Seasonality and suicide in Italy: Amplitude is positively related to suicide rates

Abstract

Background

Recent studies have reported changes in the time patterns of suicide, with conflicting findings regarding the direction of these changes: data from Italy were investigated to evaluate the influence of recent social and medicine-related changes on the seasonality of suicides in the country.

Methods

A total of 71 227 male suicides and 26 466 female suicides occurring in Italy from 1974 to 2003 were investigated with harmonic spectral analysis to extract their monthly seasonal dispersion by five-year intervals.

Results

The suicide rates of both males and females showed a rising trend, with an evident peak in the 1987–1994 period and a decrease thereafter. Seasonality of suicides, with a clear peak in spring as against the other seasons, accounted for a statistically significant proportion of total variance: around 40% among males and 39% among females. Seasonality did not change across time in a relevant way; however, an anticipation of the peak was observed in both males and females over time, with amplitude increasing or decreasing as a function of yearly suicide rates.

Limitations

Data could not be analysed according to age or to the method of suicide, since this information was not available across the whole time interval.

Conclusion

The seasonal effect on mortality by suicide is positively related to suicide rates, so much that changes in suicide rates over time correspond to changes in suicide seasonality.

Introduction

Longitudinal data provided some evidence that the seasonal effect on suicide mortality has decreased in some countries in the latest decades (Rihmer et al., 1998, Hakko et al., 1998, Jessen et al., 1998, Yip et al., 2000). As a matter of fact, studies examining very long-term time series of suicides found that the decrease in the seasonality of suicides is an ongoing process that started at the end of the 19th century (Ajdacic-Gross et al., 2005). Socio-economic and medicine-related variables may have contributed to the modification of the long-term seasonality of suicides. Oravecz and co-workers (2003), for example, found a change in the seasonal dispersion of suicides in Slovenia after 1993, when a significant increase of antidepressant sales began: the steep spring peak observed before 1993 decreased substantially, and a new autumn peak emerged (Oravecz et al., 2003, Oravecz et al., 2006).

The improvement in the recognition and treatment of mental disorders, and of affective disorders in particular, which followed the spreading of well-grounded diagnostic criteria and the availability of more effective and tolerated therapeutic compounds, is likely to have contributed to decreasing the seasonal amplitude reported by some European countries in recent years (Rihmer, 2001). However, conflicting data emerged on the fading of suicide seasonality, and in some countries a rising trend in the seasonality of suicides was reported. Using data from the Australian Bureau of Statistics on all suicides in Australia 1970–1999, Rock and co-workers (2003), for example, found that seasonal amplitude had increased in that country over time, particularly for violent suicide and suicide by males. The seasonality of suicides in the USA as well seems to have increased from the 1970s to the 1990s (Bridges et al., 2005). Negative findings, i.e. no statistically significant changes in suicide seasonal amplitude, were reported in Japan (Nakaji et al., 2004) and in Romania (Voracek et al., 2002).

Until now, no study had investigated the influence of recent changes in socio-economic and medicine-related factors on the seasonality of suicide in Italy. This study aimed at ascertaining whether the so-called seasonal effect on suicide mortality had decreased in recent years in Italy, taking into account six five-year time series: the first two (1974–1983) served as a base rate; the third and the fourth (1984–1993) saw the transformation of psychiatric assistance in Italy following the psychiatric services reorganization brought about by the so-called “Basaglia” reform, which greatly increased access to care for mentally disordered patients and was accompanied by the introduction of well-grounded criteria for diagnosis (de Girolamo and Cozza, 2000, Preti and Miotto, 2000); the last two time series, from 1994 to 2003, are the case referent term, since those years were accompanied by the introduction of second- and third-generation pharmacologic treatments for mental disorders, including SSRIs and atypical neuroleptics (Barbui et al., 2002, Barbui et al., 1999). On the basis of past results, we expected a decrease in the seasonal effect on suicide mortality in the most recent time series if the improvement in the recognition and treatment of mental disorders is a truly critical factor contributing to the decrease in seasonal amplitude, as reported in some European countries in recent years.