data. Data were extracted from the National Psychiatric Hospitalization Case Registry of the State of Israel that contains all first psychiatric admissions with schizophrenia 1989-1992 and were followed up to 1996 (N = 2293). Attempted suicide rates were: 8.5% (n = 196) at the time of first psychiatric hospitalization and 6.6% (n = 151) over the follow-up period of 4 to 7 years. Of those with a suicide attempt at first admission, 31.6% (n = 62) made a subsequent suicide attempt during the follow-up period (OR = 10.44, 95% CIs = 7.22 to 15.09). Risk profiles were derived using recursive partitioning to predict sub-groups at risk of a subsequent suicide attempt. Those characterized by an attempt at the time of first admission were college educated, female and not married (45.9% (17/37), OR = 13.46, 95% CIs = 6.89 to 26.3). The risk profiles together correctly classified 90.7% (137/151) of subsequent suicide attempts. Suicide attempts at first admission and premorbid years of education have long-term prognostic utility and risk profiles are available.

Introduction

Suicide is the leading cause of premature death among people with schizophrenia (Caldwell & Gottesman, 1990, Drake et al., 1984). Reports based on samples of persons with chronic schizophrenia, meta-analysis and mortality records have estimated that completed suicide over the lifetime of persons with schizophrenia ranges from 6% to 13% (Brown, 1997, Caldwell & Gottesman, 1990, Harkavy-Friedman et al., 1999, Inskip et al., 1998, Meltzer, 2002, Palmer et al., 2005). During early episode, that figure drops to between 3% (Clarke et al., 2006, Kuo et al., 2005, Palmer et al., 2005) and 6% in most (Fenton, 2000, Jarbin & Von Knorring, 2004) but not all (Westermeyer et al., 1991) studies. Suicide risk among patients with schizophrenia and related disorders has been estimated to be 20 times higher than control cases (Nordentoft et al., 2004). Suicide attempts are estimated to occur in among 25% to 50% of people with schizophrenia (Meltzer, 2001, Pompili et al., 2007). Meta-analysis indicates that a prior history of suicide attempts is a risk factor of future suicide (Hawton et al., 2005). Together this highlights the relevance of understanding suicide and suicide attempts in schizophrenia.

The period early in the course of schizophrenia is associated with elevated risk of suicide, particularly among those with a history of suicide attempts (Drake et al., 2006, Harkavy-Friedman et al., 1999, Hawton et al., 2005, Palmer et al., 2005, Scocco et al., 2008, Siris, 2001, Tandon, 2005, Westermeyer et al., 1991). Risk is also higher among those with histories of multiple psychiatric hospital admissions and discharges (Haw et al., 2005, Kuo et al., 2005, Qin, 2005). Indeed psychiatric hospitalization appears to be associated with two peaks of suicidal behaviors, risk that occur before first admission (Addington et al., 2002, Addington et al., 2004, Ho, 2003, Qin, 2005), and shortly after discharge (Craig et al., 2006, Pompili et al., 2005, Roy, 1982). For instance, population-based research has indicated that the highest risk period was after the first admission, during the first 5 days of in-patient care and particularly immediately after discharge (Rossau and Mortensen, 1997). This highlights the need for early episode studies aimed towards suicide prevention programs (Addington et al., 2004, Melle et al., 2006, Power et al., 2003).

Most schizophrenia suicide research focuses on chronic rather than early illness despite the heightened risk associated with the early period of illness (Clarke et al., 2006, Palmer et al., 2005). The follow-up of early episode patients in suicide research is rare, particularly in epidemiological cohorts (Craig et al., 2006, Harkavy-Friedman et al., 1999). One study, the Suffolk County Mental Health Project epidemiological study of early onset psychosis, documents suicide related behaviors in early psychosis. Predictors of suicide ideation and attempts over the 4 years following onset included: prior suicide ideation or attempt, severe depressive symptoms at index admission, lifetime substance abuse, and a younger age of hospitalization (< 28, attempts only) (Bakst et al., 2009). The suicide attempt rate at onset was 9.3% at baseline of whom 40.8% made a subsequent attempt during the 4 year follow-up period (Bakst et al., 2009). This highlights the relevance of elevated risk of a suicide attempt in the years that follow the onset of schizophrenia. Indeed, past case-control research comparing living people with schizophrenia to those who completed suicide in schizophrenia has emphasized the role of suicide attempt as a predictor of completed suicide (Pompili et al., 2009). Also, in an urban catchment study 10% of individuals with early episode psychosis attempted suicide prior to presentation. Four years later, 18% made a suicide attempt and 3% completed suicide. Suicide attempts prior to presentation related to a longer duration of untreated psychosis (Clarke et al., 2006). This research indicates that suicide behavior may be chronic among some attempters who remain alive, possibly reflecting parasuicide. Clinical research into persons with psychosis has used risk profiling to examine whether or not a person made a suicide attempt (Mann et al., 2008). That study showed that, for example, recent attempters were retrospectively characterized by suicide ideation and comorbid borderline personality disorder. We are, however, unaware of national population-based research that prospectively examines the prognostic utility of early suicide attempts during first admission in schizophrenia on subsequent outcomes in an epidemiological cohort over an extensive time period and produces risk profiles.

The current study uniquely aims to examine the correlates of suicide attempts after first hospitalization in a national population based epidemiological cohort. Specifically, we aim to examine (a) incidence rates of suicide attempts at the time of first admission with schizophrenia, (b) the extent to which suicide attempts at onset are associated with a subsequent suicide attempt, and (c) risk factors of suicide behaviors, extending past clinical research (Mann et al., 2008). The current study examines key background factors that have been shown by other research to influence the course of illness, including sex (Caldwell and Gottesman, 1990), immigration and ethnicity (Rabinowitz and Fennig, 2002) and premorbid years of education (Levine and Rabinowitz, 2009).