Disparities in Dietary Intake, Meal Patterning, and Home Food ...

3 downloads 0 Views 471KB Size Report
We examined whether young adult meal patterning, dietary intake, and home food availability differed among nonstudents, 2-year college students, and 4-year ...
RESEARCH AND PRACTICE

Disparities in Dietary Intake, Meal Patterning, and Home Food Environments Among Young Adult Nonstudents and 2- and 4-Year College Students Melissa C. Nelson, PhD, RD, Nicole I. Larson, PhD, MPH, RD, Daheia Barr-Anderson, PhD, Dianne Neumark-Sztainer, PhD, MPH, RD, and Mary Story, PhD, RD

We examined whether young adult meal patterning, dietary intake, and home food availability differed among nonstudents, 2-year college students, and 4-year college students (N=1687; mean age=20.5 years). Unadjusted analyses showed that few young adults consumed optimal diets and, compared with 4-year college students, nonstudents and 2-year students consumed fewer meals and poorer diets. After controlling for sociodemographics and living arrangements, we found that over half of the observed associations remained significant (P < .05). Nutrition interventions are needed for young adults, particularly specific at-risk groups. (Am J Public Health. 2009;99:1216–1219. doi:10. 2105/AJPH.2008.147454) Little is known about eating habits and food practices during young adulthood, and few studies to date have explored the heterogeneity of young adult lifestyles and possible contextual influences on diet. Previous research on young adults has primarily focused on youths attending 4-year colleges,1–4 with little understanding of diet-related factors among those who have less traditional college experiences (such as those who attend 2-year community or technical colleges) or who do not enroll in college after high school.4 Thus, we sought to examine differences in young adults’ dietary

intake, meal patterns, and home food availability among (1) nonstudents, (2) students at 2-year colleges, and (3) students at 4-year colleges.

METHODS Data for this cross-sectional analysis were from Project Eating Among Teens II (EAT-II), a longitudinal study of Minnesota adolescents and young adults.5 Our sample consisted of 750 men and 937 women who completed the EATII survey and the Youth–Adolescent Food-Frequency Questionnaire at follow-up as young adults (mean age=20.5 years) from 2003 to 2004. The EAT-II survey included items assessing meal and snack frequency and home food availability. Scores for home food availability were created by summing the availability of specific examples of 5 healthful and 4 unhealthful foods and beverages.6–8 Scores ranged from 5 to 20 (for healthful foods and beverages) and 4 to 16 (for unhealthful foods and beverages); higher scores indicate greater availability of healthful (or unhealthful) foods and beverages. The food frequency questionnaire was used to assess participants’ overall past-year dietary intake.9–11 Sociodemographic characteristics were selfreported and included race/ethnicity, socioeconomic status (SES), age, and parental status.12 Student status was defined as not a student, student at a community or technical college (2-year college), or student at a 4-year college. Past-year living arrangements included rented apartment or house, parent’s home, or on campus (including residence hall and fraternity or sorority house). We used linear regression to estimate differences in diet-related outcomes by student status. Models were examined (1) unadjusted, (2) adjusted for sociodemographics, and (3) additionally adjusted for living arrangement. Because responders to the EAT-II survey were demographically different from nonresponders, all analyses were adjusted for differential response rates using response propensity weights,13 as described elsewhere.14 Analyses were conducted with SAS version 8.2 (SAS Institute, Cary, NC).

RESULTS Sociodemographic factors varied by student status (Table 1). Among 4-year college students,

1216 | Research and Practice | Peer Reviewed | Nelson et al.

38% rented apartments or houses, 25% lived with parents, and 37% lived on campus (data not shown). Among 2-year students, 31% rented, 68% lived with parents, and 1% lived on campus. Among nonstudents, 38% rented and 62% lived with parents. Unadjusted models indicated that most young adults did not meet national recommendations for dietary intake (Table 2).15,16 Four-year college students reported eating meals more frequently and had better dietary intakes than did 2-year students and nonstudents. Nonstudents reported the lowest home availability of healthful foods. Overall, a majority of relationships remained statistically significant (19 of the 25 crude significant associations) in models adjusted for sociodemographic factors. After further adjustment for living arrangement, 13 of the original 25 associations continued to be significant.

DISCUSSION We examined differences in dietary factors among young adults by student status. In general, nonstudents and 2-year college students reported less frequent meals and poorer dietary intake compared with 4-year college students. Most differences were still evident after we controlled for sociodemographic factors, such as race/ethnicity and SES. Although some associations were attenuated after adding contextual factors (i.e., living arrangement) to the model, over half (52%) of the associations originally detected in the crude models remained statistically significant. In this examination of young adult dietary patterns, crude and adjusted estimates were informative in different ways. Crude estimates indicated which subgroups were most at risk and thus may be the most effective targets for intervention strategies. By contrast, adjusted estimates allowed us to further explore the extent to which the existence of dietary differences across young adult subgroups are independent of sociodemographic factors, which we know are important correlates of both student status and dietary intake. Significant differences observed in adjusted models suggest that nonstudents and 2-year college populations are important targets for nutrition interventions not only because they represent a greater number of racial/ethnic minorities and

American Journal of Public Health | July 2009, Vol 99, No. 7

RESEARCH AND PRACTICE

TABLE 1—Participant Characteristics, by Student Status: Project Eating Among Teens (EAT)-II Survey, Minnesota, 2003–2004 Total, % (No.) or Mean

Nonstudent, % (No.) or Mean

2-y College Student, % (No.) or Mean

4-y College Student, % (No.) or Mean

45 (742)

45 (333)

22 (163)

33 (245)

55 (923)

34 (316)

29 (265)

37 (342)

Low

36 (597)

55 (331)

26 (153)

19 (113)

Middle

51 (834)

32 (265)

28 (235)

40 (335)

High

13 (208)

18 (36)

16 (34)

66 (138)

White

56 (925)

32 (293)

23 (214)

45 (418)

Black

16 (263)

48 (126)

37 (97)

15 (40)

Asian Otherb

17 (276) 11 (179)

43 (118) 57 (102)

23 (64) 27 (49)

34 (94) 16 (28)

Gender Men Women Socioeconomic status,a

P