Abstract
Purpose
To examine the prospective, longitudinal associations between positive well-being during adolescence and health outcomes in young adulthood, using a large, nationally representative sample of youth.
Methods
Based upon data from the first three waves of the National Longitudinal Study of Adolescent Health (Add Health), we examined positive well-being during adolescence (averaged across Waves I-II) as a predictor of perceived young adult general health and risky health behaviors (Wave III). Each model included a full set of health and demographic baseline covariates. Missing values were assigned using multiple imputation methods (N = 10,147).
Results
Positive well-being during adolescence was significantly associated with reporting better perceived general health during young adulthood, independent of depressive symptoms. Positive well-being was also significantly associated with fewer risky health behaviors in Wave III, after adding all covariates, including depressive symptoms and baseline risky health behaviors.
Conclusion
Few studies of adolescent health have examined positive psychological characteristics, tending to focus instead on the impact of negative mood states and cognitions on health. This study demonstrates that positive well-being during adolescence predicts better perceived general health and fewer risky health behaviors during young adulthood. Aligned with the goals of the positive youth development perspective, promoting and nurturing positive well-being during the transition from childhood to adolescence may present a promising way to improve long-term health.
The past few decades have witnessed significant advances towards an integrative approach to studying health and development. Behavioral, psychological, and social processes are fundamental to a comprehensive understanding of health, disease resistance, and wellness [1]. Scholars have identified multiple determinants of health across the lifespan including genetics, health behaviors, access to health care, and social factors such as personal ties, socioeconomic inequality, and community influences [2]. However, one determinant of health that remains understudied is positive well-being, a construct that encompasses positive emotion and expectations for the future, perceived social acceptance, and high self-esteem [3–5].
Pathways from Positive Well-Being to Good Health
Over the past decade, researchers have found important associations between positive well-being and improved immune, cardiovascular, and neuroendocrine functioning, superior cognitive performance, and longevity [4–8]. It appears that happiness, positive emotions, and self-esteem might play a key role in stress biology, reducing the activation of the hypothalamic-pituitary-adrenal (HPA) axis and increasing cellular immune competence [3–5]. Another possible pathway stems from Fredrickson’s broaden-and-build theory, which suggests that positive emotions broaden people’s thoughts, actions and behaviors, while negative states heighten people’s autonomic activity and narrow their attention [7]. Indeed, neuropsychological evidence suggests that positive affect is associated with broad, flexible, creative and efficient patterns of thought; effects linked to increases in brain dopamine levels [9].
Other studies have found that positive well-being may help to enhance healthy outcomes via social and behavioral pathways [10]. Cross-sectional evidence has shown that positive well-being is associated with fewer health-compromising and risky behaviors, more physical activity, and better sleep quality [4, 5, 11]. Positive well-being may also have indirect effects on health through social mechanisms, since happy individuals have more friends, more social support and better interpersonal relationships [4], all of which have positive implications for health [12].
Despite the growing interest in positive health, few studies have examined the long-term effects of positive well-being. Additionally, almost all studies to date have focused exclusively on middle aged and elderly adult populations, where disease and mortality rates are highest. The goal of the present study is to examine the prospective, longitudinal associations between positive well-being during adolescence and health outcomes in young adulthood, using a large, nationally representative sample of youth.
Positive Well-Being in Adolescence
While positive well-being has potentially important implications across the lifespan, it may be a particularly salient factor for health trajectories from adolescence into young adulthood. For instance, developmental models of health, such as the Life Course Health Development (LCHD) framework [2], emphasize the role of risk and protective factors during developmental transitions and turning points [13, 14]. Adolescence is a period of rapid developmental change, characterized by growing autonomy and increased involvement in one’s own health through health-compromising or health-promoting behaviors [15]. During this important developmental transition, youth begin to choose their own environments and make their own decisions that will set trajectories for adult health and health behaviors. Adolescence, therefore, represents a sensitive developmental period in which psychological characteristics such as positive well-being may play a pivotal role in physical health.
Historically, the scientific study of adolescent development has defined this stage of life as “storm and stress” or as an ontogenetic disturbance of normative development [16, 17]. In fact, until recently, positive well-being in adolescence was only discussed as the absence of negative or undesirable behaviors [18]. It was not until the 1990s, with the beginning of the positive youth development perspective, that a new vision emerged for discussing young people as resources to be developed, and not as problems to be managed [19]. Positive youth development programs foster healthy development by promoting positive qualities such as self-determination, bonding, prosocial involvement, optimism, self-efficacy, and positive identity [20]. The current study seeks to contribute to this paradigm shift from deterrence to development [21] by examining the longitudinal health consequences of positive well-being.
Positive Well-Being and Depressive symptoms
An examination of the association between positive well-being and health must address the role of depression and depressive symptoms. Is positive well-being a distinct dimension of psychological health or does it simply reflect the absence of depressive symptoms [22]? The later view sees well-being and ill-being as opposite ends of a bipolar continuum [23]. Therefore, the possibility that positive emotional states are associated with good health may be explained as merely the absence of negative affect or depression. However, research suggests that well-being (e.g. happiness, self-esteem, optimism) and ill-being (e.g. depression, anxiety, pessimism) are distinct dimensions of emotional functioning, and thus positive well-being could provide health benefits independent of the negative outcomes associated with depression or depressive symptomology [24]. Additionally, studies within the past decade have found that positive affect is associated with lower risk of illness and mortality, independent of the effects of negative affect [5, 22].
Research on the long-term effects of positive well-being, particularly with child or adolescent samples, is sparse. Depressive symptoms have been linked to health problems in adolescent and young adult populations including overall poor health, health-risk behaviors, obesity, asthma, diabetes and other physical health problems [25–27]. However, none of these studies examined positive well-being as a contributing factor to health, leaving the literature ambiguous regarding the relative contributions of well-being and ill-being [22]. In order to examine the unique effects of positive well-being, this study will test whether positive well-being is associated with perceived general health and risky health behaviors, controlling for depressive symptoms.
Design and Method
The data are drawn from the first three waves of the National Longitudinal Study of Adolescent Health (Add Health), a nationally representative study of adolescents in the United States. The study used a school-based design to select a stratified sample of 80 high schools (and feeder middle schools) with selection probability proportional to size. Wave I (1994–1995) included 20,745 adolescents (aged 11–20) and their parents. All adolescents in grade 7 through 11 in Wave I were targeted about one year later for the Wave II in-home interview (n=14,738). A third wave of data collection (Wave III) took place from 2000–2001 (ages 18–27), for all respondents from the Wave I in-home sample (n=15,197). A more detailed description of the Add Health survey design can be found elsewhere [28]. The final analytic sample for the current study (N=10,147) included those who participated in the in-home surveys in all three waves. This limited the sample to adolescents in grades 7 to 11, since 12th grade students were not followed in Wave II.
Measures
Outcome measures
The dependent variables measured self-reported general health and health behaviors of the young adults in Wave III. General health was measured by a single question, “In general, how is your health?” (1=poor; 2=fair; 3=good; 4=very good; 5=excellent). To measure health-compromising behaviors, we created a risky health behavior index using young adults’ responses to six categories: low physical activity, fast food consumption, binge drinking, cigarette smoking, marijuana use, and other illicit drug use.
The variable for engagement in physical activity was created from seven questions asking about the frequency of various activities in the past week (e.g. cycling, dancing, team sports, running). Fast food consumption was measured as the number of days the respondent ate fast food in the past week. Cigarette smoking was measured by the number of days that a person smoked at least one cigarette over the past 30 days. Binge drinking was defined as consuming five or more drinks in a row. Respondents were asked how many days they participated in binge drinking over the past year. Responses in the lowest quartile for physical activity (one or fewer activities per week) and the highest quartiles for fast food (4 or more meals per week), cigarette smoking (more than 20 days per month), and binge drinking (more than once a month) each counted towards this risky health behavior index.
Respondents were also asked how many times during the past 30 days they used: 1) marijuana; 2) cocaine; 3) crystal meth; 4) other illegal drugs such as LSD, PCP, ecstasy, mushrooms, inhalants, ice, heroin, or prescription medicines not prescribed for them. One or more uses of marijuana (22% of sample) and one or more uses of any other illegal drugs in the past 30 days (7% of sample), each contributed one count towards the risky health behavior index. The final risky health behavior index ranges from 0 (low risk) to 6 (high risk), based on the six categories described above.
Positive well-being
This scale was created by taking a unit weighted average of the following ten questions in each Wave I (α=.849) and Wave II (α=.851). Since these ten questions have different response categories, the scale was constructed by using standardized values of each item. Four questions came from the Center for Epidemiologic Studies Depression Scale (CES-D) and asked about the following feelings experienced in the past week: happiness, enjoyment of life, felt just as good as other people, and hopefulness about the future (0=never or rarely; 1=sometimes; 2=a lot of the time; 3=most of the time). In past research, these four items have loaded onto a single positive affect factor [29], and have been used in other studies to measure positive affect/mood (happiness, enjoyment) as well as optimism and self-esteem [22].
The additional six questions asked respondents whether they have good qualities, have a lot to be proud of, like themselves as they are, always do things right, feel socially accepted, and feel loved and wanted (1=strongly disagree to 5=strongly agree). Most of these later items were adapted from the Rosenberg Self-Esteem scale [30] and they have been used together in other Add Health studies to measure self-esteem or self-image [31, 32]. In a series of principal component analyses (with varimax rotation) we found that the positive well-being items (positive CES-D items and self-esteem items) and negative CES-D items factor separately, when the number of components is constrained to two. However, in an unconstrained model, we observe one factor for positive CES-D items, one factor for self-esteem items, and four factors for negative CES-D items (representing different subtypes of depressive symptoms). Therefore, we also examined the independent contributions of the positive affect subscale (4 items; α=.725) and the self-esteem subscale (6 items; α=.855).
Health and demographic covariates
The remaining CES-D items were used to create the depressive symptoms scale, a unit-weighted average of the following feelings experienced in the past week: felt blue; bothered by things that don’t usually bother you; frequent crying; felt depressed, had trouble keeping mind on things; did not enjoy life; did not feel just as good as other people; felt disliked by people; felt sad; and felt too tired to do things (Wave I α=.861; Wave II α=.863). As expected, the positive well-being and depression scales were significantly negatively correlated (r = −.553, p < .001).
We also included a number pre-existing physical and mental conditions measured at Wave I in all analyses, including self-reported general health (1=poor; 2=fair; 3=good; 4=very good; 5=excellent), body mass index quartiles, a count (0–14) of nonspecific somatic symptoms experienced at least once per week (e.g. headaches, fatigue, stomach aches, pain in muscles or joints), and a number of binary indicators for diagnosis of asthma, heart problems, and physical limitations, as self-reported in Wave I. The analyses also include parents’ self-reported of obesity (yes/no) and all risky health behaviors at baseline (see Table 1). Additionally, basic demographic data were dummy coded and included as covariates: age, gender, race/ethnicity, parental marital status (two biological parents; two parents, one biological; single mother; single father; other), and mean parental education (less than high school, high school graduate/GED, some college, college graduate, and post graduate). Finally, we included a binary measure for health insurance and a continuous measure of neighborhood SES (proportion of families living below the poverty line).
Table 1.
Weighted descriptive statistics for outcomes, positive well-being, and key covariates (N = 10,147).
| Mean | S.D. | Min. | Max. | |
|---|---|---|---|---|
| Outcome Variables (Wave III) | ||||
| Perceived general health | 3.989 | .866 | 1 | 5 |
| Risky Health Behavior Index | 1.375 | 1.236 | 0 | 6 |
| Physical activity in past week | 5.968 | 6.211 | 0 | 49 |
| Fast food meals in past week | 2.483 | 2.090 | 0 | 7 |
| Binge drinking in past year | 29.615 | 54.728 | 0 | 356 |
| Number of days smoked in past 30 days | 8.834 | 13.100 | 0 | 30 |
| Marijuana use in past 30 days | 3.477 | 9.749 | 0 | 60 |
| Hard drug use in past 30 days | .517 | 3.256 | 0 | 60 |
| Positive Well-Being (Avg. Waves I-II) | ||||
| Good qualities | 4.298 | .541 | 1.5 | 5 |
| Do Right | 3.823 | .732 | 1 | 5 |
| Feel socially accepted | 4.128 | .632 | 1 | 5 |
| Feel loved | 4.326 | .589 | 1 | 5 |
| Proud | 4.334 | .589 | 1 | 5 |
| Like self | 4.046 | .810 | 1 | 5 |
| Feel just as good as others | 1.955 | .823 | 0 | 3 |
| Hopeful | 1.830 | .795 | 0 | 3 |
| Happy | 2.150 | .657 | 0 | 3 |
| Enjoy life | 2.246 | .698 | 0 | 3 |
| Depressive symptoms scale (Avg. Waves I-II) | ||||
| Bothered | .491 | .545 | 0 | 3 |
| No appetite | .457 | .571 | 0 | 3 |
| Blue | .374 | .556 | 0 | 3 |
| Distracted | .811 | .653 | 0 | 3 |
| Depressed | .485 | .611 | 0 | 3 |
| Too tired to do anything | .724 | .600 | 0 | 3 |
| Failure | .190 | .410 | 0 | 3 |
| Fearful | .355 | .572 | 0 | 7 |
| Talk less | .550 | .560 | 0 | 3 |
| Lonely | .425 | .554 | 0 | 3 |
| Others seem unfriendly | .408 | .497 | 0 | 3 |
| Sad | .548 | .553 | 0 | 3 |
| Felt disliked by others | .398 | .510 | 0 | 3 |
| Frequently cry | .434 | .582 | 0 | 4 |
| Hard time starting things | .621 | .558 | 0 | 3 |
| Life is not worth living | .141 | .371 | 0 | 3 |
| Demographic Controls (Wave I) | ||||
| Age | 14.934 | 1.542 | 11 | 20 |
| Grade | 8.932 | 1.397 | 7 | 11 |
| Male | .503 | .498 | 0 | 1 |
| White | .677 | .466 | 0 | 1 |
| Parents’ average education (years) | 12.961 | 2.619 | 6 | 18 |
| Proportion of poor families in neighborhood | .120 | .129 | 0 | .859 |
| Baseline Health Controls (Wave I) | ||||
| Perceived general health | 3.887 | .899 | 1 | 5 |
| BMI | 22.228 | 4.497 | 11.478 | 56.378 |
| Risky Health Behavior Index | 1.213 | 1.246 | 0 | 6 |
| Physical activity in past week | 5.151 | 2.535 | 0 | 10.5 |
| Fast food meals in past week | 2.145 | 1.759 | 0 | 7 |
| Binge drinking in past year | 10.456 | 44.861 | 0 | 365 |
| Number of days smoked in past 30 days | 3.910 | 9.088 | 0 | 30 |
| Number of marijuana uses in past 30 days | 1.246 | 5.984 | 0 | 60 |
| Number of hard drug uses in past 30 days | .484 | 4.037 | 0 | 60 |
Analytic Strategy
The general analytic strategy was to investigate young adult perceived general health and risky health behaviors (Wave III) as a function of positive well-being during adolescence (averaged across Waves I-II), covarying for depressive symptoms (also averaged across Waves I-II), and a wide range of baseline variables at Wave I. Because causality between positive well-being and health goes in both directions and the effects accumulate over a long period of time, prospective studies are best suited for investigating the linkage between well-being and health outcomes [33].
Analyses were conducted in Stata 11.0 [34] using survey procedures and appropriate Add Health longitudinal sampling weights, which adjust for the complex cluster sample design, unequal probability of selection, and non-response [35]. Missing values on all variables were assigned using multiple imputation methods (with multivariate normal regression to impute multiple variables simultaneously). All independent variables, outcome variables, strata, and sampling weights were used as predictors in the imputation process [34, 36]. For all categorical variables, missing values were assigned to the category with the highest imputed value [36]. A total of five imputations were used to obtain final estimates. In all analyses, continuous covariates were standardized and all coefficients were converted into odds ratios for interpretation.
Both outcomes were originally modeled using ordinal logistic regression based on a 5-level general health variable (excellent; very good; good; fair; poor) and a 5-level risky health behavior index (0; 1; 2; 3; 4 or more risky health behaviors)1. However, since both of these ordered logit models violated the proportional odds assumption2, we ran a series of logit models to capture the differential effects of positive well-being on levels of perceived general health and number of risky health behaviors. Consistent with previous research [37], most young adults in our sample are healthy, therefore we focus on the top category of general health for our analyses (i.e. excellent perceived health compared to very good, good, fair, or poor health).
Results
Descriptive Statistics
On average, participants rated their health to be very good in both Wave I (mean = 3.887, S.D. = .899) and Wave III (mean = 3.989, S.D. = .866). Youth reported, on average, 1.213 (S.D. = 1.246) risky health behaviors in Wave I and 1.375 (S.D. = 1.236) risky health behaviors at Wave III. Table 1 presents descriptive statistics for the positive well-being and depressive symptom variables, as well as a number of baseline health and demographic covariates.
Outcome 1: Perceived General Health
Table 2 shows the results from a logistic regression examining the association between positive well-being and perceived general health. Positive well-being during adolescence (Wave I-II) was significantly associated with reporting excellent health during young adulthood (Wave III). For a one standard deviation increase in the positive well-being scale, we observed the odds of reporting excellent general health versus all lower categories (very good/good/fair/poor health) combined to be 30% greater, covarying for the effects of depressive symptoms, baseline health, and demographic characteristics (OR = 1.297, p < .001). Depressive symptoms were significantly associated with reduced odds in reporting excellent health. For a one standard deviation increase in the depressive symptom scale, the odds of reporting excellent health in Wave III was 8% less than all lower categories of general health combined, after adding covariates (OR = .919, p < .001).
Table 2.
Logistic regressions predicting perceived general health at Wave III from positive well-being, depressive symptoms, and baseline covariates, reported in odds ratios (N = 10,147).
| Perceived Excellent General Health WIII (Compared to all lower categories combined: Very good; Good; Fair; Poor ) | ||||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| Positive Well-Being Scale (WI-II) | 1.297** | |||
| Positive Affect Subscale | 1.277** | 1.251** | ||
| Self-Esteem Subscale | 1.151** | 1.062 | ||
| Depressive symptoms scale (WI-II) | .919** | .894** | .873** | .917* |
| Perceived General Health WI (Very good)a | ||||
| Excellent health | 2.255** | 2.276** | 2.410** | 2.273** |
| Good health | .643** | .639** | .622** | .636** |
| Fair health | .469** | .454** | .439** | .460** |
| Poor health | .330 | .341 | .304 | .342 |
| Risky Health Behaviors WI | ||||
| Physical activity | .918* | .932* | .937† | .927* |
| Fast food | .967 | .975 | .979 | .975 |
| Binge drink | 1.049 | 1.033 | 1.035 | 1.03 |
| Smoke | .935 | .937 | .937 | .940 |
| Marijuana | .984 | .991 | .989 | .989 |
| Hard drugs | .981 | .987 | .982 | .987 |
| Health Controls WI | ||||
| Symptom count | .956 | .961 | .962 | .959 |
| Cardiac abnormality | .081* | .086** | .080** | .085** |
| Asthma | .907 | .938 | .944 | .936 |
| Physical limitation/disability | .635† | .634† | .634† | .613† |
| Body Mass Index (1st quartile)a | ||||
| BMI 2st quartile | 1.142 | 1.166† | 1.144† | 1.159† |
| BMI 3rd quartile | .960 | 1.022 | 1.010 | 1.021 |
| BMI 4th quartile | .712** | .742** | .730** | .741** |
| Parent(s) with obesity | .919 | .898 | .909 | .899 |
| Health insurance | 1.108 | 1.134 | 1.152 | 1.135 |
| Other WI controls | ||||
| Gender (1=male) | 1.287 | 1.268** | 1.315** | 1.279** |
| Race/ethnicity (White)a | ||||
| Black | 1.186† | 1.183† | 1.282** | 1.196* |
| Native American | 1.726* | 1.854* | 1.794* | 1.857* |
| Asian | .826 | .849 | .852 | .857 |
| Hispanic | 1.262* | 1.249* | 1.264* | 1.262* |
| Other race | 1.408 | 1.345 | 1.445 | 1.356 |
| Parents mean education level (high school)a | ||||
| Less than high school | 1.064 | 1.044 | 1.077 | 1.069 |
| Some college | .991 | 1.003 | .989 | 1.002 |
| College | 1.059 | 1.083 | 1.059 | 1.080 |
| More than college | 1.355** | 1.391** | 1.327** | 1.370** |
Notes:
p < 0.01,
p < 0.05,
p < 0.1,
Reference group in parenthesis.
Also controlling for age, neighborhood poverty, and family structure.
A higher score on each the positive affect (Model 2) and self-esteem (Model 3) subscales were significantly associated with better perceived general health in Wave III when analyzed separately, following the same trends as the positive well-being scale, with smaller effect sizes. However, when examined in the models together (Model 4), only the positive affect subscale predicts reporting excellent health compared to all lower categories of general health (OR = 1.251, p < .001).
Outcome 2: Risky Health Behavior Index
Positive well-being was also significantly associated with reporting fewer risky health behaviors in Wave III. For a one standard deviation increase in the positive well-being score, we observed the odds of reporting four or more risky health behaviors versus all lower categories combined (three or fewer risky behaviors) to be 17% lower, covarying for baseline risky health behaviors and the full set of covariates (OR=.833, p < .001). Similar effect sizes were reported across all models, as seen in Table 3. Additional analyses examining the positive affect and self-esteem subscales show that the positive affect subscale predicted fewer risky health behaviors across all levels of the risky health behavior index (see Table 4). The self-esteem subscale and the depressive symptoms scale, however, were not associated with risky health behaviors at Wave III.
Table 3.
Serial logit regressions predicting risky health behaviors at Wave III from positive well-being, depressive symptoms, and baseline covariates, reported in odds ratios (N = 10,147).
| Risky Health Behavior Index WIII (Versus all Lower Levels on the Index) | ||||
|---|---|---|---|---|
| 4 or More Risky Health Behaviors | 3 or More Risky Health Behaviors | 2 or More Risky Health Behaviors | 1 or More Risky Health Behaviors | |
| Positive Well-Being Scale (WI-II) | .833** | .859** | .870** | .904* |
| Depressive symptoms scale (WI-II) | .903 | .928 | .998 | 1.040 |
| Perceived General Health WI (Very good)a | ||||
| Excellent health | .887 | .801* | .880† | .857* |
| Good health | 1.113 | .981 | .956 | 1.069 |
| Fair health | .891 | .838 | .773† | .975 |
| Poor health | .828 | .772 | .771 | 1.078 |
| Risky Health Behaviors WI | ||||
| Physical activity | .962 | 1.007 | .981 | .948 |
| Fast food | 1.055 | 1.135** | 1.164** | 1.208** |
| Binge drink | 1.042 | 1.017 | 1.071* | 1.027 |
| Smoke | 1.322** | 1.436** | 1.551** | 1.646** |
| Marijuana | 1.253** | 1.212** | 1.215** | 1.154* |
| Hard drugs | .950 | .955 | .907* | .914* |
| Health Controls WI | ||||
| Symptom count | 1.127 | 1.112** | 1.057 | 1.023 |
| Cardiac abnormality | 1.159** | .504 | .816 | 3.762 |
| Asthma | .987 | .977 | .998 | .988 |
| Physical limitation/disability | .833 | .695 | .964 | 1.007 |
| Body Mass Index (1st quartile)a | ||||
| BMI 2st quartile | .991 | .857 | .926 | .881 |
| BMI 3rd quartile | .888 | .841† | .933 | .798* |
| BMI 4th quartile | 1.044 | .852 | .987 | .787** |
| Parent(s) with obesity | .822 | .841† | .861* | .937 |
| Health insurance | 1.170 | 1.008 | 1.017 | 1.081 |
| Other WI controls | ||||
| Gender (1=male) | 1.932** | 1.963** | 1.964** | 1.805** |
| Race/ethnicity (White)a | ||||
| Black | .312** | .483** | .678 | .973 |
| Native American | .592 | .754 | .624 | .870 |
| Asian | .244** | .493** | .460** | .657** |
| Hispanic | .564** | .551** | .559** | .680** |
| Other race | .744 | .635 | .703 | .992 |
| Parents mean education level (high school)a | ||||
| Less than high school | .797 | .991 | .866 | .918 |
| Some college | 1.274 | 1.273* | .995 | .927 |
| College | 1.214 | 1.241† | 1.058 | .908 |
| More than college | .890 | .929 | .914 | .752 |
Notes:
p < 0.01,
p < 0.05,
p < 0.1,
Reference group in parenthesis.
Also controlling for age, neighborhood poverty, and family structure.
Table 4.
Serial logit regressions predicting risky health behaviors at Wave III from positive affect, self-esteem, depressive symptoms, and baseline covariates, reported in odds ratios (N = 10,147).
| Risky Health Behavior Index WIII (Versus all Lower Levels on the Index) | ||||
|---|---|---|---|---|
| 4 or More Risky Health Behaviors | 3 or More Risky Health Behaviors | 2 or More Risky Health Behaviors | 1 or More Risky Health Behaviors | |
| Positive Affect Subscale | .809** | .892* | .893** | .916* |
| Self-Esteem Subscale | 1.036 | .961 | .958 | .990 |
| Depressive symptoms scale (WI-II) | .930 | .948 | 1.007 | 1.059 |
| Perceived General Health WI (Very good)a | ||||
| Excellent health | .892* | .799* | .889 | .877† |
| Good health | 1.098 | .944 | .933 | 1.082 |
| Fair health | .868 | .836 | .774† | .943 |
| Poor health | .801 | .799 | .818 | .938 |
| Risky Health Behaviors WI | ||||
| Physical activity | .954 | 1.009 | .480 | .948 |
| Fast food | 1.050 | 1.142** | 1.172** | 1.225** |
| Binge drink | 1.047 | 1.021 | 1.075* | 1.037 |
| Smoke | 1.295** | 1.429** | 1.541** | 1.643** |
| Marijuana | 1.230** | 1.205** | 1.250** | 1.218** |
| Hard drugs | .951 | .956 | .891** | .889* |
| Health Controls WI | ||||
| Symptom count | 1.155** | 1.100* | 1.043 | 1.014 |
| Cardiac abnormality | .924 | .921 | .547 | 3.358 |
| Asthma | .824 | .974 | 1.006 | .976 |
| Physical limitation/disability | 1.104 | .727 | 1.032 | .961 |
| Body Mass Index (1st quartile)a | ||||
| BMI 2st quartile | .976 | .874 | .916 | .908 |
| BMI 3rd quartile | .855 | .860 | .907 | .723* |
| BMI 4th quartile | 1.059 | .874 | .982 | .798* |
| Parent(s) with obesity | .824 | .841† | .849* | .954 |
| Health insurance | 1.093 | .984 | .995 | 1.040 |
| Other WI controls | ||||
| Gender (1=male) | 1.996** | 1.925* | 1.964** | 1.816** |
| Race/ethnicity (White)a | ||||
| Black | .284* | .473** | .706** | .957 |
| Native American | .480 | .648 | .583 | .808 |
| Asian | .253** | .504** | .483** | .653** |
| Hispanic | .584* | .559** | .563** | .660** |
| Other race | .792 | .654 | .671 | .966 |
| Parents mean education level (high school)a | ||||
| Less than high school | .829 | .975 | .863† | .916 |
| Some college | 1.251 | 1.267† | 1.002 | .957 |
| College | 1.225 | 1.235† | 1.027 | .908 |
| More than college | .836 | .933 | .910 | .745** |
Notes:
p < 0.01,
p < 0.05,
p < 0.1,
Reference group in parenthesis.
Also controlling for age, neighborhood poverty, and family structure.
Many demographic variables including gender, age, and race were also significantly related to young adults’ participation in risky health behaviors, covarying for baseline levels of risky behaviors and all health and demographic covariates (see Tables 3–4). Males were more likely to participate in risky health behaviors than females, and participation in risky health behaviors appeared to increase linearly with age. Additionally, self-identified Black, Asian, and Hispanic youth were less likely to report more risky health behaviors, compared to the White youth, holding all other covariates constant.
Discussion
Despite the growing interest in positive health, this is the first study to examine the association between perceived health (and health behaviors) and both positive well-being and depressive symptoms in an adolescent sample. Since positive and negative states are inversely correlated, associations between positive well-being and health outcomes have sometimes been conceptualized as merely the absence of depression or negative affect. However, the majority of previous studies of depression or depressive symptoms and health rarely examine the potential impact of positive well-being [22], making it impossible to compare the relative significance of distinctive measures of well-being and ill-being. This study demonstrates that positive well-being during adolescence predicts better perceived health and fewer risky health behaviors during young adulthood, independent of depressive symptoms. Furthermore, these analyses suggest that some positive characteristics (i.e. happiness, enjoyment of life, feeling just as good as other people, and hopefulness about the future) may have a stronger association with health outcomes than other characteristics (i.e. self-esteem).
Risky health behaviors represent a particularly important outcome to examine in a young adult population since it is estimated that 40% of premature deaths in the United States are related to health-compromising behaviors, with obesity, physical inactivity, and smoking identified as the most prevalent causes of mortality [38]. Although most health-intervention programs target risky behaviors directly, new policy initiatives, including the newest iteration of Healthy People (2020), recommend an increased focus on positive youth development interventions for improving adolescent health as one of the key emerging issues for the next decade [39]. Indeed, the evidence presented in this study is well-aligned with the Positive Youth Development movement, and suggests that promoting and nurturing positive characteristics in families, schools, and after-school programs, may be a promising way to improve lifelong health.
Limitations
While there are many benefits to using a nationally representative dataset such as Add Health, there are also several challenges and limitations, such as missing data. To address the potential bias from sample attrition, we weighted the regression analyses using the in-school sampling weights to take care of post-stratification issues. Additionally, we include a comprehensive set of potentially confounding baseline covariates in our models, but we cannot account for biases associated with unmeasured characteristics that change between the baseline (Wave I) and final observations (Wave III). These characteristics could affect adolescents’ positive well-being or depressive symptoms, as well as their health. Another limitation is that all study data were self-reported, so estimation results are sensitive to mismeasurement or misreporting. The results from these models should be interpreted with these limitations in mind.
Until recently, the only research on social and psychological influences on health, studied how negative feelings and experiences are linked to poor health, therefore most existing data has only limited information on positive psychological qualities. For this study, we constructed a positive well-being measure from two other scales (CES-D and Rosenberg Self-Esteem scale), based on statistical support that these variables are interrelated (e.g. high internal consistency) and theoretical underpinnings that these positive characteristics are jointly predictive of health [3–5]. More research is needed to understand how positive emotions and characteristics such as happiness, optimism, perceived social acceptance, and self-esteem relate to both perceived and diagnosed health conditions.
Despite these limitations, the present study strongly suggests that positive well-being during adolescence predicts good perceived general health in young adulthood and may serve to reduce multiple risky health behaviors, including fast food consumption, low physical activity, binge drinking, cigarette smoking, marijuana use, and other illicit drug use. Consequently, positive well-being could be a key psychological factor to cultivate during adolescence to promote positive health and reduce risky behaviors that may lead to disease later in life.
Acknowledgments
This research was supported by NICHD R01 HD053731-01A1. This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01-HD31921 for this analysis.
Footnotes
This variable was truncated at “4 or more risky health behaviors” due to the small number of responses of 5 (n=123) or 6 (n=20) in the sample.
Ordinal logistic regression assumes that the coefficients that describe the relationship between each pair of outcome groups are the same.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Singer B, Ryff C, editors. New horizons in health: An integrative approach. Washington, D.C: National Academy Press; 2001. [PubMed] [Google Scholar]
- 2.Halfon N, Hochstein M. Life course health development: an integrated framework for developing health, policy, and research. The Milbank Quarterly. 2002;80(3):433–479. doi: 10.1111/1468-0009.00019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Manderscheid RW, Ryff CD, Freeman EJ, et al. Evolving definitions of mental illness and wellness. Preventing Chronic Disease. 2010;71:1–6. [PMC free article] [PubMed] [Google Scholar]
- 4.Lyubomirsky SL, King L, Diener E. The benefits of frequent positive affect: Does happiness lead to success? Psychological Bulletin. 2005;131(6):803–855. doi: 10.1037/0033-2909.131.6.803. [DOI] [PubMed] [Google Scholar]
- 5.Dockray S, Steptoe A. Positive affect and psychobiological processes. Neuroscience & Biobehavioral Reviews. 2010;35(1):69–75. doi: 10.1016/j.neubiorev.2010.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Steptoe A, Wardle J, Marmot M. Positive affect and health-related neuroendocrine, cardiovascular, and inflammatory processes. Proceedings of the National Academy of Sciences. 2005;102(18):6508–6512. doi: 10.1073/pnas.0409174102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fredrickson B. The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist. 2001;56(3):218–226. doi: 10.1037//0003-066x.56.3.218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Danner DD, Snowdon DA, Friesen WV. Positive emotions in early life and longevity: Findings from the nun study. Journal of Personality and Social Psychology. 2001;80(5):804–813. [PubMed] [Google Scholar]
- 9.Ashby F, Isen A, Turken U. A neuropsychological theory of positive affect and its influence on cognition. Psychological Review. 1999;106:529–550. doi: 10.1037/0033-295x.106.3.529. [DOI] [PubMed] [Google Scholar]
- 10.Zimmerman MA, Copeland LA, Shope JT, et al. A longitudinal study of self-esteem: Implications for adolescent development. Journal of Youth and Adolescence. 1997;26(2):117–141. [Google Scholar]
- 11.Steptoe A, O’Donnell K, Marmot M, et al. Positive affect, psychological well-being, and good sleep. Journal of Psychosomatic Research. 2008;64(4):409–415. doi: 10.1016/j.jpsychores.2007.11.008. [DOI] [PubMed] [Google Scholar]
- 12.Cohen S. Social relationships and health. American Psychologist. 2004;59:676–684. doi: 10.1037/0003-066X.59.8.676. [DOI] [PubMed] [Google Scholar]
- 13.Blane D. The life course, the social gradient, and health. In: Marmot M, Wilkinson RG, editors. Social determinants of health. New York: Oxford University Press; 1999. pp. 64–80. [Google Scholar]
- 14.Graber J, Brooks-Gunn J. Transitions and turning points: Navigating the passage from childhood through adolescence. Developmental Psychology. 1996;32(4):768–776. [Google Scholar]
- 15.Harris K. An integrative approach to health. Demography. 2010;47(1):1–22. doi: 10.1353/dem.0.0091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hall GS. Adolescence: Its psychology and its relation to physiology. anthropology, sociology, sex, crime, religion, and education. Englewood Cliffs: Pretice-Hall; 1904. [Google Scholar]
- 17.Freud A. In: Adolescence as a developmental disturbance. Caplan G, Lebovici S, editors. New York: Basic Books; 1969. pp. 5–10. [Google Scholar]
- 18.Benson PL, Scales PC, Hamilton SF, et al. Positive youth development: Theory, research, and applications. In: Lerner RM, editor. Theoretical models of human development. Hoboken: Wiley; 2006. pp. 894–941. [Google Scholar]
- 19.Roth JL, Brooks-Gunn J. Youth development programs: Risk, prevention and policy. Journal of Adolescent Health. 2003;32(3):170–182. doi: 10.1016/s1054-139x(02)00421-4. [DOI] [PubMed] [Google Scholar]
- 20.Catalano RF, Berglund ML, Ryan JAM, et al. Positive youth development in the United States: Research findings on evaluations of positive youth development programs. The Annals of the American Academy of Political and Social Science. 2004;591:98–124. [Google Scholar]
- 21.Pittman KM, Irby M, Ferber T. Unfinished business: Further reflections on a decade of promoting youth development. In: Ventures PP, editor. Youth development: Issues, challenges and directions. Philadelphia: Public/Private Ventures; 2000. pp. 17–64. [Google Scholar]
- 22.Pressman SD, Cohen S. Does positive affect influence health? Psychological Bulletin. 2005;131(6):925–971. doi: 10.1037/0033-2909.131.6.925. [DOI] [PubMed] [Google Scholar]
- 23.Russell J, Carroll J. On the bipolarity of positive and negative affect. Psychological Bulletin. 1999;125:3–30. doi: 10.1037/0033-2909.125.1.3. [DOI] [PubMed] [Google Scholar]
- 24.Ryff CD, Dienberg Love G, Urry HL, et al. Psychological well-being and ill-being: Do they have distinct or mirrored biological correlates? Psychotherapy and Psychosomatics. 2006;75(2):85–95. doi: 10.1159/000090892. [DOI] [PubMed] [Google Scholar]
- 25.Keenan-Miller D, Hammen C, Brennan P. Health outcomes related to early adolescent depression. Journal of Adolescent Health. 2007;41(3):256–262. doi: 10.1016/j.jadohealth.2007.03.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wickrama K, Wickrama T, Lott R. Heterogeneity in youth depressive symptom trajectories: Social stratification and implications for young adult physical health. Journal of Adolescent Health. 2009;45(4):335–343. doi: 10.1016/j.jadohealth.2009.04.018. [DOI] [PubMed] [Google Scholar]
- 27.Blaine B. Does depression cause obesity?: A Meta-analysis of longitudinal studies of depression and weight control. Journal of Health Psychology. 2008;13(8):1190–1197. doi: 10.1177/1359105308095977. [DOI] [PubMed] [Google Scholar]
- 28.Harris KM, Halpern CT, Whitsel E, et al. The National Longitudinal Study of Adolescent Health: Research Design. 2009 Available from: http://www.cpc.unc.edu/projects/addhealth/design.
- 29.Sheehan T, Fifeld J, Reisine S, et al. The measurement structure of the Center for Epidemiologic Studies Depression scale. Journal of Personality Assessment. 1995;64(3):507–521. doi: 10.1207/s15327752jpa6403_9. [DOI] [PubMed] [Google Scholar]
- 30.Rosenberg M. Society and the adolescent self-image. Princeton: Princeton University Press; 1965. [Google Scholar]
- 31.Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm. Journal of the American Medical Association. 1997;278(10):823–832. doi: 10.1001/jama.278.10.823. [DOI] [PubMed] [Google Scholar]
- 32.Goodman E, Whitaker R. A prospective study of the role of depression in the development and persistence of adolescent obesity. Pediatrics. 2002;110(3):497–504. doi: 10.1542/peds.110.3.497. [DOI] [PubMed] [Google Scholar]
- 33.Kubzansky L, Kawachi I. Affective states and health. In: Berkman L, Kawachi I, editors. Social Epidemiology. New York: Oxford University Press; 2000. pp. 213–241. [Google Scholar]
- 34.StataCorp. STATA (Version Release 11.0) Stata Corporation; College Station, TX: 2009. [Google Scholar]
- 35.Tourangeau R, Shin H. National Longitudinal Study of Adolescent Health: Grand Sample Weight. Carolina Population Center, University of North Carolina; Chapel Hill: 1998. [Google Scholar]
- 36.Allison PD. Missing data. Thousand Oaks, CA: Sage; 2001. [Google Scholar]
- 37.Park N. Character strengths and positive youth development. The Annals of the Academy of Political and Social Science. 591(1):40–54. [Google Scholar]
- 38.Mokdad AH, Marks JS, Stroup DF, et al. Actual causes of death in the United States. Journal of the American Medical Association. 2000;291(10):1238–1245. doi: 10.1001/jama.291.10.1238. [DOI] [PubMed] [Google Scholar]
- 39.U.S. Department of Health and Human Services. [Accessed on April 20, 2011];Healthy People 2020. Available at: http://healthypeople.gov/2020/