Results 30 papers fulfilled inclusion criteria. There were considerable methodological differences, making direct comparison problematic and precluding meta-analysis of rates. A clear robust finding across all studies is that fall rates for both men and women increase with age. Typically age related odds ratios appear modest (1.03–1.05/annum) but this implies quite substantial changes in risk over 5 year period. Prevalence of falling is generally higher among women than men but evidence describing the relationship between socio-economic status and falls is scant. Using ecological measures the most robust studies suggest higher rates among the most deprived. Ethnicity is confounded with socio-economic and lifestyle differences (eg, rural vs urban dwelling groups). Nonetheless there are broad general trends to be found. There are differences in falls rates between ethnic groups living in the same societies for example, USA. However, it is less clear how ethnicity per se relates to falls. Caucasians living in USA have higher falls injury admission rates, than other groups but this may relate to access to services.

Conclusions It is clear that we know little about the relationship between socio-demography and falls. Future studies need to use standardised methods to permit synthesis.