Background Although long-term trends in local labour market conditions are likely to influence health, few studies have assessed whether this is so. This paper examines whether (1) trends in local employment rates have relevance for mortality and morbidity outcomes in England and (2) trends are stronger predictors of these outcomes than employment rates measured at one point in time.

Methods Using latent class growth models, local areas were classified into eight groups following distinct trends in employment rates between 1981 and 2008. Areas were also categorised in ‘octile’ groups by rank of employment rates in 2001. These area groupings were linked to a sample of 207 959 individuals from the Office of National Statistics Longitudinal Study. Associations between area groupings and risk of all-cause mortality and of reporting a limiting long-term illness at the end of the period were measured using logistic regression. Models were adjusted for individuals' socio-demographic characteristics measured in 1981 and for their residential mobility between 1981 and 2001.

Results Compared to areas with continuously high employment rates over the period, risk of mortality and morbidity was higher in areas with persistently low or declining employment rates. Findings suggest that long-term trends in local employment rates are useful as predictors of mortality and morbidity differences. These are not so clearly distinguished by only considering employment rates at one point in time.

Conclusion Poor health outcomes are associated with long-term economic disadvantage in some areas of England, reflected in employment rates, underlining the importance of efforts to improve health in areas with especially ‘deep-seated’ deprivation.