Table 4.

Intervention effects on ∆sales of vegetables and fruits

	Early intervention		Late intervention
	Estimates	95% CI	Estimates	95% CI
Vegetables
Total	10.2	−7.2 to 27.7	19.6	2.2 to 37.1*
Adjusted	9.8	−7.3 to 27.0	18.7	1.6 to 35.9*
Green leafy	16.7	−3.9 to 37.3	20.2	−0.3 to 40.8
Adjusted	16.6	−3.8 to 37.1	18.6	−1.8 to 39.0
Leafy	12.8	−8.7 to 34.2	21.4	−0.1 to 42.8
Adjusted	12.8	−8.5 to 34.1	21.4	0.1 to 42.7*
Fruit and flower	−16.1	−39.4 to 7.3	34.1	10.7 to 57.4†
Adjusted	−16.5	−39.6 to 6.7	33.0	9.9 to 56.2†
Root	9.1	−16.9 to 35.2	3.4	−22.6 to 29.4
Adjusted	8.6	−16.9 to 34.0	2.4	−23.1 to 27.8
Juice	30.3	−20.8 to 81.5	38.0	−13.2 to 89.1
Adjusted	28.9	−21.9 to 79.6	36.4	−14.4 to 87.1
Other	15.5	−1.9 to 32.9	4.4	−13.0 to 21.7
Adjusted	14.4	−2.7 to 31.5	3.0	−14.2 to 20.1
Fruits	−10.7	−28.9 to 7.4	0.1	−18.1 to 18.2
Adjusted	−11.4	−28.9 to 6.1	−0.7	−18.2 to 16.9

The interaction estimate between the intervention status and study periods on ∆sales in a linear regression model was assessed as an intervention effect. The estimate for each intervention period is equal to (mean ∆sales in the given intervention stage − mean ∆sales in the pre-intervention stages at store I) − (mean ∆sales in the given intervention stage − mean ∆sales in the pre-intervention stage at store C), which can be calculated from values in Table 3. *P < 0.05, ^†P < 0.01

Adjusted estimates were calculated based on adjusted sales in 2007, which are described in Table 3