Do monthly or seasonal variations exist in suicides in a high-risk setting?

Abstract 

An unequal distribution of suicides over months and seasons has been a consistent finding in epidemiological surveys on suicide. Jails and prisons are a high-risk setting for suicide all over the world. The high prevalence of both outward and self-directed violence in prison populations indicates dysfunctional central serotonin (5-HT) neurotransmission and, therefore, could account for an unequal distribution of suicides over months and seasons due to underlying bioclimatic factors. Within a total survey of suicides in the Austrian penitentiary system, the weekly, monthly and seasonal distribution of custodial suicides between 1947 and 1999 was studied. After an explorative comparison of suicide distribution over weekdays, months and seasons of the year by χ²-tests, a harmonic Poisson regression model was performed to detect seasonality of suicides. No unequal distribution of suicides was evident over the 53-year period. A limitation of this study was its sample size of 412, a low number compared with population-based samples, where a spring suicide peak was consistently found. An explanation for lacking seasonality could be that bioclimatic factors are less relevant in urban, industrialized areas, where jails and prisons usually are located. One of the core characteristics of penal institutions is the limited possibility for communication and social interaction. This social isolation is independent of seasonal changes. If the individual's possibilities for social interactions are limited, the influence of seasonal changes in social activities may be less relevant. This could explain the absence of seasonal changes in custodial suicide incidence.

Keywords:  Prison suicide, Seasonality, Circannual rhythm, Harmonic Poisson regression