Of 28460 people invited, 4270 (15%) signed on to the study Web site and 2540 (8.9%) participated (). Data were dropped for 27 participants whose baseline and follow-up responses were inconsistent on key factors (e.g., gender, birth date), yielding 2513 participants. Women and minority members made up 69% and 32% of enrollees, respectively (). Mean age (46.3 years; SD=10.8) and distribution by gender, race/ethnicity, and living status (married or living with a partner) were virtually identical across arms. Half (50%) reported education as high school or less than a college degree and half had a college or postcollege degree. Half (50%) were in the “contemplation” stage of change, 20% were in “preparation,” 21% were in “maintenance,” and 7% had recently become “active” in reaching minimum fruit goal intake.28
Flow chart of recruitment and retention: Making Effective Nutritional Choices (MENU) study, 2005–2006.
Baseline Description of Participants Enrolled in an Online Intervention to Improve Fruit and Vegetable Consumption, by Study Arm: Making Effective Nutritional Choices (MENU) Study, 2005–2006
Overall, the baseline daily mean fruit and vegetable intake was 4.4 servings (SD=2.8) by the 16-item measure and 3.3 (SD=1.6) by the 2-item measure (). Despite randomization, statistically significant differences were found in reported fruit and vegetable intake at baseline by study arm when the 16-item measure was used, with fewer servings in arm 2.
Follow-up participation rates were 86% at 3 months, 80% at 6 months, and 80% at 12 months. Of the 2513 enrollees, 99.9% provided complete 2-item baseline responses and 97% provided complete 16-item baseline responses. For analysis, 80% provided usable 2-item survey data at both baseline and 12 months, and 71% provided usable 16-item survey data at both assesments. Of the 725 participants missing the 16-item assessments,10% at baseline and 20% at 12 months were excluded because of implausibly high reported fruit and vegetable servings and 70% did not complete both the baseline and the 12-month assessments.
Response rates for participants with complete 16-item data were 73% for arm 1, 71% for arm 2, and 69% for arm 3 (P=.18). Proportions of those completing both assessments differed significantly (P<.05) from the baseline-only participants by location (ranging from 60%–79% by site), age (for <30 years, 68%; for ≥60 years, 75%), gender (74% for women; 66% for men), race (77% for Whites versus 59% for African Americans, and 58% for Hispanics versus 72% for non-Hispanics), education (78% for postcollege versus 61% for high school or vocational school), and baseline fruit and vegetable servings as assessed by the 2-item measure (3.4 servings versus 3.2 servings). Stage of change for eating recommended servings of vegetables varied (P<.05): of baseline participants, 80% of those in the precontemplation stage were in the final analytic data set, compared with 73% of contemplators, 70% of maintainers, 66% of preparers, and 62% in the “action” stage; no statistical difference in response was seen by baseline stage of change for fruit (P=.13).
The number of responses per person to counselor-initiated e-mails ranged from 0 to 17 (mean=6). Participation rate in e-mail support exchanges in arm 3 varied, with 22% having no exchanges, 29% having 1 to 3, 17% having 4 to 8, and 33% having 9 or more. Counselors limited their e-mail messages to 2 short paragraphs; the length of participant-sent messages ranged from a single sentence to several paragraphs.
At 12 months, average fruit and vegetable servings increased by more than 2 servings across each arm, as assessed by the 16-item measure (). The adjusted mean change was statistically greater for arm 3 participants (+2.80 servings) than for control participants (+2.34 servings) (P=.025, or P=.050 after the Hochberg adjustment). There was no statistical difference between arms 1 and 2 (P=.177) or arms 2 and 3 (P=.37). Cohen’s effect sizes (0.6–0.8) indicate medium to large within-arm changes relative to small effect sizes (0.04–0.10) for between-arm differences.
Mean Fruit and Vegetable Servings Per Day at 12 Months Versus Baseline, by Outcome Measure and Study Arm: Making Effective Nutritional Choices (MENU) Study, 2005–2006
When the 2-item measure was used, with more complete follow-up data and adjusting for baseline servings, fruit and vegetable intake increased in the tailored arm (arm 2; P=.05) and in the tailored plus e-mail counseling arm (arm 3; P=.04); both of these arms increased by an average 2.55 servings compared with 2.38 servings for the control group (). When the 2-item measure was used, the proportion of participants within study arms eating 5 or more fruit and vegetable servings daily ranged from 20% to 22% at baseline to 67% to 75% at the 3-month assessment (P<.001; ). These gains were sustained at 6 and 12 months. At follow-up, 85% to 92% of participants reached minimum fruit intake goals and 59% to 73% met minimum vegetable intake goals; differences between arms were most pronounced at 3 and 6 months. For fruits alone, statistically significant differences between the arms remained at 12 months. The intervention effect was smaller but similar for vegetables, but results were imprecise at final follow-up, presumably because of lower statistical power.
Comparison of Fruit and Vegetable Consumption Levels Over Time Using the 2-Item Measure, by Study Arm: Making Effective Nutritional Choices (MENU), 2005–2006
Assessment of frequency of Web site visits across the duration of the study demonstrated that increases in fruit and vegetable servings corresponded with more visits, regardless of arm; mean daily intake increased by 2.2, 2.4, and 3.0 servings for participants who visited infrequently (<7 Web site visits), moderately often (7–13 visits), and very often (>13 visits), respectively (P<.001) (data not shown). Program satisfaction assessments at 6 and 12 months, regardless of arm, revealed that 95% of respondents rated the program as “very good” to “excellent”; in addition, 90% reported that they would recommend the program to others and that involvement of family and others had increased over time.