Adolescent individual, school, and neighborhood influences on young adult hypertension risk.

BACKGROUND: Geographic and contextual socioeconomic risk factors in adolescence may be more strongly associated with young adult hypertension than individual-level risk factors. This study examines the association between individual, neighborhood, and school-level influences during adolescence on young adult blood pressure.
METHODS: Data were analyzed from the National Longitudinal Study of Adolescent to Adult Health (1994-1995 aged 11-18 and 2007-2008 aged 24-32). We categorized hypertension as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg. Secondary outcomes included continuous systolic and diastolic blood pressure. We fit a series of cross-classified multilevel models to estimate the associations between young adulthood hypertension with individual-level, school-level, and neighborhood-level factors during adolescence (i.e., fixed effects) and variance attributable to each level (i.e., random effects). Models were fit using Bayesian estimation procedures. For linear models, intra-class correlations (ICC) are reported for random effects.
RESULTS: The final sample included 13,911 participants in 128 schools and 1,917 neighborhoods. Approximately 51% (7,111) young adults were hypertensive. Individual-level characteristics-particularly older ages, Non-Hispanic Black race, Asian race, male sex, BMI, and current smoking-were associated with increased hypertension. Non-Hispanic Black (OR = 1.21; 95% CI: 1.03-1.42) and Asian (OR = 1.28; 95% CI: 1.02-1.62) students had higher odds of hypertension compared to non-Hispanic White students. At the school level, hypertension was associated with the percentage of non-Hispanic White students (OR for 10% higher = 1.06; 95% CI: 1.01-1.09). Adjusting for individual, school, and neighborhood predictors attenuated the ICC for both the school (from 1.4 null to 0.9 fully-adjusted) and neighborhood (from 0.4 to 0.3).
CONCLUSION: We find that adolescents' schools and individual-level factors influence young adult hypertension, more than neighborhoods. Unequal conditions in school environments for adolescents may increase the risk of hypertension later in life. Our findings merit further research to better understand the mechanisms through which adolescents' school environments contribute to adult hypertension and disparities in hypertension outcomes later in life.

Introduction

High blood pressure and hypertension are growing problems in adolescents and young adults. The estimated prevalence of hypertension in the United States in 2017–2018 was 6–10% among adolescents (10–17 years) and 22.4% among young adults (18–39 years) [1-5]. The prevalence in adolescents is as high as 30% in adolescent boys with obesity and 23–30% in girls with obesity [6]. National statistics indicate that nearly 1 in 4 young adults in the U.S. experience elevated blood pressure [7, 8]. Hypertension is higher among young adult men than young adult women aged 18–39 (31.2% compared with 13.0%) [5, 9].

The study of hypertension risk factors in adolescence and young adulthood has many important public health implications. Hypertension, one of the major modifiable risk factors for cardiovascular disease (CVD), is established early in life. In a meta-analysis of 50 cohort studies, data from diverse populations show that blood pressure tracks from childhood into adulthood [10]. In addition, hypertension that begins in childhood is associated with adverse cardiac changes and vascular damage that in turn is associated with premature cardiovascular disease in adulthood [11]. Nevertheless, studies are limited examining whether adolescent risk factors related to sociocultural contexts (e.g., schools and neighborhoods) are associated with young adulthood hypertension; and thus, evidence is required to fill this gap and inform hypertension interventions in adolescence.

Studies have shown that schools and neighborhoods—the two contexts in which adolescents spend most of their time—have important bearings on cardiovascular risk factors including hypertension, diabetes, and obesity [12-17]. Cross-sectional and longitudinal evidence demonstrates that schools are more salient than neighborhoods in explaining variation in weight gain and body mass index (two important hypertension risk factors), and that schools provide direct opportunity and support for dietary intake and exercise. However, most studies face three key limitations: (1) they use single-level analysis, which cannot capture hypertension risk at a contextual level; (2) they examine one context at a time, making it difficult to compare the relative influences of multiple contexts; and (3) they examine the association cross-sectionally during adolescence and estimate the relationship between contextual risk factors and hypertension only in adolescence, and not young adulthood. To our knowledge, no prior study has integrated these multilevel contexts during adolescence and compared their long-term influences on hypertension in young adulthood.

This study uses data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) to investigate whether school and neighborhood contexts and their characteristics during adolescence are associated with the likelihood of hypertension in young adulthood. We hypothesized that higher socioeconomic status at both the school-level and neighborhood-level would predict lower hypertension risk in young adulthood. We hypothesized that school factors would be more strongly associated with young adult hypertension than neighborhood factors [7, 15].

Methods

Data collection

The National Longitudinal Study of Adolescent to Adult Health (Add Health) is a nationally representative cohort of adolescents in the U.S. who have been followed from adolescence through adulthood to identify social, behavioral, and biological determinants of health across the life course [18]. The Add Health study design is coordinated by the Carolina Population Center, as detailed elsewhere [19]. We obtained Institutional Review Board approval to conduct secondary analyses of the Add Health data using deidentified data obtained under an Add Health Restricted-Use Data Contract at the University of California, San Francisco.

Participants

Add Health longitudinally follows a nationally representative sample of adolescents in grades 7 to 12 and ages 11–18 at baseline (Wave I; interviewed 1994–1995; N = 20,745) into adulthood. A sample of 80 high schools and 52 feeder middle schools from the United States was selected to ensure representation of U.S. schools with respect to region of country, urbanicity, school size, school type, and ethnicity. The current study uses in-home interview data from Wave I and Wave IV (aged 24–32 years; interviewed 2007–2008; N = 15,701). Of respondents who participated in both waves (N = 15,701), we excluded those missing measured blood pressure at Wave IV (n = 334); school contextual data (n = 910); neighborhood contextual data (n = 6); or individual SES measures (n = 540), resulting in a final analytic sample of 13,911 respondents.

Outcomes

We constructed all outcome variables from the Wave IV in-home interview. After the interview, participants rested in a seated position for five minutes, after which three measures of blood pressure were recorded. Trained and certified Add Health field interviewers followed a computer-assisted data collection protocol to record participants’ blood pressure. Interviewers measured blood pressure using an appropriately sized arm cuff and an automatic oscillometric monitor approved by the British Hypertension Society (BP 3MC1-PC_IB; MicroLife USA, Inc., Dunedin, FL) [20]. Three blood pressure measurements were taken at 30-second intervals from the right arm with the patient in the resting, seated position after 5 minutes of rest. The second and third measurements were double-entered and then averaged to give the final blood pressure recorded, which we used in the present study to measure hypertension status. After blood pressure measurement, the interviewer inventoried and recorded antihypertensive medications (beta-adrenergic blockers; calcium channel blockers; angiotensin converting enzyme inhibitors; angiotensin II receptor blockers; centrally or peripherally acting anti-adrenergics; vasodilators; thiazide diuretics; antihypertensive combinations) used by participants within the preceding four weeks.

In primary analyses, we classified participants as hypertensive according to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure—if they had an average measured systolic blood pressure (SBP) ≥ 140 mmHg, an average measured diastolic blood pressure ≥ 90 (DBP) mmHg, or use of antihypertensive medications [20]. In secondary analyses, we applied a more recent definition of hypertension based on the 2017 American College of Cardiology/American Heart Association Hypertension Guideline definition, and classified participants as hypertensive if they had an average SBP ≥ 130 mmHg and/or an average DBP ≥ 80 (DBP) mmHg, or use of antihypertensive medications [21]. Mean arterial pressure (MAP) was conventionally approximated as the weighted sum of systolic and diastolic blood pressure, using the following formula where the weights for SBP (1/3) and DBP (2/3) reflect the typical contributions of ventricular systole and diastole to the duration of the cardiac cycle.

Individual variables

We constructed individual covariates using data from the Wave I in-home interview, including adolescents’ biological sex (male, female), race/ethnicity (non-Hispanic Black, Hispanic, Asian and Pacific Islander, Other, Multiracial, and non-Hispanic White). At the individual-level, SES was determined based on parental education and receipt of public assistance. We used data from either the youth or caregiver interview to capture receipt of public assistance (mother currently receiving public assistance, such as welfare or not) and highest level of parental education (defined as the maximum level of education by the resident mother, resident father, or resident step-father/partner (no high school diploma or equivalent; completed high school or equivalent; completed some college, trade school or a 2-year degree; completed equivalent 4-year college degree or above). Height and weight were measured by trained interviewers at Wave IV. Young adult body mass index (BMI) at Wave IV was calculated as the ratio of weight in kilograms over height in meters squared. Age at Wave IV (in years) was calculated from the date of Wave IV in-home interview and participant’s date of birth.

School variables

We constructed school-level covariates using data from the Wave I data. Using the survey of the full sample of schools, at the school-level, we created a continuous measure of school-level SES by aggregating individual-level data. Use of individual-level data was required as information about school-level SES was not directly available. We calculated the proportion of students within each school whose mother had received public assistance or had a college degree.

Neighborhood variables

We constructed neighborhood-level covariates using data from the Wave I data. At the neighborhood level, we used data from the 1990 Census to create a neighborhood-level SES measure indicating the proportion of residents within each neighborhood who had received public assistance or had a college degree. We also calculated the proportion of students in either the school or the neighborhood who were White.

Statistical analysis

We examined bivariate associations of individual-, school-, and neighborhood-level characteristics by Wave IV hypertension status. Differences in demographics and SES by hypertension were examined using two-sample t-tests for continuous variables and chi-square tests for categorical variables. School- and neighborhood-level demographics and SES are summarized using means (standard deviations). All tests were performed at an alpha-level of 0.05.

We constructed a series of cross-classified multilevel models (CCMM) to estimate the associations between Wave IV hypertension, systolic blood pressure, and mean arterial pressure with individual-level, school-level, and neighborhood-level factors (i.e., fixed effects) and the proportion of variance in Wave IV outcomes attributable to each level (i.e., random effects) [22]. The major advantage of CCMM compared with traditional multilevel models is that it allows for estimation of the effects of multiple non-nested contexts (e.g., students may attend schools outside of their neighborhoods and schools may draw students from multiple neighborhoods).

To parse out the effects of individual-level, neighborhood-level, and school-level contributions on hypertension, SBP, DBP, and MAP we ran logistic and linear regression models with model-building proceeding in a number of steps. We first examined the independent contributions of neighborhood and school contexts on the outcome using two-level hierarchical null (or unconditional) models. These models were fit by including individuals nested within either the school- or neighborhood-level. Next, school and neighborhood contexts were examined simultaneously by allowing for cross-classification of the two contexts. A total of four null models were fit including (1) individual-only, (2) individual and school, (3) individual and neighborhood, and (4) individual, school, and neighborhood. Subsequent models incorporated this cross-classification of school and neighborhood and the adjustment for other predictors via the following model equation. For example, SBP (denoted y) for an adolescent in the study (denoted i) nested in a given school (denoted j) and neighborhood (denoted k) was modeled as: with the following fixed effect parameters: β0 refers to the overall mean SBP or MAP y across all schools and neighborhoods, β refers to the vector of individual-level covariates, β refers to the vector of school-level covariates, and β refers to the vector of neighborhood-level covariates. Random effect parameters included the following: e0 refers to the individual-level random effect variance parameter for the individual within the combination of school j and neighborhood k, u0 is the variance at the school-level and u0 is the variance at the neighborhood-level. A series of five adjusted cross-classified models were fitted. Model 1 adjusted for individual-level predictors including age, sex, race/ethnicity, parental education, and parental receipt of public assistance. Model 2 included individual-level predictors as well as the following school-level predictors: percentage of students of non-Hispanic White race, percentage of students whose parents receive public assistance and percentage of students whose parents have a college degree. Model 3 included individual predictors plus neighborhood-level predictors from the Census: percentage of residents’ White race, percentage of residents receiving public assistance and percentage of residents with a college degree. Model 4 presents adjusted model all socioeconomic individual-, school, and neighborhood-level predictors. Model 5 presents the fully-adjusted model, which included all individual-, school-, and neighborhood-level predictors including individual BMI and smoking. All models for MAP, DBP and SBP additionally adjusted for self-reported use of antihypertensive medications.

For linear regression models predicting SBP, DBP, and MAP we report parameter estimates (β) and 95% credible intervals (CI) for fixed effects, parameter estimates (95% CI) for intercepts while variance estimates (95% CI) and intra-class correlations (ICC) are reported for random effects. ICCs allow for comparison of variance parameters across contextual levels and are interpreted as the percent of variance attributable to a given level. For logistic models predicting hypertension, we present odds ratios (OR) and 95% credible intervals for fixed effects, parameter estimates (95% CI) for intercepts, and variance estimates (95% CI) and ICC for random effects [23]. Model fit was evaluated using the deviance information criterion (DIC), which is a test statistic produced by the MCMC procedure that refers to the model complexity and “badness of fit” with higher DIC values indicate a poorer fitting model [24].

Models were fit using MLwiN (version 3.00; Birmingham, UK) via Stata’s runmlwin command. MLwiN uses Bayesian estimation procedures using Markov Chain Monte Carlo (MCMC) methods with non-informative priors and a Metropolis-Hastings sampling algorithm allowing for simultaneous modeling of non-hierarchically nested contexts [24-27]. All univariate and bivariate analyses were preformed using Stata version 16 (College Station, TX).

Results

The final analytic sample included a total of 13,911 participants from Wave IV from 128 schools and 1,917 neighborhoods. As shown in Table 1, the mean age of the analytic sample in Wave IV was 28.9 (SD = 1.7), 53.1% of the participants were non-Hispanic White, 20% were non-Hispanic Black, and 16% were Hispanic. The mean age of the analytic sample in Wave I was 15.6 (SD = 1.7). Participants’ average SBP, DBP, and MAP were 124.5 (13.6), 79.0 (10.2), and 94.2 (10.7) mmHg, respectively; 509 (3.7%) of Wave IV study sample reported use of antihypertensive medications. Of the 13,911 Wave IV participants included in this study, 7,111 (51%) young adults were classified as hypertensive. For all outcomes, S3 Table presents results for null cross-classified multilevel models.

Table 1

Individual-, school-, and neighborhood-level Wave I (1994–1995) characteristics of participants in young adulthood at Wave IV (2008–2009; N = 13,926) of the National Longitudinal Study of Adolescent to Adult Health.

N (%)	Wave IV Total Sample (N = 13,926)	Wave IV Hypertension (N = 2,881)	Wave IV No Hypertension (N = 11,045)
Individual-level (N = 13,926)				P-Value
Age (years), Mean (SD)	28.94 (1.72)	29.14 (1.71)	28.89 (1.72)	P<0.001
Sex				P<0.001
Female	7,373 (53)	1,028 (14)	6,345 (86)
Male	6,553 (47)	1,853 (29)	4,700 (71)
Race/ethnicity				P<0.001
Non-Hispanic White	7,395 (53)	1,482 (51)	5,913 (80)
Non-Hispanic Black	2,831 (20)	669 (23)	2,162 (76)
Asian	768 (6)	164 (6)	604 (79)
Hispanic	2,190 (16)	407 (14)	1,783 (81)
Other	172 (1)	41 (1)	131 (76)
Multiracial	570 (4)	118 (4)	452 (79)
BMI, kg/m2 (WIV)				P<0.001
Under Weight (<18.5)	192 (1)	16 (0.6)	176 (1.6)
Normal Weight (18.5–24.9)	4,395 (32)	459 (16)	3,936 (35)
Overweight (25–29.9)	4.199 (30)	818 (29)	3,381 (30.6)
Obese (≥30)	5,125 (37)	1,568 (55)	3,546 (32)
Unknown	10 (0.1)	4 (0.1)	6 (0.05)
Current Smoking (WIV)				P<0.001
No	8,916 (64)	1,735 (61)	7,181 (65)
Yes	4,887 (35)	1,101 (38)	3,786 (34)
Unknown	108 (1)	30 (1)	78 (1)
Antihypertensive medications
No	13,417 (96)	2,372 (82)	11,045 (100)
Yes	509 (4)	509 (18)	0 (0)
Parent Receipt of Public Assistance				P = 0.18
No	12,703 (91)	2,605 (90)	10,098 (91)
Yes	1,223 (9)	276 (10)	947 (9)
Parental Education				P = 0.30
Less than high school	1,664 (12)	338 (11)	1,326 (12)
High school graduate/GED	3,611 (26)	815 (28)	2,796 (25)
Some College	4,154 (30)	876 (30)	3,278 (29)
College graduate or beyond	4,497 (32)	852 (29)	3,645 (33)
School-level (N = 128)
	Mean (SD)	Median	Minimum	Maximum
Percent of students Non-Hispanic White	47.5 (25.5)	55.0	0	85.9
Percent of parents receiving public assistance	10.4 (9.4)	7.2	0	45.4
Percent of parents with college degree	31.7 (16.9)	28.3	5.5	91.2
Neighborhood-level (N = 1,917)
Percent of residents Non-Hispanic White	67.1 (32.5)	79.7	0	100
Percent of residents receiving public assistance	10.5 (9.6)	7.2	0	61.8
Percent of residents with college degree	23.8 (14.6)	20.3	1.1	77.8

Hypertension (140/90 mmHg)

In the null cross-classified model (Model 4, S3 Table), the random effects for school- and neighborhood-levels were 5% and 1% respectively. Table 2 shows the series of adjusted cross-classified models predicting hypertension (140/90 mmHg) among young adults in Wave IV. In the model adjusting only for individual factors (Model 1), the random effects for school- and neighborhood-levels were 1.08% and 0.3%, respectively. These indicate that the majority of variation in hypertension is due to individual or unmeasured variation, with a small percentage attributable to the school and a negligible percentage to the neighborhood. When school-level predictors were added to the model (Model 2), the school-level variance slightly increased to 1.2% attributable to school while the neighborhood variance remained stable with 0.3% of the variance being attributable to the neighborhood. Model 3 introduces neighborhood predictors into the individual-only model, and variance contributions of the school and neighborhood were similar to Model 2 (1.2% and 0.3%, respectively). In the fully-socioeconomic and fully-adjusted CCMM (Model 4) accounting for individual, school, and neighborhood-level predictors, ICCs for the school and neighborhood decreased to 0.9 and 0.3%. Comparing the variance parameters from the fully-adjusted model (Table 2, Model 5) to the null model (S3 Table, Model 4), adjusting for individual, school and neighborhood predictors attenuated the random effects for both the school (from 1.4 null to 0.9 fully-adjusted) and neighborhood (from 0.4 to 0.3).

Table 2

Logistic cross-classified multilevel models (CCMM) predicting hypertension from individual-, school- and neighborhood-level factors in the National Longitudinal Study of Adolescent to Adult Health, Wave IV (WIV), 2008–2009 (N = 13,926).

Hypertension (140/90)	Model 1	Model 2	Model 3	Model 4	Model 5
	Individual Cross-Classified	Individual and School Cross-Classified	Individual and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified
Fixed effect estimates Odds Ratios (95% Credible Interval)
Intercept	-3.55 (-3.94, -3.11)	-3.34 (-3.90, -2.98)	-3.38 (-4.15, -2.50)	-2.79 (-3.37, -2.11)	0.04 (0.02, 0.10)
Individual-level
Age, years (WIV)	1.09 (1.07, 1.11)	1.08 (1.06, 1.10)	1.09 (1.06, 1.12)	1.06 (1.04, 1.09)	1.05 (1.02, 1.07)
Female	0.41 (0.37, 0.44)	0.41 (0.37, 0.44)	0.41 (0.37, 0.44)	0.41 (0.37, 0.44)	0.40 (0.36, 0.44)
Race/ethnicity
Non-Hispanic White	REF	REF	REF	REF	REF
Non-Hispanic Black	1.28 (1.13, 1.42)	1.35 (1.17, 1.55)	1.19 (1.02, 1.37)	1.21 (1.03, 1.42)	1.21 (1.03, 1.42)
Asian	1.06 (0.86, 1.30)	1.14 (0.91, 1.40)	1.03 (0.83, 1.24)	1.11 (0.88, 1.39)	1.28 (1.02, 1.62)
Hispanic	0.93 (0.80, 1.07)	0.97 (0.83, 1.14)	0.90 (0.77, 1.04)	0.96 (0.82, 1.13)	0.92 (0.77, 1.08)
Other	1.23 (0.82, 1.78)	1.28 (0.84, 1.86)	1.21 (0.81, 1.75)	1.15 (0.84, 1.73)	1.17 (0.79, 1.66)
Multiracial	1.14 (0.99, 1.32)	1.17 (0.92, 1.26)	1.10 (0.88, 1.35)	1.14 (0.91, 1.40)	1.13 (0.90, 1.40)
Parent receipt of public assistance	1.11 (0.96, 1.30)	1.09 (0.92, 1.26)	1.08 (0.91, 1.26)	1.07 (.92, 1.23)	1.06 (0.89, 1.24)
Parental Education
Less than high school	REF	REF	REF	REF	REF
High school graduate / GED	1.14 (0.98, 1.32)	1.13 (0.98, 1.29)	1.14 (0.98, 1.32)	1.12 (0.96, 1.28)	1.11 (0.94, 1.30)
Some college	1.03 (0.88, 1.18)	1.02 (0.89, 1.17)	1.03 (0.87, 1.19)	1.01 (0.87, 1.17)	1.02 (0.85, 1.21)
College graduate or beyond	0.92 (0.78, 1.07)	1.00 (1.00, 1.02)	0.93 (0.78, 1.08)	0.92 (0.78, 1.07)	0.98 (0.81, 1.21)
BMI, kg/m2 (WIV)
Under or Normal Weight	-	-	-	-	REF
Overweight	-	-	-	-	1.92 (1.70, 2.16)
Obese	-	-	-	-	3.87 (3.47, 4.31)
Unknown	-	-	-	-	9.64 (1.77, 28.01)
Current smoking (WIV)
No	-	-	-	-	REF
Yes	-	-	-	-	1.15 (1.04, 1.25)
Unknown	-	-	-	-	1.36 (0.84, 2.04)
School-level, per 10%
Percent of students Non-Hispanic White	-	1.03 (1.00, 1.07)	-	1.05 (1.02, 1.08)	1.06 (1.01, 1.09)
Percent of parents receiving public assistance	-	1.06 (0.96, 1.16)	-	1.02 (0.93, 1.13)	1.02 (0.92, 1.13)
Percent of parents with college degree	-	0.98 (0.94, 1.04)	-	0.97 (0.93, 1.02)	0.98 (0.95, 1.03)
Neighborhood-level, per 10%
Percent of residents Non-Hispanic White	-	-	0.99 (0.97, 1.02)	0.96 (0.93, 0.98)	0.97 (0.94, 1.01)
Percent of residents receiving public assistance	-	-	1.07 (0.96, 1.17)	1.05 (0.96, 1.14)	1.05 (0.95, 1.16)
Percent of residents with college degree	-	-	0.99 (0.95, 1.04)	1.01 (0.95, 1.06)	1.05 (0.99, 1.10)
Random effect and variance estimates (95% Credible Interval) [ICC, %]
School	0.04 (0.02–0.06) [1.08]	0.04 (0.01–0.07) [1.20]	0.04 (0.01–0.07) [1.20]	0.03 (0.00–0.06) [0.90]	0.03 (0.00, 0.06) [0.90]
Neighborhood	0.01 (0.00–0.02) [0.30]	0.01 (0.00–0.02) [0.30]	0.01 (0.00–0.03) [0.30]	0.01 (0.01–0.03) [0.30]	0.01 (0, 0.03) [0.30]
Fit statistics (DIC)	13669.94	13670.47	13671.14	13671.22	13033.32

In the fully-adjusted CCMM, we found significant associations between hypertension and individual-level fixed effects for age, female sex, race/ethnicity, BMI, and current smoking. For every additional year in age, young adults had 1.05 higher odds of hypertension (95% CI: 1.02, 1.07). Females were less likely to have hypertension than males (OR = 0.40, 95% CI: 0.36, 0.44). Both Non-Hispanic Black (OR = 1.21; 95% CI: 1.03–1.42) and Asian (OR = 1.28; 95% CI: 1.02–1.62) students had higher odds of hypertension as compared to non-Hispanic White students. Compared to students with under or normal weight BMIs, students with overweight BMI (OR = 1.92; 95% CI: 1.70–2.16) or obese BMI (OR = 3.87; 95% CI: 3.47–4.31) had increased odds of hypertension. Moreover, current smokers had higher odds of hypertension (OR = 1.15; 95% CI: 1.04–1.25).

At the school level, we detected an association between hypertension and the percentage of non-Hispanic White students in the school (OR for 10% higher = 1.06; 95% CI: 1.01–1.09). Moreover, compared to the individual, school, and neighborhood only cross-classified models, the fully adjusted cross-classified models accounting for individual, school, and neighborhood fixed effects had the lowest DIC value indicating a stronger fitting model. There was no association with the percentage of students in the school whose parent had received public assistance or percentage of students in the school whose parents have a college degree and young adult hypertension. Neighborhood-level fixed effects were not associated with hypertension.

Results from cross-classified logistic models predicting a more recent definition of hypertension (SBP/DBP of ≥130/80 or use of antihypertensives) during young adulthood (Wave IV) are presented in S1 Table and were comparable to the findings from logistic regression with our primary hypertension classification. For example, we found significant associations between hypertension and individual-level fixed effects for age (OR = 1.06; 95% CI: 1.05–1.08), female sex (OR = 0.33; 95% CI:1.05–1.08), overweight BMI (OR = 1.87; 95% CI: 1.71–2.04) or obese BMI (OR = 3.65; 95% CI: 3.33–3.98), current smoking (OR = 1.11; 95% CI: 1.02–1.20). School and neighborhood-level fixed effects were not associated with hypertension (130/80 mmHg). Comparing the variance parameters from the fully-adjusted model (Table 3, Model 5) to the null model (S3 Table, Model 4), adjusting for individual, school and neighborhood predictors attenuated the variance contributions for both the school (from 1.4 null to 0.6 fully-adjusted) and neighborhood (from 0.4 to 0.3).

Table 3

Linear cross-classified multilevel models (CCMM) predicting systolic blood pressure from individual-, school- and neighborhood-level factors in the National Longitudinal Study of Adolescent to Adult Health, Wave IV (WIV), 2008–2009 (N = 13,926).

Systolic Blood Pressure (mmHg)	Model 1	Model 2	Model 3	Model 4	Model 5
	Individual Cross-Classified	Individual and School Cross-Classified	Individual and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified
Fixed effect estimates β (95% CI)
Intercept (SE)	123.39 (119.34, 127.32)	123.83 (119.53, 127.99)	124.59 (120.32, 128.79)	125.22 (120.84, 129.62)	120.49 (116.35, 124.67)
Individual-level
Age, years (WIV)	0.21 (0.07, 0.34)	0.20 (0.06, 0.33)	0.21 (0.07, 0.34)	0.18 (0.05, 0.31)	0.11 (-0.02, 0.24)
Female	-9.93 (-10.36, -9.51)	-9.93 (-10.35, -9.51)	-9.94 (-10.35, -9.52)	-9.95 (-10.37, -9.53)	-9.50 (-9.89, -9.10)
Race/ethnicity
Non-Hispanic White	REF	REF	REF	REF	REF
Non-Hispanic Black	1.88 (1.25, 2.50)	2.15 (1.45, 2.87)	1.54 (0.76, 2.31)	1.68 (0.91, 2.45)	1.35 (0.59, 2.08)
Asian	1.88 (-1.45, 0.63)	-0.16 (-1.28, 0.93)	-0.53 (-1.59, 0.52)	-0.25 (-1.35, 0.86)	0.59 (-0.50, 1.65)
Hispanic	-0.69 (-1.41, 0.06)	-0.48 (-1.28, 0.33)	-0.84 (-1.62. -0.06)	-0.52 (-1.32, 0.27)	-0.96 (-1.68, -0.22)
Other	1.24 (-0.62, 3.12)	1.45 (-0.54, 3.41)	1.14 (-0.84, 3.11)	1.38 (-0.60, 3.27)	0.98 (-0.90, 2.85)
Multiracial	0.009 (-1.07, 1.09)	0.15 (-0.93, 1.27)	-0.08 (-1.15, 1.02)	0.07 (-1.03, 1.18)	-0.29 (-1.34, 0.74)
Parental education
Less than high school	REF	REF	REF	REF	REF
High school graduate / GED	0.026 (-0.37, 1.17)	0.35 (-0.42, 1.13)	0.17 (-0.61, 0.97)	0.39 (-0.35, 1.17)	0.31 (-0.42, 1.03)
Some college	0.39 (-0.95, 0.57)	-0.19 (0.99, 0.58)	0.45 (-0.32, 1.24)	-0.13 (-0.91, 0.65)	-0.05 (-0.79, 0.69)
College graduate or beyond	-0.19 (-1.86, 0.27)	-1.01 (-1.78, -0.22)	-0.05 (-0.85, 0.72)	-0.86 (-1.67, -0.06)	-0.25 (-1.01, 0.48)
Parent receipt of public assistance	0.26 (-0.52, 1.05)	0.28 (-0.51, 1.07)	0.02 (-0.61, 0.97)	0.21 (-0.59, 1.00)	0.14 (-0.61, 0.92)
Antihypertensive medication	7.58 (6.45, 8.69)	7.57 (6.45, 8.70)	7.55 (6.42, 8.67)	7.52 (6.38, 8.65)	5.63 (4.55, 5.55)
BMI, kg/m2 (WIV)
Under or Normal Weight					REF
Overweight					5.03 (4.52, 5.55)
Obese					9.42 (8.93, 9.89)
Unknown					16.96 (9.66, 24.11)
Current smoking (WIV)
No					REF
Yes					1.07 (0.64, 1.49)
Unknown					0.95 (-1.33, 3.22)
School-level, per 10%
Percent of students Non-Hispanic White		0.09 (-0.06, 0.25)		0.17 (-0.01, 0.36)	0.15 (-0.01, 0.15)
Percent of parents receiving public assistance		-0.19 (-0.70, 0.32)		-0.19 (-0.74, 0.35)	-0.29 (-0.80, -0.29)
Percent of parents with college degree		-0.17 (-0.40, 0.07)		-0.08 (-0.35, 0.18)	0.02 (-0.22, 0.17)
Neighborhood-level, per 10%
Percent of students Non-Hispanic White			-0.06 (-0.19, 0.06)	-0.14 (-0.29, 0.01)	-0.07 (-0.21, -0.07)
Percent of parents receiving public assistance			-0.03 (-0.50, 0.44)	0.06 (-0.43, 0.56)	0.16 (-0.29, 0.16)
Percent of parents with college degree			-0.30 (-0.54, -0.04)	-0.23 (-0.50, 0.03)	-0.02 (-0.26, -0.02)
Random effect and variance estimates (95% Credible Interval) [ICC, %]
Individual	157 (153, 160) [99]	156 (152, 160) [99]	156 (152, 160) [99]	156 (152, 160) [99]	141 (138, 145) [99]
School	1 (0.6, 2.2) [0.9]	1 (0, 2) [0.7]	1 (0, 1.9) [0.6]	1.16 (0.5, 2.1) [0.7]	0.9 (0.3, 1.6) [0.6]
Neighborhood	0.1 (0, 0.4) [0.1]	0.4 (0, 1.9) [0.3]	0.4 (0, 1.7) [0.4]	0.06 (0.01, 0.2) [0.3]	0.1 (0, 0.4) [0.1]
Fit statistics (DIC)	109841.9	109850.1	109847	109839.5	108468.56

Systolic and diastolic blood pressure blood pressure

In our null models predicting systolic blood pressure, not accounting for fixed effects at any level, individual level random effects accounted for 98.5% of the variance, school for 1.1% and neighborhoods for 0.4% (S3 Table, Model 4). With the inclusion of individual level fixed effects, the individual level random effects did not substantially increase (from 98.5% in the null model to 99% in the individual-only model). The same trend held for models adding school-level effects (school-level random effect: 0.7%) and adding neighborhood-level fixed effects (neighborhood-level random effect: 0.4%). In null models predicting diastolic blood pressure, not accounting for fixed effects at any level, individual level random effects accounted for 99% of the variance, school for 0.1% and neighborhoods for 0.1% (S3 Table, Model 4). Similar to results for systolic blood pressure, with the inclusion of fixed effects at each level, the individual, school, or neighborhood-level random effects did not substantially change (Table 4). We then examined the associations of characteristics of individuals, schools and neighborhoods with young adult systolic blood pressure, and found that compared to hypertension, many significant relationships did not remain. For example, students with overweight BMI (β = 5.03; 95% CI: 4.52–5.55) or obese BMI (β = 9.42; 95% CI: 8.93–9.89) had higher systolic blood pressure. At the neighborhood level, in the fully adjusted individual, school, and neighborhood cross-classified model, (Table 3, Model 5) percent of students Non-Hispanic White was associated with lower systolic blood pressure (β = -0.07; 95% CI: -0.21, -0.07). Other neighborhood-level factors and school-level fixed effects were not associated with either systolic or diastolic blood pressure.

Table 4

Linear cross-classified multilevel models (CCMM) predicting diastolic blood pressure from individual-, school- and neighborhood-level factors in the National Longitudinal Study of Adolescent to Adult Health, Wave IV (WIV), 2008–2009 (N = 13,926).

Diastolic Blood Pressure (mmHg)	Model 1	Model 2	Model 3	Model 4	Model 5
	Individual Cross-Classified	Individual and School Cross-Classified	Individual and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified	Individual, School, and Neighborhood Cross-Classified
Fixed effect estimates β (95% CI)
Intercept (SE)	69.61 (66.45, 72.73)	69.98 (66.56, 73.27)	70.27 (66.91, 73.60)	70.79 (67.29, 74.21)	67.74 (64.41, 71.15)
Individual-level
Age, years (WIV)	0.40 (0.29, 0.51)	0.39 (0.29, 0.50)	0.40 (0.30, 0.51)	0.38 (0.28, 0.49)	0.35 (0.24, 0.45)
Female	-4.75 (-5.08, -4.42)	-4.74 (-5.07, -4.42)	-4.75 (-5.07, -4.42)	-4.75 (-5.08, -4.43)	-4.48 (-4.79, -4.16)
Race/ethnicity
White	REF	REF	REF	REF	REF
Black	1.02 (0.53, 1.51)	1.20 (0.65, 1.76)	0.88 (0.28, 1.49)	0.95 (0.35, 1.55)	0.76 (0.17, 1.35)
Asian	0.70 (-0.10, 1.52)	0.89 (0.03, 1.74)	0.67 (-0.16, 1.48)	0.85 (0.01, 1.72)	1.41 (0.55, 2.26)
Hispanic	-0.34 (-0.91, 0.24)	-0.20 (-0.81, 0.44)	-0.40 (-0.99, 0.19)	-0.24 (-0.86, 0.37)	-0.51 (-1.08, 0.07)
Other	0.31 (-1.14, 1.77)	0.42 (-1.11, 1.94)	0.28 (-1.26, 1.81)	0.39 (-1.14, 1.86)	0.13 (-1.35, 1.61)
Multiracial	0.08 (-0.75, 0.93)	0.19 (-0.65, 1.06)	0.06 (-0.76, 0.92)	0.16 (-0.71, 1.02)	-0.07 (-0.91, 0.74)
Parental education
Less than high school	REF	REF	REF	REF	REF
High school graduate / GED	0.28 (-0.31, 0.88)	0.26 (-0.33, 0.87)	0.32 (-0.28, 0.92)	0.29 (-0.29, 0.88)	0.21 (-0.36, 0.78)
Some college	0.01 (-0.57, 0.61)	0.05 (-0.57, 0.65)	0.10 (-0.52, 0.71)	0.08 (-0.52, 0.69)	0.12 (-0.46, 0.70)
College graduate or beyond	-0.66 (-1.28, -0.04)	-0.55 (-1.14, 0.07)	-0.48 (-1.09, 0.15)	-0.46 (-1.09, 0.16)	-0.07 (-0.67, 0.51)
Parent receipt of public assistance	0.36 (-0.25, 0.98)	0.32 (-0.29, 0.93)	0.31 (-0.31, 0.92)	0.28 (-0.33, 0.90)	0.22 (-0.37, 0.83)
Antihypertensive medication	5.29 (4.42, 6.16)	5.28 (4.41, 6.16)	5.27 (4.40, 6.14)	5.25 (4.27, 6.13)	4.05 (3.21, 4.90)
BMI, kg/m2 (WIV)
Under or Normal Weight					REF
Overweight					2.80 (2.39, 3.21)
Obese					5.99 (5.61, 6.36)
Unknown					8.04 (2.27, 13.69)
Current smoking (WIV)
No					REF
Yes					0.94 (0.60, 1.28)
Unknown					0.06 (-1.74, 1.85)
School-level, per 10%
Percent of students Non-Hispanic White		0.01 (-0.04, 0.21)		0.13 (-0.02, 0.28)	0.11 (-0.02, 0.25)
Percent of parents receiving public assistance		-0.05 (-0.46, 0.38)		-0.04 (-0.48, 0.39)	-0.11 (-0.53, 0.31)
Percent of parents with college degree		-0.22 (-0.41, -0.02)		-0.16 (-0.38, 0.04)	-0.09 (-0.29, 0.11)
Neighborhood-level, per 10%
Percent of residents Non-Hispanic White			-0.02 (-0.12, 0.09)	-0.07 (-0.19, 0.05)	-0.03 (-0.14, 0.08)
Percent of residents receiving public assistance			-0.01 (-0.36, 0.37)	0.01 (-0.38, 0.39)	0.06 (-0.30, 0.41)
Percent of residents with college degree			-0.24 (-0.43, -0.03)	-0.14 (-0.35, 0.08)	0.01 (-0.20, 0.19)
Random effect and variance estimates (95% Credible Interval) [ICC, %]
Individual	101.31 (92.3, 96.6) [0.9]	94.48 (92.1, 96.6) [0.9]	94.36 (92.1, 96.6) [0.9]	94.46 (92.3, 96.7) [0.9]	88.46 (86.42, 90.58) [0.9]
School	1 (0.5, 1.5) [0.01]	1 (0.5, 1.6) [0.01]	1 (0.6, 1.6) [0.01]	0.9 (0.4, 1.4) [0.02]	0.70 (0.34, 1.16) [0.1]
Neighborhood	0.04 (0, 0.3) [0.01]	1 (0.55, 1.6) [0.01]	1 (0.6, 1.6) [0.01]	0.8 (0.5, 1.4) [0.01]	0.07 (0.01, 0.31) [0.1]
Fit statistics (DIC)	102829.1	102826.9	102829.1	102829.69	101919.40

Results from cross-classified linear models predicting MAP during young adulthood (Wave IV) are presented in S2 Table, and were comparable to the findings from linear regressions of systolic and diastolic blood pressure (Tables 3 and 4).

Discussion

To our knowledge, this is the first study to compare the influence of individual adolescent factors, schools and neighborhoods on young adult blood pressure and hypertension outcomes simultaneously. This study adds to previous literature on contextual influences on adolescent and young adult development by exploring the relative contributions of both school-level and neighborhood-level socioeconomic characteristics to young adult blood pressure and hypertension using a school-based sample of US adolescents. We found that the variation in hypertension was largely explained at the individual level with only small but significant contributions at the school- and neighborhood-level. These results suggest that the between-level variation in hypertension was due largely to the observed individual characteristics across schools and neighborhoods, and that more of the variability in hypertension was attributable to the school-level characteristics than the neighborhood-level characteristics.

Individual-level characteristics, particularly older ages, Non-Hispanic Black race, Asian race, male sex, BMI, and current smoking, were associated with increased hypertension among young adults. These findings at the individual-level add to the literature demonstrating that individual-level risk factors in adolescents influence hypertension risk later in life. Consistent with previous evidence, we found that hypertension risk increases with age and is higher for young adult men than women, and Black compared to White young adults [5, 9]. Moreover, this is consistent with the substantive body of literature indicating that Non-Hispanic Black Americans develop hypertension earlier in life than White Americans and provides further evidence that the racial/ethnic disparities in hypertension risk factors can appear as early as adolescence [28]. There is significant evidence showing racial/ethnic disparities in hypertension among young adults are linked to disparities in obesity, physical activity, and healthcare access, among other risk factors for hypertension. Moreover, these findings may be explained by adolescents’ exposures to everyday discrimination and racism. Several studies have found associations between reports of discrimination and self-reported everyday discrimination with hypertension including a systematic review evaluating the association between perceived racial discrimination with hypertensive status and systolic, diastolic, and ambulatory blood pressure [29-31].

Similar to other studies of contextual influences on adolescents’ cardiovascular risk factors, we found that school-level influences are related to adult health outcomes [13, 15, 32–34]. Of note, at the school level, we found that having a higher proportion of non-Hispanic White students was associated with higher hypertension risk into young adulthood 14 years later at follow-up. This finding is in the opposite direction for the findings for individual-level race/ethnicity with hypertension and neighborhood-level race/ethnicity for systolic blood pressure, which suggests an increased risk for students of color. This finding also suggests that unequal conditions for adolescents at the school level may increase the risk of hypertension later in life. This finding aligns with previous research showing the intersection between the social determinants of health and disparities by race/ethnicity are rooted in structural racism that results in inequitable access to resources required for health and well-being including uneven access to quality schools, better neighborhoods, and quality medical care [35]. Moreover, exclusionary policies such as redlining have had the effect of reducing the quality of local schools. This school-level finding may also reflect influences of other attributes of adolescents’ school environment including the food environment and access to physical activity during school hours. For example, in a study using Add Health data, investigators found differences in physical activity levels in Hispanic and Non-Hispanic Black adolescents as compared to Non-Hispanic White adolescents and that these differences were largely attributable to the schools the adolescents attended [36]. Relatedly, in an adjusted analysis of Study of Cardiovascular Risks in Adolescents (ERICA) of students enrolled in public and private schools located in urban and rural areas of Brazil, investigators found that consumption of meals prepared on the school premises was associated with adolescents’ hypertension risk (OR = 0.79, 95% CI: 0.69–0.92), implying that the school food environment in adolescence may influence their cardiovascular health [37].

Limitations

This study has limitations that merit acknowledgement. First, analyses are based on a study that selected adolescents using school-based sampling resulting in a large proportion of small neighborhoods. Although 45% of neighborhoods at Wave I contained a single respondent, prior work using Add Health has indicated no issue with bias in the random effect estimates as a result of small neighborhood sizes [38]. Second, limited school and neighborhood-level measures during adolescence were available and thus this study may miss other contextual attributes at the school and neighborhood level that may influence young adult hypertension risk measures of the built environment and access to green space. Nevertheless, Add Health is one of the few large, national samples of adolescents in the US that collected school and neighborhood-level data along with follow-up into young adulthood. Data were unweighted in these analyses because complex sample weighting techniques for CCMMs are not well-established. Nonetheless, strengths of the study included a large, national sample, and longitudinal study design. Given the discordance in young adult hypertension between NHANES and Add Health studies, some have questioned the accuracy and reliability of blood pressure in Add Health. However, one study found that, compared to NHANES, Add Health’s terminal digit preference of blood pressure is infrequent, bias is low, short-term reliability is good to excellent, and comparable to that found in well-known, exam center-based studies of cardiovascular disease [2]. Therefore, our study’s findings provide further evidence that the prevalence of hypertension among Add-Health Wave-IV participants indicates an unexpectedly high risk of cardiovascular disease among U.S. young adults and deserves further scrutiny [2].

Conclusion

In conclusion, we find that adolescents’ schools and individual-level factors influence young adult hypertension, more than neighborhoods. Our study contributes to the sparse literature examining multiple contextual contributors to young adult hypertension and indicate that the individual and school-level adolescent contexts may be the most important environments. Understanding the relative importance of these various contexts is important for developing targeted interventions to reduce hypertension risk factors in adolescents, hypertension in young adulthood, and cardiovascular disease later in life. Understanding these contexts can inform implementation strategies for hypertension prevention and health promotion efforts at the individual and school levels. Our findings merit further research to better understand the mechanisms through which adolescents’ school environments contribute to adult hypertension and disparities in hypertension outcomes later in life.

Series of adjusted cross-classified multilevel models predicting hypertension (130/80) based on the 2017 ACC/AHA guidelines.

(DOCX)

Click here for additional data file.

Series of adjusted cross-classified multilevel models predicting mean arterial pressure (MAP), defined as the weighted sum of systolic and diastolic blood pressure.

(DOCX)

Click here for additional data file.

Null cross-classified multilevel models for all outcomes including hypertension (140/90), hypertension (130/80), systolic blood pressure, diastolic blood pressure, and mean arterial pressure (MAP).

(DOCX)

Click here for additional data file.

31 Jan 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Giacomo Pucci

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, all author-generated code must be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

3. Thank you for stating the following in the Competing Interests section:

"I have read the journal's policy and the authors of this manuscript have the following competing interests: Michelle Odden is a consultant for Cricket Health, Inc. The remaining authors have indicated no conflicts of interest to disclose."

We note that you received funding from a commercial source: Cricket Health, Inc

Please provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc.

Within this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript considers secondary data analysis and fitting of cross-classified multilevel models (CCMM) to AddHealth data (kind of restrictive in nature wrt. access, yet nationally representative) to evaluate hypertension risk in young adults. The study is timely, and the Bayesian statistical methods employed seems reasonable, and an intelligent use (in an attempt to model the data hierarchy via random effects, which the Bayesian paradigm elegantly accommodates!). The study was approved by the respetive IRB. My questions are as follows:

(a) AddHealth is a wonderful resource; how are missing data (responses) considered within the proposed CCMMs ?

(b) Any effort in producing a sample size/power statement with a desired effect size wrt the study goals will be highly appreciated, and helpful in determing the size of the analytical data for future analysis. The sample size statement should consider the primary response variable, and possible hierarchy, mimicking AddHealth. If a hierarchical design is not feasible, some ballpark estimates would be helpful.

(c) CCMMs are wonderful tools. However, the CCMM equation appears restrictive in terms of the random error term. Specifically, how can one guarantee that the SBPs are usually Gaussianly distributed? Provide references or arguments is support.

(d) Presence of 2 random effects terms often complicates the analysis. Can authors justify that they are better off putting 2 random effects (instead of one), via a lower DIC value? The data design suggests the hierarchy using random effects; there maybe situations where model fit may suggest otherwise.

Reviewer #2: Abdel Magid and colleagues investigate the association of individual, school and neighboorod influences with blood pressure traits and hypertension in a longitudinal study, the National Longitudinal Study of Adolescent to Adult Health.

Different Cross-classified multilevel models were built to understand whether the health outcomes are influenced by multiple social and physical contexts. I find the methods and data analysis performed in a considerable way.

However some parts of the manuscript should be improve and presented in a clearer way.

1) Abstract: Age range could be included. ICC has not been introduced.

2) "Individual, School, and Neighborhood Variables" the description of which variables are included in the school and neighboorod level should be reported in a clearer way.

3) Page 10, line 268: "were examined using two-sample t-tests for continuous variables and chi-square tests for categorical variables". It would be good to have the p-value added in the Table 1.

4) Page 10 line 280: "were fit using MLwiN (version 3.00; Birmingham, UK) via Stata’s runmlwin command". This part could integrated with the one in line 330.

5) Page 12, line 317: DBP is not written

6) Do the authors think could it be a good idea to represent some results graphically?

7) Pag 13, line 341: Is the number reported as hypertensive correct?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

4 Mar 2022

Responses are in bold and changed text is indicated by italics, below

JOURNAL REQUIREMENTS

PLOS ONE’s STYLE REQUIRMENTS

• Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

o To adhere to PLOS One’s style requirements, we have made the following changes: bold type 18pt font for level 1 headings, bold type 16pt font for level 2 headings, manuscript title in sentence case, removed author’s titles, spelled out “USA” to “United States of America”, removed corresponding author’s physical address and kept corresponding author’s email.

CODE SHARING

• Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, all author-generated code must be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

o In accordance with PLOS ONE’s specific guidelines on code sharing we have shared our author-generated code underpinning the findings in this manuscript on Open Science Framework. This code can be found at: https://osf.io/86fpn/?view_only=57084773c8b642a497c1d6c8806ab0ee

COMPETING INTERESTS SECTION

• Thank you for stating the following in the Competing Interests section: "I have read the journal's policy and the authors of this manuscript have the following competing interests: Michelle Odden is a consultant for Cricket Health, Inc. The remaining authors have indicated no conflicts of interest to disclose."We note that you received funding from a commercial source: Cricket Health, Inc
Please provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc. Within this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared. Please include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf

o We confirm that this competing interest does not alter our adherence to all PLOS ONE policies on sharing data and materials. The “Potential Conflicts of Interests” section now reads as: “Potential Conflicts of Interest: Michelle Odden is a consultant for Cricket Health, Inc. The remaining authors have indicated no conflicts of interest to disclose. This does not alter our adherence to PLOS ONE policies on sharing data and materials.” This updated statement is also included in our cover letter.

REVIEWER COMMENTS

REVIEWER #1

• The manuscript considers secondary data analysis and fitting of cross-classified multilevel models (CCMM) to AddHealth data (kind of restrictive in nature wrt. access, yet nationally representative) to evaluate hypertension risk in young adults. The study is timely, and the Bayesian statistical methods employed seems reasonable, and an intelligent use (in an attempt to model the data hierarchy via random effects, which the Bayesian paradigm elegantly accommodates!). The study was approved by the respetive IRB. My questions are as follows:

o Thank you for your succinct summary and thoughtful, insightful review. We appreciate your careful review of our research.

• (a) AddHealth is a wonderful resource; how are missing data (responses) considered within the proposed CCMMs ?

o We conducted a complete case analysis and note in the Methods how many indviduals were dropped. Total indviduals in Wave IV is 15,701 and we include 13,911 in our analysis. In the final sample, there was no additional missing data.

• (b) Any effort in producing a sample size/power statement with a desired effect size wrt the study goals will be highly appreciated, and helpful in determing the size of the analytical data for future analysis. The sample size statement should consider the primary response variable, and possible hierarchy, mimicking AddHealth. If a hierarchical design is not feasible, some ballpark estimates would be helpful.

o We found small effect sizes for both school and neighborhood context (intercept variance parameters). A larger sample size may have allowed us to detect small effect sizes, though we were constrained by the Add Health sample size and school and neighborhood structure. Literature on sample size for CCMMs are sparse though one study (Chung 2018; linked below) indicated good credible interval coverage across a wide range of number of higher level units, group sizes within those units, and extent of cross-classification. Per these guidelines, Add Health has a very large number of neighborhoods (n=1900) and a sufficient number of schools (n=128). Future studies using similar datasets with non-hierarchical nesting of students within schools and neighborhoods may have more variability at the contextual level and would therefore require lower sample sizes. Citation: Chung, H., Kim, J., Park, R. and Jean, H., 2018. The impact of sample size in cross-classified multiple membership multilevel models. Journal of Modern Applied Statistical Methods, 17(1), p.26. https://digitalcommons.wayne.edu/cgi/viewcontent.cgi?article=2491&context=jmasm

• (c) CCMMs are wonderful tools. However, the CCMM equation appears restrictive in terms of the random error term. Specifically, how can one guarantee that the SBPs are usually Gaussianly distributed? Provide references or arguments is support.
I

o In our sample, SBP and DBP are very normally distributed (mean and median nearly identical, SDs are not large compared to the means, etc.). For example, SBP mean=124.5 (SD=13.6) vs. median=123.5DBP mean=79.0 (SD=10.2) vs. median=78.5. We acknowledge that SBP are not necessarily usually Gausianly distributed in adult populations and usually have a wide range. Moreover, in other samples, it may be necessary to account for non-normality by transforming BP values in order to use linear regression or to use a different distribution in the model. Neverthless, in our specific case the age range is relatively small and constrained to young adulthood where things are a bit less variable/skewed.

• (d) Presence of 2 random effects terms often complicates the analysis. Can authors justify that they are better off putting 2 random effects (instead of one), via a lower DIC value? The data design suggests the hierarchy using random effects; there maybe situations where model fit may suggest otherwise.

o Though it is usually not advised to compare model fit between multilevel models where the number of levels/model structurure differs (it is okay to compare models with the same structure but differing predictors). Nevertheless, given these models are fit using maximum likelihood, we have included the DIC values in Supplementary Table 3 for the crude cross-classified multilevel models for all outcomes. The best fitting null models were a bit variable, but given the data structure and our research objective of examinining the simultaneous impact of school and neighborhood, we used the cross classified models across all outcomes. Additionally, while the neighborhood ICCs are realtively small, they are not zero, and therefore an important context to account for as we have in our study. Moreover, there is a signfigant amount of literature regarding the importance of comprehensively accountying for the study design such as we have done with the linear and logistic CCMMs. As included in our citation of Dunn et al Health and Place 2015, ignoring one of the levels is equated to misattributing the variance to the level that was included in the analysis (i.e. school-only ‘absorbs’ the neighborhood effect and makes it seem like it is a school level effect. Once you compare to the CCMM model, you can see the variance is parsed between school and neighborhood). The citation below further supports the importance of comprehensively accountying for the study design with CCMMs. While we acknowledge that this matters more and is more apparent when the variance is larger, this is also more apparent in linear models (such as in SBP and DBP models included in this study). Moreover, as shown in supplementary table 3, according to the DIC values the CCMM is best fitting for all outcomes except the 140/90 definition of hypertension. Therefore, this provides further evidence that the CCMMs were better on the whole.

o Citation: Ren, W., 2011. Impact of design features for cross-classified logistic models when the cross-classification structure is ignored. The Ohio State University. https://etd.ohiolink.edu/apexprod/rws_etd/send_file/send?accession=osu1322538958&disposition=inline

REVIEWER #2

• Abdel Magid and colleagues investigate the association of individual, school and neighboorod influences with blood pressure traits and hypertension in a longitudinal study, the National Longitudinal Study of Adolescent to Adult Health.
Different Cross-classified multilevel models were built to understand whether the health outcomes are influenced by multiple social and physical contexts. I find the methods and data analysis performed in a considerable way.
However some parts of the manuscript should be improve and presented in a clearer way.

o Thank you for your succinct summary and thoughtful, insightful review. We appreciate your careful review of our research.

• 1) Abstract: Age range could be included. ICC has not been introduced.

o We have included the following updated sentence in the abstract “Data were analyzed from the National Longitudinal Study of Adolescent to Adult Health (1994-1995 ages 11-18 and 2007-2008 aged 24-32).” The age ranges in each Add Health wave are detailed here: https://addhealth.cpc.unc.edu/documentation/study-design/ We also added the following sentence in the Abstract-Methods section “For linear models, intra-class correlations (ICC) are reported for random effects.

• 2) "Individual, School, and Neighborhood Variables" the description of which variables are included in the school and neighboorod level should be reported in a clearer way.

o We have reported the description of the three levels of variables in three sections in the methods section. This section now reads as:

“Individual Variables

We constructed individual covariates using data from the Wave I in-home interview, including adolescents’ biological sex (male, female), race/ethnicity (non-Hispanic Black, Hispanic, Asian and Pacific Islander, Other, Multiracial, and non-Hispanic White). At the individual-level, SES was determined based on parental education and receipt of public assistance. We used data from either the youth or caregiver interview to capture receipt of public assistance (mother currently receiving public assistance, such as welfare or not) and highest level of parental education (defined as the maximum level of education by the resident mother, resident father, or resident step-father/partner (no high school diploma or equivalent; completed high school or equivalent; completed some college, trade school or a 2-year degree; completed equivalent 4-year college degree or above). Height and weight were measured by trained interviewers at Wave IV. Young adult body mass index (BMI) at Wave IV was calculated as the ratio of weight in kilograms over height in meters squared. Age at Wave IV (in years) was calculated from the date of Wave IV in-home interview and participant’s date of birth.

School Variables

We constructed school-level covariates using data from the Wave I data. Using the survey of the full sample of schools, at the school-level, we created a continuous measure of school-level SES by aggregating individual-level data. Use of individual-level data was required as information about school-level SES was not directly available. We calculated the proportion of students within each school whose mother had received public assistance or had a college degree.

Neighborhood Variables

We constructed neighborhood-level covariates using data from the Wave I data. At the neighborhood level, we used data from the 1990 Census to create a neighborhood-level SES measure indicating the proportion of residents within each neighborhood who had received public assistance or had a college degree. We also calculated the proportion of students in either the school or the neighborhood who were White.”

• 3) Page 10, line 268: "were examined using two-sample t-tests for continuous variables and chi-square tests for categorical variables". It would be good to have the p-value added in the Table 1.

o We have included p-values in table 1.

• 4) Page 10 line 280: "were fit using MLwiN (version 3.00; Birmingham, UK) via Stata’s runmlwin command". This part could integrated with the one in line 330.

o Sentences in 280 and 330 have now been combined at the end of Statistical Analysis section and now reads “Models were fit using MLwiN (version 3.00; Birmingham, UK) via Stata’s runmlwin command. MLwiN uses Bayesian estimation procedures using Markov Chain Monte Carlo (MCMC) methods with non-informative priors and a Metropolis-Hastings sampling algorithm allowing for simultaneous modeling of non-hierarchically nested contexts.24-27All univariate and bivariate analyses were preformed using Stata version 16 (College Station, TX).”

• 5) Page 12, line 317: DBP is not written

o This sentence now reads: “All models for MAP, DBP, and SBP additionally adjusted for self-reported use of antihypertensive medications.”

• 6) Do the authors think could it be a good idea to represent some results graphically?

o Given that the school and neighborhood variation contributions are less than <1.5%, we do not think representing the results graphically would add much to the results above and beyond what is included in the tables.

• 7) Pag 13, line 341: Is the number reported as hypertensive correct?

o We have double checked this number, and yes it is correct: “Of the 13,911 Wave IV participants included in this study, 7,111 (51%) young adults were classified as hypertensive.” We would like to note that this number is based on the more consertavitve, newer definition of >130/80. We examined both commonly used thresholds for hypertension included 130/80 and 140/90. The number hypertensive has also been found in the literature, including in this AddHealth study published in PLoS One: Nagata JM, Ganson KT, Cunningham ML, et al. Associations between legal performance-enhancing substance use and future cardiovascular disease risk factors in young adults: A prospective cohort study. PLoS One. 2020;15(12):e0244018. Published 2020 Dec 15. doi:10.1371/journal.pone.0244018.

Submitted filename: Cover and Rebuttal Letter_upl.docx

Click here for additional data file.

28 Mar 2022

Adolescent Individual, School, and Neighborhood Influences on Young Adult Hypertension Risk

PONE-D-21-33507R1

Dear Dr. Nagata,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Giacomo Pucci

Academic Editor

PLOS ONE

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: Thanks to the authors to answer clearly and exhaustively all my comments.

I have no further issue to raise.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

5 Apr 2022

PONE-D-21-33507R1

Adolescent individual, school, and neighborhood influences on young adult hypertension risk

Dear Dr. Nagata:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Giacomo Pucci

Academic Editor

PLOS ONE