Research report
Seasonal changes in 24-h patterns of suicide rates: a study on train suicides in The Netherlands

https://doi.org/10.1016/S0165-0327(00)00308-6Get rights and content

Abstract

Background: Annual patterns in suicide rates, peaking near the summer solstice, are well documented. It has been suggested that day length or total hours of sunshine has an impact on suicide rates. If these environmental factors are involved, we would expect changes in the daily pattern of suicide rates to occur over the year. To test this hypothesis, the 24-h patterns of suicide rate were investigated as a function of time of year. Method: Detailed information about the exact time of suicides in The Netherlands is only available for train suicides. Therefore, information concerning age, sex, time and place of occurrence of all verified train suicides over 15 years in The Netherlands (n=2830) was obtained from The Netherlands Railways archives. Results: Daily patterns in train suicides show systematic variations of two kinds. First, independently of time of year, suicide rates at night drop to about 10% of their daytime values. Second, there are two daily peaks in the patterns which shift their timing over the year, with one peak occurring shortly after sunset, and the other one consistently occurring 9–10 h earlier. Both peaks shift with the 5.5-h shift in sunset time. Limitations: Train suicidal behaviour may not represent fatal suicidal behaviour in general. Conclusions: There are pronounced and systematic daily variations in train suicide rates in The Netherlands. One of these is related to clock time, the others are related to the light–dark cycle. The consistency of the patterns suggests a strong environmental influence on train suicidal behaviour. Research on 24-h patterns of suicide rates should control for time of year.

Introduction

Seasonality of suicidal behaviour has been an object of study since the previous century. The observed peaks in suicide rates near the summer solstice have been a matter of interest and debate for over 150 years (Kevan, 1980; Aschoff, 1981; Massing and Angermeyer, 1985). It has been suggested that variations of psychosocial stress over time of day and season may account for some of the temporal variations in suicide rates (Durkheim, 1897). More recently, the idea has been put forward of temporal variations in susceptibility to psychosocial stressors. This variation in susceptibility might be attributable to changes in central 5-HT (serotonin) neurotransmission (Maes et al., 1993). One argument in favour of this hypothesis is the observation of a significant temporal relationship between plasma l-tryptophan (a precursor of 5-HT) levels in healthy subjects and rates of violent suicide (Maes et al., 1995).

Whatever the mechanism by which the seasonal pattern of suicide rates is caused, there is a significant correlation between day length and suicide rate, and between hours of sunshine and suicide rate (Aschoff, 1981). If day length somehow would be causally involved in suicidal behaviour, it would be expected that the 24-h patterns of suicide rates would vary in the course of the year. Such circadian variations of suicidal behaviour (i.e., variations over the 24-h day) have received little attention. Uncertainty about the exact timing of many suicides complicates research in this field (Barraclough, 1976). Both Durkheim (1897) and Morselli (1882) noted that suicide is largely restricted to the daytime hours. These observations are confirmed by more recent major studies (Williams and Tansella, 1987; Maldonado and Kraus, 1991; O’Donnell and Farmer, 1994; Schmidtke, 1994) showing the lowest suicide rates between midnight and 06:00 h. However the temporal distribution outside this nocturnal period is unclear. Long intervals of three or more hours, chosen by most authors as the unit of temporal analysis, preclude a detailed analysis of 24-h patterns. Only two studies investigated the possible interaction between daily patterns of suicide rate and season (Schmidtke, 1994; Altamura et al., 1999). In the most recent study (Altamura et al., 1999) regarding different, but predominantly violent, types of suicide in Italy (39° N), no significant interactions were found between time of day of suicide and the quarters of the year. In the other study (Schmidtke, 1994), carried out in Germany (47°30′–54° N), the 24-h patterns of train suicide rates in the summer and winter season were compared. The peak of train suicides in males, was found to shift from 18:00–21:00 h in winter to 21:00–00:00 h in summer.

The present study focuses on train suicide as well. The advantage of investigating this type of suicide is that the exact moment of occurrence is known. Specific legal procedures connected with every train accident, i.e., collision with object, ensure detailed recording of circumstances and time. As a consequence train suicides are the best documented suicides in terms of time in The Netherlands and a number of other countries. In The Netherlands, train suicides accounted for 10–14% of all suicides in the period from 1980 to 1994. These percentages are higher than those reported in other European countries, like Germany and Austria, where they are 5.1 and 5.7 (Schmidtke, 1994; Deisenhammer et al., 1997), respectively. Still these percentages are small. Therefore it remains to be investigated whether the conclusions of the present study can be generalized to suicides in general.

Here, we specifically address the question of whether 24-h patterns of train suicide rates can be identified, and if so, whether they change with time of year. The results will be discussed in the context of possible environmental and biological factors that might be involved in processes underlying the suicidal behaviour under study.

Section snippets

Method

The Netherlands is a small, densely populated country with approximately 15 million inhabitants. It is located between 51 and 53°30′ N, in a temperate climate zone where the length of day varies considerably over the seasons. At this latitude sunrise varies from 08:46 h (Central European Time) in mid-winter to 05:18 h (Central European Summer Time) at the summer solstice. Sunset varies from 16:30 to 22:03 h. So the length of the day varies from 7 h and 44 min to 16 h and 45 min.

In the period

Results

The annual pattern of train suicides in The Netherlands, accumulated over 15 years, is presented in Fig. 1. Averaging the results and dividing this by 15 yields a mean of 14.5 cases per 4 weeks. Obviously, the data do not show the expected peak near the summer solstice. Poisson regression analysis indicates that there is no significant time trend across the months of the year (χ2=17; df=11; P=0.11).

When we analyse the same data relative to clock time, a pronounced 24-h pattern emerges (Fig. 2).

Discussion

In spite of the fact that the data on train suicides in The Netherlands, collected over 15 years, do not demonstrate a pronounced seasonal dependence when plotted as a function of time of year, nor show a pronounced diurnal pattern when plotted as a function of clock time, the data do reveal a strong and significant seasonal modulation of diurnal suicide patterns when plotted two-dimensionally. The diurnal distribution changes over the year in such a way, that peaks in summer occur at clock

Acknowledgments

We thank the N.V. Nederlandse Spoorwegen for kindly providing access to the data used in this study. We thank Professor Dr. R.H. van den Hoofdakker and Professor Dr. J.C. Gillin for commenting on a previous version of the manuscript. We thank Dr. J. Neeleman for his statistical advice. We are also grateful to A.M. Steensma for setting up the database and performing preliminary analyses.

References (21)

There are more references available in the full text version of this article.

Cited by (67)

  • The situational conditions of suicide in transit environments: An analysis using CCTV footage

    2021, Journal of Transport and Health
    Citation Excerpt :

    Suicides do not happen at random in time. In terms of temporal patterns, Dinkel et al. (2011) found that the incidence of jumping from the train station platform was more common during 9:01 a.m. to 6:00 p.m., and at night, lying on the tracks was most frequent. van Houwelingen and Beersma (2001) found that in the Netherlands suicide levels were lower at night-time than day-time.

  • Injury severity analysis of pedestrian and bicyclist trespassing crashes at non-crossings: A hybrid predictive text analytics and heterogeneity-based statistical modeling approach

    2021, Accident Analysis and Prevention
    Citation Excerpt :

    Generally, fatalities among trespassers are observed to exhibit geographic and temporal clustering (Pelletier, 1997; Savage, 2007). Interestingly, independent of time of year, suicide rates at night dropped by 10 % while the suicide rates were almost constant over daytime hours (Van Houwelingen and Beersma, 2001). Train-involved trespassing crashes are more likely to result in a fatality.

  • The role of seasonality and photoperiod on the lethality of suicide attempts: A case-control study

    2019, Journal of Affective Disorders
    Citation Excerpt :

    This is consistent with previous studies reporting more season- and circadian- related fluctuations of suicide rates among patients with mood disorders, in comparison to those with schizophrenia (Valtonen et al., 2006). Our findings confirm previous evidence showing that suicidal behaviors exhibit both circadian (van Houwelingen and Beersma, 2001; van Houwelingen et al., 2010) and seasonal rhythms (Germain and Kupfer, 2008). Both in attempted (Valtonen et al., 2006) and completed suicides (van Houwelingen and Beersma, 2001) time patterns were detected.

  • The androgen model of suicide completion

    2019, Progress in Neurobiology
  • Characteristics of railway suicides in Canada and comparison with accidental railway fatalities: Implications for prevention

    2017, Safety Science
    Citation Excerpt :

    Analyses using these categories do not provide information on the differences between accidents and suicides (Evans, 2011; Lerer and Matzopoulos, 1996; Ozdogan et al., 2006; Pelletier, 1997). The second approach has been to analyse only fatalities that are classified as suicides (Hudson, 1999; Baumert et al., 2006; Deisenhammer et al., 1997; Van Houwelingen and Beersma, 2001; Van Houwelingen et al., 2010, 2013; Erazo et al., 2004a). The third approach is to report on all railway fatalities and provide comparison data on environmental and personal variables for both accidents and suicides (Lin and Gill, 2009; Radbo et al., 2005; Rådbo et al., 2012; Silla and Kallberg, 2012; Gershon et al., 2008).

View all citing articles on Scopus
View full text