Objective: To investigate the effect of material deprivation on the winter rise in mortality and temperature dependent variations in mortality.
Design: Ecological comparison of seasonal mortality at electoral ward level. Main outcome measures were ratios of winter to rest of the year mortality rates (seasonality ratios) and monthly deaths as the outcome variable in a model with monthly average temperature and Townsend score as main predictors.
Setting: Croydon, London, United Kingdom.
Subjects: All deaths of Croydon residents for the period 1990-1995.
Main results: There was a clear relation between overall mortality and deprivation. There was no evidence of a relation between age and sex standardised seasonality ratios and Townsend scores for all deaths (Kendall's tau = -0.066, p = 0.63) or cardiovascular deaths or respiratory deaths. There was no evidence of an interaction between Townsend score and temperature in the model of ward mortality rates (p = 0.73). These findings were not affected by exclusion of deaths of nursing and residential home residents.
Conclusion: This study provides no evidence of an effect of deprivation on excess winter mortality or temperature dependent variations in mortality. The findings question simple assumptions about the relation between deprivation and excess winter mortality and highlight the need for further study to guide interventions.