Low Health Literacy Is Associated With Energy-Balance-Related Behaviors, Quality of Life, and BMI Among Rural Appalachian Middle School Students: A Cross-Sectional Study

Abstract

BACKGROUND:

Many studies document associations between low health literacy (HL) and poor health behaviors and outcomes. Yet, HL is understudied among adolescents, particularly from underserved, rural communities. We targeted rural adolescents in this cross-sectional study and explored relationships between HL and (1) energy-balance-related health behaviors and (2) body mass index (BMI) and quality of life (QOL).

METHODS:

Surveys were administered to 7th graders across 8 middle schools in rural Appalachia. HL was assessed using the Newest Vital Sign. Energy-balance-related behaviors and QOL were assessed using validated instruments. Height and weight were objectively measured. Analyses were conducted using the Hodges-Lehmann nonparametric median difference test.

RESULTS:

Of the 854 adolescent students (mean age = 12; 55% female), 47% had limited HL. Relative to students with higher HL, students with lower HL reported significantly lower frequency of health-promoting behaviors (water, fruit and vegetable intake, physical activity, sleep), higher frequency of risky health behaviors (sugar-sweetened beverages, junk food, screen time), and had higher BMI percentiles and lower QOL (all p < .05).

CONCLUSIONS:

Low HL is associated with energy-balance-related behaviors, BMI, and QOL among rural, Appalachian adolescents. Findings underscore the relevance of HL among rural middle school students and highlight implications for school health.

Health literacy (HL) is a key aspect of health behavior decision making. HL is defined as an individual’s ability to access and understand basic health information to make important health decisions.¹ Low HL skills have consistently been linked to poor health outcomes among adults.² In the United States, an estimated 88% of adults may lack the HL skills needed to manage their health and prevent disease.³ However, there remains a gap in understanding the prevalence of low HL among adolescents and its impact on their current health behaviors and outcomes as well as their health status as adults.

In the last decade, research pertaining to adolescent HL has begun to emerge.^4,5 Adolescent HL studies have primarily focused on risky health behaviors, such as sexual behaviors and alcohol, tobacco, and substance use.⁵ Yet, the observed relationships between adolescent HL and these health behaviors have been mixed.^6–8 Furthermore, few adolescent HL studies have focused on energy-balance-related behaviors, such as nutrition, physical activity, sedentary activity, and sleep. This is important, as the proportion of overweight or obese adolescents in the United States has increased from 24.7% to 30.4% over the past 3 decades and is a serious public health concern.⁹ Of the 5 known studies that have explored the relationship between adolescent HL and obesity or energy-balance-related behaviors, 3 have found a strong correlation between low HL and obesity or greater weight,^6,10,11 3 observed significant associations between low HL and unhealthy diets,^6,12,13 one found no relationship between HL and physical activity,¹² and one found no associations with sleep.¹³ Although other studies have observed a significant relationship between time spent watching television and obesity in adolescents,¹³ no known studies have examined the relationship between HL and sedentary behaviors. Finally, no studies have explored the relationship between adolescent HL and quality of life (QOL).^4,5 This is an important consideration, given the existing literature has shown adolescents with higher physical activity levels, lower sedentary behaviors, and healthier body mass index (BMI) report higher QOL.¹⁴ Understanding how HL relates to energy-balance-related behaviors, BMI, and QOL is an important next step, particularly among health disparate adolescent populations.

Furthermore, HL is understudied among adolescents residing in rural, medically underserved areas. This is problematic, as these adolescents face a myriad of barriers to health, including limited resources and access to health care, low socioeconomic status, and low educational attainment. Individuals who face these inequalities are more likely to have low HL.¹⁵ Compounding this issue, adolescence encompasses a wide age range (ages 10–19), and populations represented in current HL literature vary substantially from elementary to middle school to high school-aged students.⁵ This broad age range makes it difficult to draw conclusions and consider implications regarding the role of adolescent HL in health behavior decision making. Collectively, these inconsistencies and gaps in the literature highlight the need and opportunity to better understand the prevalence and consequences of limited HL among rural adolescents.

In rural, Appalachian Virginia, the targeted region of this research, there are striking disparities in health outcomes and determinants of health. The region has among the poorest scores for health behaviors in the state,¹⁶ and counties within the region rank low and very low in terms of health opportunity index.¹⁷ These health disparities are further compounded by high-poverty rates,^18,19 low educational attainment,²⁰ lack of resources and evidence-based prevention programs, and limited access to medical care.^19,21,22 Compared to other areas of Virginia, Appalachian communities have a higher prevalence of obesity and obesity-related chronic disease.^23,24,25 Understanding the relationship between adolescent HL and energy-balance-related behaviors is critical to inform future HL interventions in Appalachia and other underserved rural areas.

This cross-sectional study targets 7th grade adolescents in Appalachia Virginia. The primary objectives are to determine the prevalence of limited HL and to explore the relationship between adolescents’ HL status and energy-balance-related behaviors, including nutrition, physical activity, sedentary activity, and sleep. Secondary objectives are to explore how HL is related to BMI and school-related QOL among adolescents. We hypothesize that students with low HL would report lower frequency of health-promoting behaviors and higher frequency of risky-health behaviors, relative to students with higher HL. We further hypothesize that students with low HL would have a higher BMI percentile and report poorer QOL, compared to students with higher HL.

METHODS

Participants and Procedures

This cross-sectional study is a secondary analysis and uses baseline data (collected September 2018 and 2019) from a cluster randomized control trial of an intervention being conducted in middle schools in rural Appalachia Virginia.²⁵ The 4 counties represented in this study classify among the lowest county health rankings in Virginia.¹⁶ Out of 133 ranked counties, the 4 counties range from 72 to 123 for health outcomes and 79 to 132 for health factors. Eight middle schools were recruited based on the following criteria (1) within the Central Appalachia region of southwest Virginia, (2) 7th grade student body of approximately 80–200 students, and (3) 8th grade students located in the same school building as 7th grade students. These criteria were based on logistical and power calculations for the intervention trial.²⁵ School administrators and teachers were informed of the research protocol and agreed to the study procedures.²⁵ The University of Virginia Institutional Review Board approved this study. Prior to participation, caregivers provided written or verbal consent and students provided written assent. Students received small incentives for participating (highlighters, t-shirts).

Instrumentation

A survey consisting of validated measures was administered to 7th grade students during their usual PE/Health class time. The survey took approximately 45 minutes to complete. Trained research staff facilitated the survey and read each question aloud to students and assessed height and weight data on a separate day.

Demographics.

Self-reported sex and age, measured by birthday were assessed.

Health literacy status.

HL status was assessed using the Newest Vital Sign (NVS).²⁶ The NVS is an objective HL assessment using the Nutrition Facts label. It has been validated for use in adolescents.^27,28 The NVS is scored by summing the total number of correct responses out of 6 questions and has 3 HL category cut points.

Energy-balance-related behavior measures.

Nutrition behaviors included adolescents’ self-report of sugar-sweetened beverages (SSB), water, fruit and vegetables, and junk food.

The Beverage Intake Questionnaire (BEVQ-15), which has been validated in youth, was used to assess SSB and water intake.^29,30 The instrument assesses frequency of consumption across 7 response categories and portion sizes across 6 response categories. Five SSB types (regular soda/pop, sweetened juice drinks, sweetened tea, energy drinks, sweetened coffee) and one water item were assessed. Standardized scoring procedures were used to compute total ounces per day for SSB and water.^30,31

Questions from the Family Life, Activity, Sun, Health, and Eating (FLASHE) Teen Diet survey were used to assess daily fruit and vegetable and junk food intake over the past 7 days.³² Daily fruit and vegetable intake were computed from 5 items: fruit, green salad with or without vegetables, other non-fried vegetables, cooked beans, non-fried potatoes. Daily junk food intake was assessed with 5 items including fried potatoes; candy or chocolate; sweets/treats such as cookies, cakes, cupcakes, doughnuts, brownies, poptarts; ice cream or other frozen desserts; and regular potato chips, corn chips, or cheese puffs. Standardized scoring procedures were applied.³³

Physical activity, screen time, and sleep were assessed as single-item questions from the validated Middle School Youth Risk Behavior Surveillance System (YRBSS) Survey.³⁴ Physical activity was reported as the number of days per week students were physically active: “any kind of physical activity that increased your heart rate and made you breathe harder” for at least 60 minutes per day. Two screen time questions assessed hours per day of TV watched and hours per day of video games, computer games, and computer, tablet, and smartphone use not related to schoolwork using a 7-point scale (range = none to 5 or more hours per day). Sleep was reported as hours of sleep on an average school night on a 7-point scale (range = 4 hours or less to 10 or more hours).

Quality of life measures.

A single-item question from the Behavioral Risk Factor Surveillance System (BRFSS) survey asked students to rate their perception of their overall health on a 5-point Likert-type scale ranging from 1 (poor) to 5 (excellent).³⁵ The Pediatric Quality of Life inventory (PedsQL) was used to assess school-related QOL.³⁶ This included 5, 5-point Likert-type scale items: “it is hard to pay attention in class,” “I forget things,” “I have trouble keeping up with my schoolwork,” “I miss school because of not feeling well,” “I miss school to go to the doctor or hospital” which were scored on a 0–100 scale.

Height, weight, and BMI percentile.

Height and weight were collected in a private area of the school. Height was measured by having students stand without shoes on a research-grade portable stadiometer. Height was assessed twice to the nearest half centimeter and both measurements were averaged to obtain a final height. Adolescents’ weight was obtained without shoes and using a research-grade calibrated digital Tanita scale. BMI percentile was calculated using the Center for Disease Control and Prevention’s age and sex-specific growth charts, which includes student’s date of birth, sex, weight, height, and date of data collection.³⁷

Data Analysis

(SPSS Version 26.0, Armonk, NY) was used for summary statistics. Stata Version 16.0 was used to examine differences in student health behaviors, BMI, and QOL by HL status. Due to the non-normality of most of our outcome variables and the clustering within schools, the Hodges-Lehmann nonparametric median difference test was used to examine distribution differences while taking into account the school clustering in inference.³⁸ This test is robust and useful in exploring differences along the distribution of data. Two different options for categorizing adolescents by their HL levels were explored. Option 1 included students who scored 0–1 on the NVS (n = 400) (limited HL) versus students who scored a 2–6 (n = 454) (possibly limited to adequate HL). Option 2 included students who scored 0–3 (n = 702) (limited to possibly limited HL) versus students who scored a 4–6 (n = 152) (adequate HL). Findings and interpretations were remarkably similar, so option 1 is reported. These groupings allowed for results to be explored along the distribution of the data, with health-promoting behaviors (higher scores are desired) explored among the lower percentile distributions (20–50%), and risky health behaviors (lower scores are desired) explored among the higher percentile distributions (50–80%). Significance level was set at p < .05.

RESULTS

Sample Characteristics

Of 1360 eligible students, 874 (64%) students enrolled in the study, and 862 (63%) completed the baseline survey. Of those, 8 students were excluded (one outlier in reported SSB intake and 7 due to errors and missing data). In total, 854 students with complete data were included in this study. Table 1 illustrates descriptive characteristics among students. In brief, students’ mean HL score was 1.86 (SD = 1.62) with the majority scoring in high likelihood of limited HL and possibility of limited HL categories. Also, about one-half of students were in the overweight or obese weight categories.

Table 1.

Student Demographic Factors, Health Literacy Status, BMI, Energy-Balance-Related Behaviors, and Quality of Life (n = 854)

	n*	%
Sex
Male	380	45
Female	460	55
Age (years)
11	34	4
12	655	79
13 or older	136	17
Health literacy†
Limited HL (score 0–1)	400	47
Possibility of limited HL (score 2–3)	302	35
Adequate HL (score 4–6)	152	18
BMI
Underweight (< 5th percentile)	14	1.8
Normal or Healthy weight (5th-<85th percentile)	363	45.7
Overweight (85th-<95th percentile)	159	20.0
Obese (≥95th percentile)	259	32.6
	Mean	SD
Health-promoting behaviors
Water (ounces)	30.7	29.1
Fruits and vegetables (servings/day)	2	1.8
Physical activity (days/week)	4.6	2.2
Sleep (hours/day)	7.4	1.5
Risky health behaviors
SSB (ounces/day)	36.8	42.7
Junk food (servings/day)	2.4	2.1
Screen time (TV hours/day)	1.5	1.5
Screen time, video game/smartphone use (hours/day)	2.9	1.8
Health status and QOL
Health status (scaled 1, lowest-5, highest)‡	3.6	0.9
School-related QOL (0, lowest-100, highest)§	67.4	17.3

^*Does not always add up to 854 due to missing data.

^†Measured with the Newest Vital Sign and scored using validated procedures.

^‡Measured with the Behavioral Risk Factor Surveillance System single-item question.

^§Measured with the PedsQL and scored using validated procedures.

Health-Promoting Behaviors by Health Literacy Status

The outcome differences examined are between students with limited HL and those with the possibly limited to adequate HL. Therefore, for health-promoting behaviors, the lower percentiles of the differences reveal the concerning disparity experienced by the vulnerable population of interest (ie, those with limited HL engaged in much less health-promoting behaviors as compared to those with possibly limited to adequate HL). The 95% confidence intervals that do not overlap with zero show the nonoverlap between the 2 HL groups and statistically confirm the disparity. As indicated in Table 2, relative to students with the possibly limited to adequate HL, students with limited HL had significantly lower health-promoting behaviors for each of the 4 behaviors (ie, those lower percentiles of the group differences exhibit statistically significant nonoverlaps).

Table 2.

Differences in Students’ Health-promoting Behaviors and Risky Behaviors by Limited Health Literacy Status Versus Possibly Limited to Adequate Health Literacy Status^† (n = 854^‡)

	Percentile Differences (95% CI§)
Health-Promoting Behaviors	20%	30%	40%	50%
Water (ounces)	−30.9* (−35.7,−26.3)	−20* (−24,−16)	−12* (−16,−8)	−2.9 (−8, 0)
Fruits and vegetables (servings/day)	−1.4* (−1.7,−1.2)	−.9* (−1,−.7)	−.4* (−.6,−.3)	0 (−.3, .2)
Physical activity (days/week)	−3* (−4,−3)	−2* (−2,−2)	−1* (−2,−1)	0 (−1, 0)
Sleep (hours/day)	−2* (−2,−2)	−1* (−2,−1)	−1* (−1,−1)	0 (0, 0)
Risky health behaviors	50%	60%	70%	80%
SSB (ounces)	4.6* (1.2, 8)	12.9* (8.3, 18)	24* (17.2, 31.7)	42* (29.7, 54.8)
Junk food (servings/day)	0 (−.1, .3)	0.5* (.3, .7)	1* (.7, 1.3)	1.8* (1.4, 2.4)
Screen time (TV hours/day)	0 (0, 0)	.5 (0, 1)	1* (1, 1.5)	2* (1.5, 2)
Screen time (video game/smartphone use hours/day)	0 (0, 0)	1 (0, 1)	1.5* (1, 2)	2.5* (2, 3)

^*Significant at p < .05.

^†NVS score limited HL = 0–1, possibility of limited HL = 2–3, adequate HL = 4–6.

^‡Variations in sample size range from 795–854 and are due to missing data.

^§95% CI accounts for school clustering; cohorts range in size from 40 to 158.

Water.

The lowest 20th percentile of the group differences shows that students with limited HL drank 30.9 less ounces of water than those with possibly limited to adequate HL (95% CI, −36 to −26 oz), p < .05. Significant differences were also observed at the 30th and 40th percentiles, with 20 less ounces and 12 less ounces, respectively.

Fruit and vegetable intake.

Similarly, the group differences in daily fruit and vegetable intake showed significant differences among the lowest 20th, 30th, and 40th percentiles. The largest disparity (as shown in the 20th percentile of the differences) shows that students with limited HL consumed 1.4 less servings of fruits and vegetables per day than those with possibly limited to adequate HL (95% CI, −1.7 to −1.2 servings), p < .05.

Physical activity.

Significant group differences in daily physical activity were seen among the 20th, 30th, and 40th percentiles. The lowest 20th percentile of the group differences shows that limited HL students had 3 less active days per week than those with possibly limited to adequate HL (95% CI, −4 to −3 days), p < .05.

Sleep.

The lowest 20th percentile of the group differences shows that students with limited HL had 2 less hours of sleep per night than those with possibly limited to adequate HL (95% CI, −2 to −2 hours), p < .05. Significant group differences were also seen among the 30 and 40th percentiles.

Risky Health Behaviors by Health Literacy Status

For risky health behaviors, the higher percentiles represent the vulnerable population of interest (ie, those with limited HL engaged in riskier health behaviors as compared to those with possibly limited to adequate HL). As shown in Table 2, relative to students with the possibly limited to adequate HL, students with limited HL had significantly higher risky behaviors for all variables (ie, those higher percentiles of the group differences exhibit statistically significant nonoverlaps).

Sugar-sweetened beverages.

The highest 80th percentile of the group differences shows that students with limited HL drank 42 oz more of SSB per day than those with possibly limited to adequate HL (95% CI, 29.7 to 54.8 oz), p < .05. Significant differences were also observed at the 60 and 70th percentiles.

Junk food.

Similarly, the group differences in junk food intake showed significant differences among the 60th, 70th, and 80th percentiles. The largest differences (as shown in the 80th percentile) show that students with limited HL consumed 1.8 more servings of junk food per day than those with possibly limited to adequate HL (95% CI, 1.4 to 2.4 servings), p < .05.

Screen time.

The highest 80th percentile of differences shows that students with limited HL watched 2 hours more TV per day than those with possibly limited to adequate HL (95% CI, 1.5 to 2 hours), and had 2.5 more hours of daily time spent on video games, computer use and smartphone use (95% CI, 2 to 3 hours), p < .05. Significant group differences were also observed at the 70th percentile.

Measured BMI, Self-Rated Health, and Quality of Life by Health Literacy Status

For measured BMI, the higher percentiles of differences represent the vulnerable population of interest. For self-rated health status and school-related QOL, the lower percentiles of differences represent the vulnerable population of interest. As indicated in Table 3, relative to students with the possibly limited to adequate HL, students with limited HL had significantly higher BMI and significantly lower health status and QOL.

Table 3.

Differences in BMI, Self-rated Health Status, and Quality of Life (QOL) by Limited Health Literacy Status Versus Possibly Limited to Adequate Health Literacy Status^† (n = 854^‡)

	Percentile Differences (95% CI§)
BMI	50%	60%	70%	80%
BMI percentile	−0.4 (−2.2, 1)	4.1* (1.3, 7.9)	13.2* (7.7, 20.2)	30* (21.5, 37.7)
Health status and QOL	20%	30%	40%	50%
Health status (scaled 1, lowest-5, highest)||	−1* (−1,−1)	−1* (−1,−1)	0 (0, 0)	0 (0, 0)
School-relatedQOL (0, lowest-100, highest)¶	−25* (−25,−20)	−15* (−20,−15)	−10* (−10,−5)	−5 (−5, 0)

^*Significant at p < .05.

^†NVS score limited HL = 0–1, possibility of limited HL = 2–3, adequate HL = 4–6.

^‡Variations in sample size range from 795–854 and are due to missing data.

^§95% CI accounts for school clustering; cohorts range in size from 40 to 158.

^∥Measured with the Behavioral Risk Factor Surveillance System single-item question.

^¶Measured with the PedsQL and scored using validated procedures.

BMI status.

The highest 80th percentile of the group differences shows that students with limited HL had 30 more BMI percentile points than those with possibly limited to adequate HL (95% CI, 21.5 to 37.7 percentile points), p < .05. Significant differences were also observed at the 60 and 70th percentiles.

Self-rated health.

The lowest 20th percentile of the group differences shows that students with limited HL had 1 less unit of self-rated health than those with possibly limited to adequate HL (95% CI, −1 to −1 units), p < .05. Significant differences were also observed at the 30% percentile.

Quality of life.

The 20th, 30th, and 40th percentiles of the group differences showed significant differences in QOL. The greatest difference was observed among the 20th percentile, in which students with limited HL had lower QOL (25 points) than those with possibly limited to adequate HL (95% CI, −25 to −20 points), p < .05.

DISCUSSION

Consistent with our hypothesis and relative to students with higher HL, students with lower HL reported lower frequency of health-promoting behaviors, higher frequency of risky health behaviors, and had higher BMI percentile and poorer QOL. Our study represents one of the most comprehensive evaluations examining the relationships among HL and a broad range of energy-balance-related behaviors and outcomes among adolescents. Furthermore, this is the first known HL and energy-balance-related behaviors study among underserved, rural Appalachian adolescents. Given the limited research available on this topic, our research fills an important gap in the literature.

Importantly, our findings highlight inequities among Appalachian adolescents. Nearly half of students (47%) had limited HL. Other known adolescent studies that have used the NVS have found limited HL ranging from 11% to 34%.^8,10,39 Likewise, relative to national recommendations and data, our sample of adolescents generally reported lower health-promoting behaviors and higher risky health behaviors. For example, the 2015–2020 Dietary Guidelines for Americans recommend that adolescents consume 3.5 to 5 servings of fruits and vegetables per day.⁴⁰ Adolescents in our study consume a mean of 2 servings of fruits and vegetables per day (SD = 1.8). Further, our students consume high amounts of SSB (36.8 oz or about 450 cal of SSB per day) compared to national averages that indicate US adolescents ages 12–19 years consume on average 225 cal of SSB per day.⁴¹ Finally, the prevalence of overweight (20.0%) and obesity (32.6%) among students is higher than national averages of 18.1% and 20.6%.⁴² Collectively, our data highlight the disparities in health behaviors and outcomes among Appalachian adolescents, and the importance of addressing limited HL.

Study findings can also be interpreted within the context of other adolescent-focused HL studies. Across all health-promoting behaviors, limited HL students had significantly lower frequencies of water intake, fruit and vegetable intake, physical activity, and sleep. These results support previous findings that low HL is associated with unhealthy diets and poor nutrition behaviors.^6,12,13 Our findings of physical activity and sleep are novel in the adolescent HL literature. Though one previous study examined exercise, it did not find associations with adolescent HL.¹² To our knowledge, no previous studies have examined the association between adolescent HL and sleep.

Furthermore, we found that limited HL students engaged in all risky health behaviors at levels significantly higher than students with possibly limited to adequate HL. The findings of low HL students reporting unhealthier diets are reflected in the literature.⁶ Specifically, other adolescent HL studies report low HL adolescents are less likely to exhibit healthy nutrition behaviors¹² and nutrition literacy significantly impacts nutrition habits.¹³ To our knowledge, no other adolescent HL studies have examined the relationship between HL and sedentary activity, such as screen time.

Adolescents with limited HL had significantly higher BMI percentiles than their higher HL peers. Other adolescent studies have documented similar relationships among low HL and weight and obesity.^6,10,11 Given that our overall adolescent sample had disproportionately higher BMI percentiles compared to national averages,⁴² the differences among low HL students are particularly concerning. Targeting HL skills to improve energy-balance-related behaviors among adolescents should be a public health priority, especially among rural Appalachian schools.

Finally, we found significant associations between low HL and indicators of QOL. These results support the adolescent literature showing associations between low HL and low self-rated health status.^6,12 Although the adolescent health behavior literature has shown strong associations between health behaviors and school performance, no known studies have examined the relationship between adolescent HL and school-related QOL. This is important, as energy-balance-related behaviors and health status influence school functioning and academic success.⁴³

Schools are an opportunistic setting to improve HL skills, as they maximize reach to students. Furthermore, students’ health behaviors and wellness directly influence academic achievement and success.⁴⁴ Given the existing adolescent HL literature and findings from our study, prioritizing school-based health interventions and curricula designed to develop HL skills may be an important strategy for improving health and obesity-related outcomes among adolescents in health disparate, rural communities. Building adolescent HL within schools aligns with state standards of learning. For example, in Virginia, the new 2020 standards for health education focus on building “the knowledge and skills to make healthy decisions” and “the ability to access, evaluate, and use health information.”⁴⁵ More importantly, HL aligns with larger school health frameworks such as the Whole School, Whole Community, Whole Child (WSCC) Model.^46,47 Notably, within health education, achieving HL is one of the main goals by building knowledge, attitudes, and skills to support positive health behavior decision making. Numerous studies have documented the importance of the WSCC Model within schools, and the critical connection between health behaviors and students’ academic achievements.^43,48,49 Finally, although these frameworks and models emphasize the importance of HL, prospective studies and randomized control trials are needed to study the associations among HL and energy-balance-related behaviors and health outcomes over time.

Limitations

Several study limitations should be considered when interpreting findings. First, the self-reported survey is subject to social desirability and response bias. Second, measuring HL and its multiple domains, functional, communicative/interactive, and critical, is complex. The NVS has been validated in adolescents, measures functional HL, and has been used in numerous studies;^8,10,39,50 yet, it does not assess all HL domains. However, there are currently no validated HL measures to assess each domain. Third, this study does not address potential confounding variables, such as socioeconomic status. However, several adolescent HL studies document significant associations between HL and health behaviors even when controlling for sociodemographic characteristics.^6,8,12 Fourth, due to the cross-sectional design, findings do not reflect causality. Finally, given the rural Appalachian school districts targeted in this study and exclusive focus on 7th grade students, findings may lack general-izability. Despite these limitations, this research is strengthened by its large sample size of students across 8 middle schools in 4 counties, use of validated instruments, rigorous protocols to collect data, and robust hypotheses-driven analyses.

Conclusions

This research fills an important gap in the dearth of literature on adolescent HL. It also highlights the relevance of HL among adolescents residing in the underserved, rural Appalachian region. Adolescents with greater HL skills have healthier behaviors, are at a healthier weight, and have better self-rated health and school-related QOL. Future school-based studies should focus on interventions to promote HL and examine relationships between HL and energy-balance-related health behaviors over time.

IMPLICATIONS FOR SCHOOL HEALTH AND EQUITY

Adolescence is a critical time period for developing positive health behaviors. Schools play an important role in supporting adolescents’ health behavior decision making which ultimately leads to lifelong health habits. Schools also have the unique opportunity to teach students HL skills in their curricula. Given the range of HL among adolescents, it is impractical to segment students by HL status to deliver health education. Consistent with state standards and the WSCC framework, this study has several implications that should be considered for improving adolescent HL and school health:

• Schools should prioritize teaching for behavior change to improve energy-balance-related behaviors, rather than solely relying on teaching for knowledge gain and information.

• Educators should focus curricula on improving health communication skills and building health behavior decision making skills around health-promoting behaviors and risky-health behaviors.

• Health education in schools should take into account universal HL practices, such as encouraging questions, using the teach-back method, and creating action plans.⁵¹

Considering HL within health education can benefit under-resourced schools, such as those with limited staffing and with financial and time constraints. This is particularly relevant for schools in underserved, rural areas who often lack teaching resources and updated health curricula. The detrimental impact of COVID-19 on both the health care and education systems, as well as the expected surge in the obesity crisis,^52,53 further highlights the importance of prioritizing adolescent HL. Addressing HL in schools may be an important strategy for improving health behaviors and abating energy-balance-related disparities. Improving HL skills may enable students to develop lifelong healthy behaviors that can also influence the health and wellness of their peers, families, and communities.