Large-scale and high-resolution analysis of food purchases and health outcomes

EPJ Data Science

Table 2 Linear regressions that predict the three chronic diseases from item weight and nutrient diversity (plus control variables such as income, gender, age, education level)

Feature	Coefficient	Std. error	p-value
Hypertension
α (intercept)	0.5582	0.064	<0.001
Calorie consumption	0.30228	0.070	<0.001
Nutrient diversity	−0.5182	0.069	<0.001
Income	0.0615	0.041	0.131
%Females	−0.2210	0.038	<0.001
Average age	0.1627	0.037	<0.001
Education	−0.2309	0.041	<0.001
Durbin–Watson stat. = 2.033		Adj \(R^{2} = 0.377\)
Cholesterol
α (intercept)	0.5465	0.059	<0.001
Calorie consumption	0.1395	0.064	0.03
Nutrient diversity	−0.4943	0.064	<0.001
Income	0.0017	0.037	0.96
%Females	−0.2364	0.035	<0.001
Average age	0.11790	0.034	0.001
Education	−0.1785	0.037	<0.001
Durbin–Watson stat. = 1.981		Adj \(R^{2} = 0.344\)
Diabetes
α (intercept)	0.7582	0.038	<0.001
Calorie concentration	0.1301	0.028	<0.001
Nutrient diversity	−0.6353	0.043	<0.001
Income	−0.0790	0.034	0.019
%Females	−0.3693	0.031	<0.001
Average age	−0.0606	0.030	0.042
Education	0.1047	0.032	0.001
Durbin–Watson stat. = 1.964		Adj \(R^{2} = 0.585\)