Tobacco consumption and antidepressant use are associated with the rate of completed suicide in Hungary: An ecological study

Abstract

The suicide rate of Hungary is the highest in the world averaged over the last century but it has shown a very pronounced decrease since 1987. To explore the background of this decrease we investigated the associations between some known suicide-related factors (i.e. tobacco use, antidepressant use and alcohol consumption at the population level) and the suicide rate between 1985 and 2008. The total number of man-hours worked per year by psychiatrists in the outpatient service system and real GDP growth were also monitored in our study. A time series analysis model was constructed to investigate the associations between the above variables and the suicide rate. In the unadjusted model annual tobacco consumption was significantly associated with the suicide rate in a positive manner, while antidepressant use and man-hours were significantly associated with the suicide rate in a negative manner. After adjustment, the associations remained significant only for tobacco consumption and antidepressant use. Neither alcohol consumption nor real GDP growth was associated with the suicide rate in any models. Our results from group-level data confirmed the role of smoking in suicidal behavior previously suggested mainly by studies using individual-level data and also corroborated the results of previous ecological studies concerning the inverse association between antidepressant use and suicide rate. These findings and the results of previous studies – investigating the relationship between smoking and the risk of suicidal behavior at the individual-level – may suggest that programs to prevent tobacco use or to address the widespread recognition and treatment of depression may also prevent suicidality.

Introduction

According to the data of the World Health Organization (WHO), approximately one million people die globally from suicide every year, making suicide the tenth-leading cause of death worldwide. Furthermore in some countries, suicide is among the top three causes of death for young people (Hawton and van Heeringen, 2009, SUPRE, Bertolote and Fleischmann, 2002).

Hungary – a country located in Central Europe – has a “traditionally” high rate of suicide. The national suicide rate began to rise in 1955 and reached a peak in 1980–1986 (≈45/100 000/year). Since then the suicide rate has been consistently falling, reaching a rate of 24.6/100 000/year in the year 2008 (Hungarian Central Statistical Office, 2008, Statistical Yearbook of Hungary, 2009; http://www.who.int/mental_health/media/hung.pdf). The absolute extent of this decrease (≈20/100 000 inhabitants during 20 years) is very unusual. According to the suicide database of the WHO – comprising suicide data of approximately 100 countries – only the suicide rates of three countries (Denmark, Estonia and Latvia) show a similarly marked decrease in the recent decades (http://www.who.int/mental_health/prevention/suicide/country_reports/en/index.html). Although Hungarian suicide mortality has declined by almost 50% since 1986 it is still in the world’s “top ten” (Hungarian Central Statistical Office, 2008).

Suicide is a multicausal behavior in which both environmental and genetic factors are known to be involved (Hawton and van Heeringen, 2009, Rihmer et al., 2007, Döme et al., 2010b, Nock et al., 2008, Mann, 2003, Ernst et al., 2009, Sisask et al., 2010, Giltay et al., 2010, Harris and Barraclough, 1997, Schneider, 2009, Lundin and Hemmingsson, 2009, Chang et al., 2010).

The majority of epidemiological studies – both with prospective or retrospective design and mainly with individual-level databases – have found that smoking is an independent risk factor for suicidal behavior (including completed suicide) both in the general and psychiatric patient populations and that the association between smoking and suicide risk is dose-dependent (accordingly, a few review articles suggested that the inclusion of cigarette smoking in the clinically explorable suicide risk factors is necessary) (Döme et al., 2010a, Rihmer, 2007, Rihmer et al., 2002, Rihmer et al., 2007, Shah and Bhatia, 2010).

Some previous studies have shown that the marked rise in antidepressant prescription in Hungary – which is a proxy marker of better recognition and treatment of depression – was significantly associated with the major decline in the suicide mortality (Sebestyen et al., 2010, Rihmer, 1997, Rihmer, 2001, Viola et al., 2008, Kalmar et al., 2008). The aim of our ecological study was to explore the additional background of the steep drop in the Hungarian completed suicide rate using the time series analysis method. The use of time series analysis using national aggregated data is frequent in suicide research (see for example Chang et al., 2009, Pridemore and Snowden, 2009, Kapusta et al., 2009, Wheeler et al., 2008, Baldessarini et al., 2007, Bramness and Walby, 2009). Interpretation of the results of ecological studies may be limited if the changes of two phenomena are highly correlated at the population level but not correlated at the individual-level (Tu and Ko, 2008). Therefore, we took care to choose variables which were previously shown to be associated with suicide risk at the individual-level (e.g. smoking and alcohol consumption) or atleast where such associations are scientifically plausible at the individual-level (e.g. indicators of the psychiatric health care system and antidepressant consumption) (Harris and Barraclough, 1997, Schneider, 2009, Stone et al., 2009, Bronisch et al., 2008). Furthermore, we selected our independent variables from different fields of suicide risk factors (e.g. indicators of psychiatric services/treatment, economy and substance use) in order to decrease the chance of statistical collinearity between the independent variables.