Spatial analysis of dietary cost patterns and implications for health

Background: Chronic diseases such as obesity, cardiovascular disease, diabetes and cancer are
a large burden on society, for which diet is the leading modifiable risk factor. ‘Diet’ can
encompass a variety of aspects of food consumption. Dietary patterns arguably best reflect
food as consumed. The determinants of diet are varied and include cost and availability of
foods. The aim of this research is to investigate if dietary patterns and diet cost vary spatially
and whether this influences health outcomes, specifically obesity and breast cancer.
Methods: Using data from the UKWomen’s Cohort Study, data driven dietary patterns were
previously determined. Cost of diet was assigned using a food cost database. Spatial measures
for Government Office Region, North South, Urban/ Rural and the Output Area Classification
were assigned matched to postcode of the women.Weight status is calculated from self
reported height and weight. Time to event analysis investigated association between dietary
pattern, diet cost and breast cancer incidence at each spatial scale.
Results: There is some spatial variation in dietary patterns both between and within regions. A
healthy in more expensive per day than a less healthy one: £6.63 compared to £3.29. The
overweight/ obese pay more for their food. Urban and Northern areas have significantly higher
BMI than Rural and Southern areas respectively. Those in areas Constrained by Circumstance
have highest prevalence of overweight and obesity. There is some spatial variation in breast
cancer incidence and variation by dietary pattern. In postmenopausal women, positive
association exists between weight status and risk of breast cancer incidence. Methods used for
estimating small area dietary patterns and health outcomes may be applicable for use in other
developed populations.
Conclusion: Understanding determinants of dietary patterns remains important for public
health and making healthy diets accessible to all is important. However, while expensive
dietary patterns reflect a healthier diet, they do not appear to be the mechanism for which
obesity prevalence and breast cancer incidence occur. Geodemographic classifications,
combined with other spatial measures could aid more effective targeting of public health
nutrition policy.