Does weather affect self poisoning?

Objective: Many studies have attempted to investigate possible connections between the weather and attempted suicide with varying results. The aim of this study was to investigate the possible correlations between the number of admissions to the Welsh National Poisons Unit and local weather conditions between 1989 and 1999. Methods: Admissions data for self-poisoned patients was extracted from the WNPU database. For the purpose of this study the data were recorded as monthly and seasonal totals. The local monthly weather data was obtained from the Met Office for mean temperature, barometric pressure, sunshine amount, cloud cover and humidity. Rainfall data were obtained from the Hadley research centre website (England and Wales total) and the International Station of Meteorological Summary (Cardiff). Average monthly weather data were plotted against monthly admissions to determine any possible correlations. The individual monthly data for each year were also plotted. The same process was carried out for the seasonal admissions against seasonal weather data. Results: The scatter graph for total mean monthly temperature shows a strong positive relationship with admissions (r = + 0.77). However when monthly data were plotted separately for all of the years this correlation disappeared. The same was found for total admissions against average sunshine amount in hours (r = + 0.62). This correlation disappeared when the individual months were plotted. No correlations could be found between admissions and barometric pressure or humidity. The comparison of total figures for cloud cover showed no significant correlation. No patterns could be determined for either of the sets of rainfall data analysed when plotted against monthly admissions. Seasonal associations could be found for temperature (r = 0.959), cloud cover (r = − 0.79), sunshine hours (r = 0.74) and humidity (r = − 0.65) however these associations disappeared when data for individual seasons were compared. No association could be found between seasonal admissions and rainfall or barometric pressure. Conclusion: Significant associations between admissions to the poisons ward and local weather conditions could not be established in Cardiff despite superficial correlations. The implication is that admissions due to poisoning is influenced by other factors, for example social, financial, religious, ethnic or religious. It is also possible that limitations regarding admissions including availability of beds, location with respect to A&E departments and absence of ITU facilities have an influence. A combination of these factors could be great enough to conceal any possible weather influences, or it may be that meteorological conditions indeed have no influence.