Multiple Tailored Messages are Effective in Increasing Fruit and Vegetable Consumption Among Callers to the Cancer Information Service

ABSTRACT

Results are reported from a large (n = 3,402) four-group randomized trial to increase fruit and vegetable consumption among callers to the National Cancer Institute's (NCI's) Cancer Information Service (CIS) using tailored print materials. Following a baseline telephone interview, which included a brief educational message (BEM), participants were assigned randomly within CIS offices to one of four groups: single untailored (SU) group—one untailored set of materials; single tailored (ST) group—one tailored booklet; multiple tailored (MT) group—four tailored materials; and multiple retailored (MRT) group—four tailored materials with retailoring based on new information obtained at 5 months follow-up. Follow-up telephone interviews were conducted at 5 (n = 2,233) and 12 months (n = 1,927) after baseline. The main outcome measure was self-reported fruit and vegetable consumption using a seven-item food frequency questionnaire. At 12 months follow-up, there was a significant linear trend across groups of 0.21 servings (p = 0.0002). Specific nested hypotheses then were tested and revealed significant mean serving differences between SU (5.07) vs. MT (5.64) (p = 0.002) and SU vs. MRT (5.71; p < 0.001). Although the mean for ST (5.40) was greater than that for SU (5.07), the difference was not statistically significant (p = 0.07), and no difference was found between MT vs. MRT (p = 0.69). A higher proportion of recipients of tailored materials reported reading all of the materials and believing that they were written especially for them. No differences by experimental condition were found for the perceived usefulness or motivational impact of the print materials. In this trial, MT print materials were more effective at increasing fruit and vegetable (FV) consumption than were SU materials. The intervention mechanisms responsible for this effect merit further research. Retailoring did not produce a significant difference when compared with longitudinal baseline tailoring.

Notes

^a p < 0.05 for chi-square test for independence of rows and columns at 5 months.

^b p < 0.05 for Mantel-Haenszel chi-square test for trend at 5 months. 

^c p <0.05 for Mantel-Haenszel chi-square test for trend at 12 months.