Background New Deal for Communities (NDC), a 10-year area-based regeneration initiative begun in 1999 in 39 disadvantaged neighbourhoods in England, has the potential to reduce health inequalities because it focuses on key social determinants of these inequalities: unemployment, crime, education, housing and the physical environment. This study assesses whether the NDC initiative impacted on health inequalities across the socioeconomic spectrum of areas in England.

Methods The study primarily utilises secondary data including household surveys undertaken as part of a national evaluation of the NDC initiative led by Sheffield University. Representative samples of 500 residents per NDC area and matched comparator area in 2002, 2004, 2006 and 2008 provide data on health, lifestyles, wellbeing, demography and social determinants of health. The Health Survey for England in the same years provides information for representative samples drawn from across the socioeconomic spectrum, here categorised as highest, middle and lowest tertile of area deprivation. Logistic regression was used to estimate baseline differences in health and its social determinants, the time trend and differential time trends in NDC and non-intervention areas adjusted for sex, age, ethnicity and baseline education based on over 125,000 observations across eight years. Outcomes of interest included self-rated health, mental health inventory, smoking, employment status and housing tenure.

Results Initial analyses indicated a higher likelihood of smoking in non-intervention areas of medium deprivation (OR; 95%CI: 1.33; 1.10, 1.60), non-intervention areas of high deprivation (1.80; 1.48, 2.19), matched comparator areas (1.77; 1.46, 2.15) and NDC areas (1.97; 1.68, 2.31) compared with reference areas of low deprivation. Smoking declined over time but there was no evidence of differential rates of decline in the different areas. The likelihood of unemployment was considerably higher in all areas compared with the reference (for example, 4.29; 3.55, 5.18 in NDC areas). The likelihood of unemployment dropped over time in NDC areas but not in other areas (OR for time by NDC interaction: 0.93; 0.89, 0.96). Results for other outcomes are also available.

Conclusion The study indicates that the NDC intervention may have contributed to reducing a key social determinant of health inequalities, namely unemployment. Similar reductions were not observed in non-intervention high, medium or low deprivation areas in England which may eventually feed through into a narrowing of health inequalities between NCD and other areas. This initial work will be extended to examine impact across different intervention approaches and historical and current contexts.