Adolescence is a time of transition [1
] in which youth begin to establish life-long diet and physical activity habits. During this time, physical activity typically decreases [2
], while independence in making food choices increases [5
]. In addition, adolescence is typically a time of rapid growth, requiring adequate nutrition to achieve full physiological potential [6
]. Thus, the diet and physical activity patterns established during these years not only impacts youth’s current health and body size, but can continue into adulthood and impact over health, weight and risk for chronic disease [7
]. Research shows that adolescents have poor dietary habits and eating patterns [8
] and do not meet the current dietary recommendations [9
]. Thus, understanding how to design and implement age-appropriate and accessible health promotion programs to address these issues is needed [7
]. To be successful, these programs must engage youth, while supporting healthy eating behaviors and a physically active lifestyle [9
]. Adolescent athletes participating in high school (HS) sports represent a potential target group for interventions to promote and build these healthy life-skills and behaviors [10
]. These active adolescents already know about physical activity, training hard, and working toward a goal, and are open to implementing diet behaviors that can help them improve sport performance. Why not capitalize on this interest by teaching sport nutrition education to improve eating behaviors for sport performance, health, and chronic disease prevention?
Based on 2017 data, 54.3% of United States (US) HS students participated in sports [11
]. One of the most popular youth team sports is soccer, especially in Latino youth [12
]. In 2016–2017, over 24,000 HS students played soccer [13
]. Unfortunately, adolescent athletes do not always make healthful food choices [14
] or have the best food options available to them at sporting events [16
]. Active youth, especially those engaged in intense sport participation, have unique energy and nutritional needs [17
]. Thus, the eating patterns of adolescent athletes need to meet their growth and developmental needs, and the energy and nutritional demands associated with general physical activity and sport participation [18
]. Adolescent female athletes also need to meet the energy demands of menses and reproductive health [21
]. Based on the 2014 Sports Dietitians Australia Position Statement on Adolescent Athletes [17
], adolescents involved in sport need diets that provide adequate energy (i.e., calories), protein, carbohydrate, unsaturated fat, iron, calcium, vitamin D and fluids. Athletes who restrict energy intake to maintain a low body weight [22
] or have diets high in processed foods [24
] are at the greatest risk for poor energy and nutrient intakes. Female athletes are especially at risk for poor energy [22
] and micronutrient intakes, including calcium [28
], B-vitamins [31
], zinc and iron [32
]. The International Olympic Committee (IOC) consensus statement on relative energy deficiency in sport (RED-S), highlights the numerous health and performance consequences of inadequate energy intake in athletes [26
]. The negative impact of restricted energy intake on bone health in female athletes, including adolescent female athletes, is well documented in the Female Athlete Triad [21
]. One potential factor contributing to poor energy and nutrient intakes in adolescent athletes is the timing of meal consumption, both throughout the day and around sport practice or games. Consistent meal patterns may preserve lean body mass [17
] and replenish stored carbohydrate (glycogen), which is an important fuel for exercise and brain function [17
]. Skipping meals, particularly breakfast, has been identified as a concern among adolescents in general [34
], but may be especially problematic for adolescent athletes. If breakfast is skipped, then only one meal opportunity will occur prior their typical early afternoon sport practice or game. This means athletes start their sport practice being inadequately fueled for intense physical activity. If adolescent athletes can learn how to fuel their body for sport and health by selecting healthy foods and appropriate beverages, they may establish and carry these diet behaviors into adulthood [37
Research examining sport nutrition knowledge (SNK) and behaviors in adolescent athletes is primarily cross-sectional and uses club- or elite-level athletes. Generally, this research shows that adolescent athletes have low SNK (ranging from 55–74%) [38
], and the SNK they do have does not consistently predict food choices [38
]. Recently, Manore et al. [42
] examined SNK, attitudes and beliefs in 535 HS soccer players. They found that SNK was low (46%), especially in Latino youth (39%) and those participating in National School Breakfast/Lunch Program (NSLP) (41%). Little research has explored SNK, attitudes or belief differences in active adolescents based on sex, race/ethnicity, or socioeconomic status (SES). Yet researchers repeatedly highlight the diet-related disparities that exist in adolescent populations based on sex [17
], race/ethnicity [45
], and SES [46
]. Finally, based on population census data, it is projected that youth of color will represent more than 50% of the US adolescent population by 2060 [47
]. Thus, there is a need for researchers to include more diverse youth when examining sport nutrition issues in adolescent athletes, especially adequate energy and intake of nutrient dense foods [27
Intervention programs focused on nutrition or behavior change in the adolescent athlete population are also limited. Available interventions have focused primarily on improving disordered eating, body image [10
] and restrictive dietary behaviors [49
] among female athletes, reducing alcohol and sport supplement use among male athletes [50
], and improving hydration in active youth [51
]. These interventions ranged from 1 day to 8 weeks. Only three interventions have targeted changing SNK and dietary practices in adolescent athletes [53
]. However, these studies all lacked a control group, had small sample sizes (n
= 11–49), focused on individual sports (swimming, cycling, sailing, etc.), and varied in duration and how SNK scores were reported and measured. Walter et al. [55
] focused on training undergraduate nutrition student to work with HS athletes and reported no changes in SNK and behaviors. To date, no team-based interventions have focused on changing SNK and behaviors in HS athletes, beyond changing hydration practices. Capturing adolescents while they are still active and engaged in youth sports provides a ‘window of opportunity’ to cultivate life-skills that support life-long health and obesity prevention, such as healthy eating behaviors, meal planning, grocery shopping and cooking skills. This is also an opportunity to teach youth athletes how to fuel and hydrate their body for sport, physical activity and health, and to discern if sport foods and supplements are needed.
The WAVE~Ripples for Change: Obesity Prevention in Active Youth (WAVE) intervention program was developed for HS soccer players to teach sport nutrition and life-skills (e.g., meal planning, shopping on a budget, food preparation/cooking skills, and gardening) to support sustainable healthy eating and to encourage physical activity outside of sport. Thus, the purpose of this 2-year study was to examine the impact of a sport nutrition education and life-skills intervention on SNK, attitudes/beliefs and behaviors among HS soccer players and to determine differences in outcomes based on sex, SES and race/ethnicity.
2. Materials and Methods
2.1. WAVE Program Overview
The WAVE program was a 2-year integrated (research, education, and extension) obesity prevention intervention targeting HS soccer players (aged 14–19 years). The intervention was age-specific and included health assessments, face-to-face sports nutrition lessons, experiential learning and team-building workshops (TBWs). The WAVE educational objectives were to teach life-skills (e.g., meal planning, shopping on a budget, food preparation/cooking skills, and gardening) and sports nutrition education to support sustainable healthy eating among HS soccer players. Further details of the larger study can be found elsewhere [42
]. Eligibility criteria included: (1) age 14–19 years; (2) enrolled in HS soccer; (3) living with a parent/caregiver; (4) no medical conditions preventing a normal diet; (5) internet access during the 2-year study; and (6) proficiency in English. Figure 1
shows the experimental design of WAVE program for this manuscript only, see the full WAVE experimental design in Supplement Figure S1
. Here we present the 2-year intervention data for changes in SNK, attitudes/beliefs and behaviors for participants who completed all questionnaires at each of three time points. The WAVE program and follow-up evaluations were approved by the Oregon State University (OSU) Institutional Review Board (#6317).
2.2. Recruitment and Participants
We used a multi-step recruitment process for this study. First, soccer coaches and their schools were recruited through OSU 4-H Soccer Program. Second, soccer players were recruited through their coaches. Third, parents were recruited at soccer parent meetings. Soccer teams were then assigned (non-randomized) to either intervention (n = 278; 9 schools) or comparison (n = 110; 4 schools) group based on geographical location within the Willamette Valley, Oregon.
The WAVE project recruited 864 HS soccer players with 72% (n
= 620) enrolled and submitting youth assent and parent consent forms [59
]. Figure 2
describes WAVE program youth participant attendance for each sport nutrition related activity for the 2-year period. For this manuscript, participants were included if they completed two questionnaires (described below) at baseline (time 1), end of year 1 (time 2) and end of year 2 (time 3) (n
= 217; 14–18 years; 13 schools, 24 soccer teams). Only the intervention group (n
= 153 (IG)) received all aspects of the intervention, which is described in Figure 1
. The comparison group (n
= 64 (CG)) also completed all assessment activities, but did not receive the intervention. Overall, the participants (n
= 217) for this study were 64.0% female, 71.4% in grades 9 and 10, predominately Latino (47.5%) and White (44.2%), and reported playing soccer an average of 6.9 years. As a group, 46.5% participated in NSLP and 65.4% reported no sport injuries during the last 12-months.
2.3. Assessments and Questionnaires
All participants completed the questionnaires either at school or while attending soccer camp. WAVE participant’s data were collected using two questionnaires: (1) a demographic, health history and sport experience questionnaire designed for the study, including questions regarding number of years playing soccer, sport related injuries, and general diet behaviors; and (2) a validated SNK questionnaire [39
]. Participants also indicated their participation in the NSLP, which was used as a proxy for low household income [23
]. The SNK questionnaire had been previously validated in HS rugby players in Ireland [39
]. The questionnaire consisted of 40 questions, including questions regarding training schedule. Table 1
outlines the key themes and topic areas addressed in the sport nutrition questionnaire. A SNK score was calculated for each athlete by adding the total number of correct questions (n
= 10) covering four domains (hydration, protein/carbohydrate, supplements and pre/post exercise food selection). The SNK questionnaire was administered in the presence of the researchers to minimize discussion of responses between participants, and no members of the coaching staff were present. Cronbach’s alpha was used to estimate internal reliability for the nutrition knowledge subsection, which was determined to be 0.51. Baseline results for the SNK, and attitudes and behaviors related to sport nutrition for WAVE participants can be found at Manore et al. [42
The 2-year intervention included sport nutrition lessons, experiential learning in the classroom, and TBWs, which were delivered to IG soccer teams during fall soccer season and summer camps. The WAVE HS sports nutrition curriculum [60
] was delivered to IG teams and coaches by a registered dietitian nutritionist (RDN), trained in sports nutrition and who had played collegiate/professional soccer prior to the intervention, all lessons were pilot-tested and revised based on input from athletes and sports nutrition experts (RDNs, Certified Specialist in Sport Nutrition (CSSD)). Topics covered were as follows: hydration; pre/during/post-exercise fueling; body composition/image; maintaining muscle and staying healthy; and eating well while dining out (see Table 2
). Lessons involved experiential learning such as role playing, cooking demonstrations and food tastings, meal and snack planning around training and games. Newsletters reinforced lessons and provided recipes/tips to meet sports fuel and nutrition needs. Three life-skill training sessions were delivered to teams via the TBWs (~1–1.5 h each) and focused on meal planning, shopping on a budget, food preparation/cooking skills, and gardening [58
]. See Figure 1
for WAVE Program timeline.
2.5. Statistical Analysis
The statistical analysis was completed in two phases. Phase one investigated the differences between those who completed all aspects of the intervention (n = 217; completers), and those lost to follow up (n = 340; incompleters). Phase two examined the difference between the groups (IG, n = 153; CG, n = 64) over time (3 time points) on SNK, attitudes and beliefs, and behaviors.
The outcomes (sociodemographic, psychosocial and behavioral variables) for participants who completed the 2-year intervention (n = 217; completers) were compared to those lost to follow-up (n = 340; non-completers). We used multivariate analysis to examine retention probability. Results showed that participants assigned to the IG were 2 times more likely to complete the 2-year intervention compared to CG peers (Odds Ratio (OR) 2.02, 95% Confidence Interval (CI) 1.29, 3.18, p = 0.002). Furthermore, females were 1.7 times more likely to participate throughout the intervention compared to males (OR 1.67, 95% CI 1.14, 2.46, p = 0.009). Participants in the 12th grade at baseline were less likely to complete the intervention compared to 9th-grade participants at baseline (OR 0.036, 95% CI 0.01, 0.10, p = 0.001), since 12th graders graduated before the intervention was over. No other differences based on socio- demographic factors was observed between those who completed the full protocol and those who did not. Daily breakfast was the only outcome variable (e.g., SNK, attitudes/beliefs, behaviors) that differed between completers and non-completers. At baseline, completers were 1.5 times more likely to report eating breakfast daily compared to non-completers (OR 1.56, 95% CI 1.05, 2.33, p = 0.000). Successful completion of the program did not depend on geographical location. All participants were recruited from two counties. Retention among the counties was similar (39.2% vs. 39.0%; X2 = 0.0046, p = 0.95).
For all participants (n = 217), descriptive statistics (mean, standard deviation (SD)) were calculated for baseline demographic characteristics. Independent t-tests were used to examine differences between IG and CG on age, number of years playing soccer, age started preparing meals and SNK scores. Chi-squared tests were used to examine differences between groups (IG, CG) for discrete variables (sex, race/ethnicity, year in school, no injuries in the past 12 months, participation in the NSLP, prepares meals for self). Chi-squared tests were also used to compare outcome variables related to attitudes, beliefs and behaviors between the groups over time.
To determine the impact of the intervention, outcome variables were compared between groups (IG, CG) over the three time periods. Mean differences between groups for variables related to SNK, attitudes/beliefs and behaviors were determined through longitudinal data analysis using generalized estimating equations. Each model contained the independent variables of time (time 1 = baseline, time 2 = end of year 1, time 3 = end of intervention; treated as a categorical variable) and group (IG or CG), and the interaction of group and time (time, group and group × time). Each model used age, race/ethnicity, sex, and SES as adjustment variables. During the 2-year study, the IG spent 4 h in sport nutrition lessons and 3 h in TBW. The impact of frequency of participation (e.g., hours) on SNK score and change score was examined with the addition of a dose variable (total hours of participation) in the model. We assumed that missing data were missing at random. All analyses were conducted using Stata SE 14.2 (College Station, TX, USA).
Participant demographic characteristics are presented in Table 3
. Overall, 60% of participants were female (p
= 0.005), and 46.5% participated in the NSLP (p
= 0.057), with more males (46.2%) than females (36%) (p
= 0.020), and more Latinos (82%) than White (14%) (p
< 0.0001). At baseline, the IG had played soccer longer (1.5-year longer, p
= 0.008) than CG. Results comparing the main outcome variables at baseline between groups showed that the IG reported that athletes had different nutritional requirements (IG = 60.6% vs. CG = 41.3%; X2
= 6.55, p
= 0.010), and that muscle mass is important to performance (IG = 68.3% vs.CG = 52.4%; X2
= 4.77, p
= 0.029). In addition, a greater number of IG reported consuming sugar sweetened beverages prior to physical activity (IG = 68.6% vs. CG = 51.6%; X2
= 5.68; p
= 0.017). No other differences between IG and CG were observed at baseline.
The WAVE program provided seven hours of nutrition education (sport nutrition lessons (4 h), TBW (3 h)) to the IG, with a mean participation of 4.1 h (SD = 1.6 h). Overall, there were no differences in participation hours based on sex, but the NSLP participants had a lower rate of participation than non-NSLP participants (3.6 h vs. 4.4 h; p = 0.001), and participants identified as White had higher participation (4.4 h) than Latino (3.5 h) (p = 0.0068).
3.1. Sport Nutrition Knowledge
The SNK results for each time period are presented in Table 4
, with a maximum possible score of 10 at each time point. At baseline (time 1) there was no statistical difference in the sex-race-income adjusted mean SNK score between groups. However, the IG significantly increased their SNK scores by ~10% from time 1 to time 3 (p
≤ 0.001), while there were no changes in the CG (see Table 4
). Within the IG at time 3, the final SNK scores were as follows: non-NSLP participants (64.3% vs. 54.5% NSLP), White (63.2% vs. 55.1% Latino) and males (61.1% vs. 59.1% females). The significant change over time in total SNK scores in the IG vs. CG is attributed to the significant improvement in the scores of the IG (time 1 = 51.6%; time 3 = 59.1%) vs. GC (time 1 = 44.8%; time 3 = 50.2%) female athletes (e.g., significant group × sex interaction). Among males, there was no difference in the sex-race-income-adjusted mean SNK score between the groups at any time period. For females, there was no statistical difference between IG and CG for the sex-race-income-adjusted mean SNK score at time 1. No other differences by sub-population (NSLP participation nor race/ethnicity) were observed in SNK score changes over time.
3.2. Attitudes and Beliefs Relevant to Sport Nutrition
shows the attitudes and beliefs relevant to sport performance between groups. At time 1, most participants (IG = 84.5%; CG = 88.9%) reported that diet was important to performance. These responses increased to 92.1% and 93.6% by time 3 in the IG and CG, respectively. Throughout the study, a greater percentage of IG (time 1 = 60.6%; time 3 = 73.0%) players reported that athletes have different nutritional requirements than their peers compared to the CG (OR 2.00 (CI = 1.28, 3.16; p
= 0.003). In addition, soccer players in the IG (time 1 = 68.3%; time 2 = 70.7%, time 3 = 62.2%) were twice as likely as the CG players to agree with the statement that muscle mass is important for performance (p =
0.002). No differences based on sex, race/ethnicity or NSLP participation status were observed.
Examination of changes in beliefs over time showed that at time 2, soccer players in the IG were twice as likely as CG players (CI = 1.20, 5.70; p
= 0.015) to report that they try to eat for performance (IG = 46.6% vs.CG = 33.3% CG). At time 3, IG players were three times likely as CG players (CI = 2.59, 7.77; p
= 0.002) to report that they try to eat for performance (48.7% IG vs. 30.2% IG). By time 3, soccer players within the IG (31.6%) were less likely to report that their diet met nutritional requirements than the CG (47.6%) (OR = 0.43; CI = 0.18, 0.99; p
= 0.48). Differences over time by sex, race/ethnicity or NSLP participation status were not observed. No other statistically significant difference in changes over time were observed in attitudes and beliefs relevant to sport nutrition (see Table 5
3.3. Dietary Behaviors Related to Sport Performance
Baseline values and changes over time among the five dietary behaviors are provided in Table 6
. At baseline, self-report of dietary behaviors related to sport performance were similar between the groups, except for the consumption of sugar sweetened beverages (SSB) 1–4 h prior to physical activity (p
= 0.017) (Table 6
). Among the IG, 68.6% reported consuming SSB, compared to 51.6% of the CG participants, with no differences based on NSLP or sex. For the IG, the consumption of SSB 1–4 h before physical activity decreased by 17% over the intervention, but there was no time × group interaction. Over half of the participants (53.7%) reported consuming breakfast daily at time 3, regardless of group assignment (X2
= 0.0123, p
= 0.912). There were no differences between groups for consumption of breakfast over time. There was a significant group × time interaction for the change in ‘eat lunch 5 or more days a week’ at time 2 (p
= 0.027) and time 3 (p
= 0.040). Throughout the intervention, the proportion of athletes within the IG reporting eating lunch at least 5 days per week remained at 92.2–93.4%, while the percentage of CG participants reporting this behavior decreased from 98.4% (time 1) to 90.6% (time 3). There were no group or time differences for fueling behavior before/after physical activity. For the IG group, eating within 1 h of physical activity increased by 10% from time 1 (33.1%) to time 3 (43.7%), while there was no change in the CG (time 1 = 38.7%; time 3 = 33.3%); however, the differences in change over time between groups were not statistically significant when controlled for sex, race/ethnicity and NSLP participation status. For both groups, the proportion of participants reporting ‘eat within 1 h after physical activity’ did not significantly change across the intervention (71–79%) (see Table 6