The majority of patients declared their household income per annum (204/218; 93.5%): 71/204 (34.8%) patients were fully subsidised, 89/204 (43.6%) earned a yearly household income of less than USD 6600,  9/204 (4.4%) earned a yearly household income of between USD 6600 and USD 23000 and 35/204 (17.2%) earned in excess of USD 23000 per year. Univariate analysis did not demonstrate SES as being a significant determinant of outcome (p=0.05) but multivariate analysis revealed that patients who came from homes where household income exceeded USD 23000 per year had a significantly lower risk of death (HR 0.071; p=0.039) (TABLE 3). We also demonstrated a significant negative correlation between SES and stage (Spearman rho = -0.153; p=0.03), in that patients with higher SES presented with lower stage tumours. However, in the cox model, we adjusted for both SES and stage, revealing that the significant impact of SES on survival was in excess of any effect that could be explained by lower stage disease at diagnosis alone.