The burden of physical activity on type 2 diabetes public healthcare expenditures among adults: a retrospective study

Table 3 Logistic regression results to determinate association with public healthcare expenditures in type 2 diabetics (Brazil, 2009).

Independent variable		Expenditure category (higher tertile)
		i	ii	iii	iv	v	vi	vii
		Laboratory tests	Medication	T2DM medication	Other medication	Medical consultation	Nursing consultation	Overall expenditure
		OR_adj1 (95%CI)	OR_adj2 (95%CI)	OR_adj1 (95%CI)*	OR_adj3 (95%CI)	OR_adj2 (95%CI)	OR_adj2 (95%CI)	OR_adj2 (95%CI)
PAL	Sedentary	1.00	1.00	1.00	1.00	1.00	1.00	1.00
	M. Active	0.72 (0.29-1.78)	0.70 (0.28-1.77)	0.71 (0.28-1.77)	0.40 (0.15-1.05)	0.33 (0.25-1.61)	1.02 (0.39-2.67)	0.70 (0.28-1.77)
	Active	0.38 (0.12-1.16)	0.69 (0.25-2.03)	0.72 (0.25-2.03)	0.19 (0.06-0.64)	0.26 (0.08-0.87)	0.87 (0.29-2.62)	0.69 (0.25-2.03)

* = insulin use was significantly associated with T2DM in chi-square test, however, it was not inserted in logistic regression as adjust; OR = odds ratios; OR_adj1 = odds ratio adjusted by gender and skin color; OR_adj2 = odds ratio adjusted by gender, skin color and insulin use; OR_adj3 = odds ratio adjusted by gender, skin color, age and waist-to-hip ratio; 95%CI = 95% confidence interval; T2DM = type 2 diabetes mellitus; PAL = physical activity level; M. Active = moderately active