Family Obligations and Asthma in Youth: The Moderating Role of Socioeconomic Status

Abstract

Objective:

Fulfilling family obligations— providing instrumental help to and spending time with family— is a common aspect of family relationships. However, whether fulfilling these obligations links with physical health remains unclear. This study investigated whether fulfilling family obligations was associated with asthma outcomes among youths, and whether these associations differed depending on family socioeconomic status (SES).

Method:

Participants were 172 youths aged 8 to 17 years (M_age = 12.1; 54% male) who had been physician-diagnosed with asthma. Youths reported on family obligation frequency, completed a clinical measure of asthma control (Asthma Control Test; ACT), and completed a measure of airway inflammation (fractional exhaled nitric oxide). Parents completed a measure of their child’s asthma control (ACT) and reported on family income.

Results:

Fulfilling family obligations was not associated with asthma outcomes (βs < .14, ps > .075). However, SES (family income) interacted with family obligations such that fulfilling family obligations was associated with greater airway inflammation (interaction term β = −.17, p = .023) and poorer parent-reported asthma control (interaction term β = .15, p = .039), only among youth from lower SES backgrounds, but not among youth from higher SES backgrounds. Exploratory analyses suggest that these interactions were robust against covariates and were largely consistent across age and across the two dimensions of family obligation behaviors.

Conclusions:

Findings from this study suggest that among youth from lower SES backgrounds, engaging in more frequent family obligation behaviors may have negative repercussions in terms of their asthma.

Asthma, a chronic lung disease that involves inflammation of the airways, is one of the most common chronic illnesses among youth (Zahran, Bailey, Damon, Garbe, & Breysse, 2018). According to recent Morbidity and Mortality Weekly Reports by the Center of Disease Control and Prevention, asthma affects the lives of approximately 6 million American youth, accounting for about 13.8 million missed school days per year (Zahran et al., 2018) and 14% of all emergency department visits among youths aged 5 to 17 years (Ashman, Rui, & DeFrances, 2018). Health care use is high among youths with asthma: 71% take routine care visits, 5% were hospitalized, and 54% reported at least one asthma attack in the past year which resulted in 17% having to visit emergency department or urgent care (Zahran et al., 2018). In addition, asthma disproportionately affects youths from different demographic groups: youths from families with lower income as well as African American and Latin American youths (vs. European American youths) were more likely to be diagnosed with asthma (Zahran et al., 2018). Importantly, empirical evidence suggests that the social environment in which youth are situated is linked to asthma (Chen, Chim, Strunk, & Miller, 2007; Chen et al., 2017b; Klinnert et al., 2001). This paper focuses on the links between family social environment and asthma.

Research has long shown that youth’s interactions and relationships with their families are important for youth’s health (Amato & Keith, 1991; Chen, Brody, & Miller, 2017a; Repetti, Taylor, & Seeman, 2002; Wegman & Stetler, 2009). For instance, family environments marked by conflict, lack of parental support, and chaos have been linked with worse physical health-related outcomes, such as heightened inflammation, among youths (Miller & Chen, 2010; Repetti et al., 2002; Schreier, Roy, Frimer, & Chen, 2014; Troxel & Matthews, 2004). Furthermore, family functioning and interactions also have clinical relevance in the context asthma; for example, conflicts with family (Chen, Bloomberg, Fisher Jr, & Strunk, 2003; Chen et al., 2006), lack of family support (Chen et al., 2007), and poorer parent-child relationships in the form of insecure child attachment (Bleil, Ramesh, Miller, & Wood, 2000), parenting difficulties (Klinnert et al., 2001), and low parental support (Miller, Gaudin, Zysk, & Chen, 2009b) have been linked with worse asthma outcomes. As such, it is not surprising that many current asthma practice recommendations are aimed at the family (and not just the individual with asthma), such as keeping a clean house, and thus asthma management becomes a part of family life (Drotar & Bonner, 2009; Fiese, Wamboldt, & Anbar, 2005; Rosales, McQuaid, & Koinis-Mitchell, 2017). In sum, the different ways youths interact with their families have implications for their physical health not only among healthy populations, but also among clinical populations like asthma patients.

One normative, but understudied, way in which youths interact with their families day-to-day is through fulfilling family obligations. Family obligations refer to the behaviors youth engage in to provide tangible help to their family (e.g., cooking, cleaning home, and caring for siblings) and to spend time with them (e.g., visiting extended family members) (Fuligni, Tseng, & Lam, 1999; Hardway & Fuligni, 2006). These behaviors can be normative components of youth’s everyday interactions with their families (Bianchi & Robinson, 1997; Cogle, Tasker, & Morton, 1982; Fuligni et al., 1999; Gager, Sanchez, & Demaris, 2009; Hofferth & Sandberg, 2001), with daily diary estimates indicating that youth can spend approximately seven hours a week fulfilling family obligation behaviors (Fuligni, Yip, & Tseng, 2002; Gager et al., 2009; Hardway & Fuligni, 2006; Larson & Verma, 1999). This amount of time spent fulfilling family obligations was shown to be comparable to other activities significant to youth, such as studying or socializing with peers (Hardway & Fuligni, 2006), and has been linked with psychological well-being. In the present study, we examined whether these normative obligatory behaviors have implications for physical health, specifically in youth with asthma, and investigated whether the health correlates of family obligations may be different depending on whether youth come from low or high socioeconomic status family backgrounds.

Family Obligations and Psychological Well-Being

The majority of studies that have examined the link between engaging in family obligation behaviors and health have focused on psychological well-being. In some studies, engaging in family obligation behaviors has been linked with positive psychosocial outcomes, such as greater positive mood, greater satisfaction with life, and better interpersonal outcomes (East, Weisner, & Reyes, 2006; Kuperminc, Wilkins, Jurkovic, & Perilla, 2013; Telzer & Fuligni, 2009). However, in other studies, engaging in family obligation behaviors has been linked with negative psychosocial outcomes, such as greater feelings of demand and burden, more stress, and poorer academic achievement (Early, Cushway, & Cassidy, 2006; East et al., 2006; Kuperminc et al., 2013; Telzer & Fuligni, 2009; Tseng, 2004; Vasquez-Salgado, Greenfield, & Burgos-Cienfuegos, 2015). These seemingly contradictory patterns may be explained by the meaning youth derive from fulfilling family obligations (East, 2010; Goodnow & Lawrence, 2001; Kuperminc et al., 2013; Telzer & Fuligni, 2009; Weisner, 2001). For example, in one study, helping the family provided youth with a sense of role fulfillment (i.e., the sense of being a good child or a good sibling), which in turn was associated with greater positive mood, despite experiencing greater demands from family (Telzer & Fuligni, 2009). Thus, although fulfilling family obligations can be stressful and demanding for youth, it may also be associated with greater psychological well-being when it provides opportunities for role fulfillment in the family.

Family Obligations and Physical Health

The existing theory and empirical research on the link between family obligations and physical health is relatively limited and, to some extent, inconclusive about whether such a link exists. Some work has theorized that because engaging in more family obligation behaviors can increase feelings of stress and demand (East et al., 2006; Kuperminc et al., 2013; Telzer & Fuligni, 2009) and because these psychological states have been linked with worse physical health (McEwen, 2000; Miller, Chen, & Cole, 2009a; Miller, Cohen, & Ritchey, 2002), engaging in family obligation behaviors may be linked with worse physical health. Indeed, one study found that youth who spent more time providing help for the family had higher levels of systemic inflammation, as measured by C-reactive protein and soluble interleukin 6 receptor (Fuligni et al., 2009). However, empirical support for this hypothesis that family obligations will be linked to worse health does not emerge consistently. The one other study that has explored the relationship between family obligations and physiological outcomes showed that doing household chores was not associated with youth’s cortisol profiles (McHale et al., 2012). Thus, previous literature on the link between family obligations and physical health is scarce and findings are inconclusive among the emerging studies.

The Role of Context: Socioeconomic Status

The mixed findings for the link between family obligations and health suggests that there may be contextual moderators influencing the relationship, such that the health implications of family obligations may depend on the family context in which these behaviors take place. One context variable that may be relevant is socioeconomic status (SES). The family environments associated with low SES are very different from those associated with high SES (Chen & Miller, 2013). For example, families from lower (vs. higher) SES backgrounds are often characterized as having less stability in daily family routines (Jensen, James, Boyce, & Hartnett, 1983), more frequent conflicts (Conger & Elder, 1994), and higher parental demands for youth (Conger & Elder, 1994; McLoyd, 1990). These SES-related differences in family environments may have differential implications for the experience and meaning of fulfilling family obligations (Lehman, Taylor, Kiefe, & Seeman, 2005). For example, in low SES families, family members may be more dependent on youth helping out, increasing the demand for youth (Burton, 2007; Tseng, 2004). Moreover, in low SES households, family obligations may be perceived as mandatory, making the experience more stressful and burdensome (Burton, 2007; Evans, Gonnella, Marcynyszyn, Gentile, & Salpekar, 2005). For instance, knowing that there is no one else to take care of younger children, youth from lower SES families may experience this task as particularly demanding. As such, fulfilling family obligation may be associated with worse health among youths from lower (vs. higher) SES background. Consistent with this hypothesis, one study of healthy youths found higher levels of inflammation (i.e., higher levels of inflammation and higher pro-inflammatory cytokine production in response to bacterial stimulation) among youths who both provided a lot of help to their families and perceived their families as making a lot of demands on them, as compared to youths who provided more help, but experienced less demand (Levine, Hoffer, & Chen, 2017). Further, low SES families may have more conflict and difficulties communicating with one another (Conger & Elder, 1994). Having to fulfill family obligations in high conflict contexts may make the experience even more stressful. Following this reasoning, one study found that providing help for family was associated with cigarette and alcohol use only among adolescents from households with high levels of conflict, but not among adolescents from households with low levels of conflict (Telzer, Gonzales, & Fuligni, 2014). Taken together, we similarly hypothesize that engaging in family obligations may be more strongly associated with worse health outcomes among youths from lower (vs. higher) SES background (McEwen, 2012; Miller et al., 2009a).

Current Investigation

The present study investigated the relationships among SES, family obligations, and physical health in youth. Because of the high prevalence and impact of asthma in youth and because of the significant role of the family in asthma, asthma provided a useful model for studying the interplay between SES and family obligations in a disease context in children. In a sample of youth who were physician-diagnosed with asthma, we tested whether SES would interact with family obligations to predict multiple asthma outcomes, including clinical measures of asthma control and a non-invasive measure of airway inflammation (fractional exhaled nitric oxide). We hypothesized that family obligation behaviors carried out in low SES environments would be associated with worse asthma outcomes, whereas family obligation behaviors in higher SES environments would not be associated with asthma outcomes. Additional exploratory analyses examined whether the SES by family obligation interactions were robust against additional potential confounds (medical, environmental, and family structure covariates) and whether the interactions were consistent across age and across the two dimensions of family obligation behaviors (helping the family vs. spending time with the family).

Method

Participants.

Participants were 172 youths (59% European American), aged 8 to 17 years, who had been physician-diagnosed with asthma. Youths were recruited to be part of a larger research project aimed to investigate the neighborhood, family, and individual contributors to asthma disparities. This paper focused on a family factor and its links with clinical measures of asthma and airway inflammation. Youths were recruited from the Chicago metropolitan area (i.e., non-rural regions) through one health care system, NorthShore University HealthSystem, and one federally-qualified health center, Erie Family Health Center. Families were fluent in English, youths had no current or history of any chronic physical illnesses (except for asthma), had no acute respiratory illness and were not on oral steroids at the time of visit. Youth visited the research lab with a parent (88% mothers) and completed all assessments in a single visit. Because our sample includes younger youth, a reading assessment prior to administering questionnaires was conducted for youth under the age of 12. If youth had difficulty with the reading screener, a research assistant read all questions out loud and helped explain terms to youth. In addition, for all youth, the research assistant remained available for questions throughout the questionnaire period. Youths gave written assent, and parents provided written consent. This study was approved by Northwestern, NorthShore, and Erie Institutional Review Boards (See Table 1 for a summary of sample characteristics).

Table 1.

Demographic Features and Descriptive Statistics (N=172).

	Mean (SD)	Percent (Frequency)
Age	12.06 (2.47)
Sex - Male		41.3 (71)
Race/Ethnicity
European American		58.7 (101)
African American		23.3 (40)
Asian American		5.8 (10)
Latino/Hispanic American		7.6 (13)
Other ethnicities		4.7 (8)
Family income
Less than $5000		1 (1)
$5000 – $19,999		6.4 (11)
$20,000 – $34,999		6.4 (11)
$35,000 – $49,999		10.5 (18)
$50,000 – $74,999		12.2 (21)
$74,999 – $99,999		12.8 (22)
$100,000 – $149,999		22.1 (38)
$149,999 – $199,999		10.5 (18)
Higher than $199,999		18.6 (32)
Asthma severitya	2.45 (0.87)
Use of inhaled corticosteroidsb	2.61 (3.08)
Use of beta agonistsb	1.71 (2.30)
Home environment controlc	3.85 (2.97)
Exposure to second-hand smoked	0.54 (1.28)
Family size	4.14 (1.24)
Birth order – first born		43.0 (74)
Family obligatione	3.48 (0.72)
Airway inflammation (ppb)f	29.09 (25.23)
Parent-reported asthma controlg	20.62 (3.51)
Youth-reported asthma controlg	19.20 (3.59)

Note.

^aAsthma severity ranged from 1 to 4, higher score indicated more severe asthma.

^bUse of medications refers to the number of times in past week.

^cHome environment control ranged from 1 to 9, higher score indicated better home environment control.

^dExposure to second-hand smoke refers to the average number of days per week.

^eFamily obligation was rated on a 1 to 5 scale, with higher scores indicating more engagement of family obligation behaviors.

^fFor youth 12 or older, airway inflammation lower than 25 ppb is considered low, between 25 to 50 ppb is considered intermediate, and higher than 50 ppb is considered high.

^gParent-reported and youth-reported asthma control ranged from 8 to 25, with higher scores reflecting greater asthma control. Asthma control scores greater than 19 indicate well-controlled asthma.

Measures.

Family Obligation.

Youths completed an adapted version of the current assistance measure (Fuligni et al., 1999) to report how often they typically engaged in family obligation activities. This scale includes 11 items describing two dimensions of family obligation: providing tangible help for the family (e.g., run errands that the family needs done, help out around the house) or spending time with family members (e.g., spend time at home with family, spend time with extended family members). Items were rated on a 1 (almost never) to 5 (almost always) scale and averaged to create a family obligation score, such that higher score indicates more frequent engagement of family obligation activities, α = .81. The current assistance measure has demonstrated good internal consistency (Fuligni et al., 1999) and has been linked with more positive family relationships (Fuligni & Pedersen, 2002; Fuligni et al., 1999).

Airway Inflammation.

The concentration of nitric oxide in the airways has been shown to reflect the amount of inflammation in the airways (Kharitonov, O’Connor, Evans, & Barnes, 1995; Van Den Toorn et al., 2001). The concentration of nitric oxide in exhaled breath, or fractional exhaled nitric oxide (FeNO), was measured using an electrochemical analyzer (the NIOX system; Aerocrine AB, Solna, Sweden) following the American Thoracic Society Guidelines. Specifically, after one inhalation and one exhalation, youths inserted a mouthpiece, inhaled to total lung capacity, and then exhaled at a pressure of 10 – 20cm H₂O for 10 seconds to maintain a constant flow rate of 50 ± 5 mL/s. Youths performed this test while standing and were presented with computer-based visual cues that facilitated maintenance of a constant flow rate. To ensure accuracy of the assessment, participants were not allowed to exercise or consume any food or caffeine one hour prior to the test. Because there were two extreme outliers (i.e., >3 SDs above the mean), FeNO was winsorized such that values of the two outliers were replaced with the next highest value.

Asthma Control.

Asthma control was reported by both the parent and the youth using the Asthma Control Test™ (ACT), a scale commonly used in clinical settings (Nathan et al., 2004). This scale has been shown to have good test-retest reliability (Schatz et al., 2006) as well as convergent and discriminant validity (Nathan et al., 2004; Schatz et al., 2006) among asthma patients twelve years or older. For youth-reports, we chose to administer only one version of the ACT (the one intended for those twelve years or older) for all youths across all ages so that statistical analyses for the same measure could be conducted on the entire sample. For parent-reports, similar to other studies (duRivage et al., 2017), the same ACT items completed by youths were adapted for parents to report on their child’s asthma condition (e.g., During the past 4 weeks, how often has your child had shortness of breath). The ACT includes five items assessing asthma symptoms, use of rescue medications, the effects of asthma on daily functioning, and perceived asthma control over the past four weeks, rated on a 5-point scale and summed such that higher scores indicated better asthma control (α_youth-report = .72; α_{parent-report} = .80). Consistent with previous literature (duRivage et al., 2017), parent- and youth-reported asthma control were moderately correlated, r = .49, p < .001.

Socioeconomic Status (SES).

SES was measured by interviewing parents about their family income. Specifically, parents reported the total family income received from all members living in the household and from all sources of income (e.g., wages, government assistance, alimony) over the past 12 months. Family income is an established measure of SES (“The MacArthur Research Network on Socioeconomic Status and Health,” 2008; Oakes & Rossi, 2003), and consistent with measures of SES in our previous studies (Chen, McLean, & Miller, 2015; Schreier & Chen, 2010; Schreier et al., 2014).

Covariate Variables.

Demographic variables, medical variables, environmental variables, and family structure variables were assessed as covariates. Demographic variables included youth age, sex (male vs. female), and race/ethnicity (European American vs. non-European American). Medical covariates included asthma severity and use of medications. Asthma severity was determined through a combination of symptoms and medications, as recommended by the National Asthma Education and Prevention Program/Expert Panel Report 2 guidelines and ranged from 1 (mild intermittent) to 4 (severe) (Bacharier et al., 2004). Use of medications referred to use of inhaled corticosteroids and use of beta agonists (number of times in the past week). Environmental variables include exposure to second-hand smoke (average number of days per week) and home environment control (a 9-point interviewer rating of environment control based on the Family Asthma Management System Survey (McQuaid, Walders, Kopel, Fritz, & Klinnert, 2005), a semi-structured interview with youths and parents that probed exposures to environmental triggers, such as pets, and efforts to improve environment, such as use of air filters). Family structure variables included family size (number of people living in the family home) and youth’s birth order.

Statistical Analyses

To test whether SES moderated the association between family obligation and asthma outcomes, a series of hierarchical regression analyses were conducted following the recommendations of Aiken & West (1999). Specifically, asthma outcomes (i.e., airway inflammation, parent-reported asthma control, and youth-reported asthma control) were predicted from (1) demographic variables; (2) SES and family obligation; and (3) the interaction between SES and family obligation. Demographic variables, including race/ethnicity (given the substantial overlap between SES and race/ethnicity in the United States), gender, and age were included as covariates in all models. All continuous predictors were mean-centered and dichotomous predictors were effect-coded. Significant interaction effects were followed up with tests of simple slopes between family obligation and asthma outcomes at ±1 SD from the mean of SES. Secondary analyses tested the robustness of results by additionally controlling for medical (asthma severity and medication use), environmental (home environment control and exposure to second-hand smoke), and family-structure (family size and birth order) covariates. Secondary analyses also tested whether results were consistent across dimensions of family obligation behaviors by separately testing the time spent with family by SES interaction and the family assistance by SES interaction. Finally, to test whether results were consistent across age, we tested the three-way interaction of age by SES by family obligations.

Results

Preliminary Analyses.

Simple correlations among study variables are presented in Table 2. Older age was associated with greater airway inflammation, male (vs. female) youth reported better asthma control, and European American (vs. non-European American) youth had less airway inflammation and better parent-reported asthma control.

Table 2.

Simple correlations among study variables (N = 172).

		1	2	3	4	5	6	7	8	9	10	11	12	13	14
1.	Age	-
2.	Gender	.13	-
3.	Race/Ethnicity	−.06	−.01	-
4.	Asthma severity	−.07	−.05	−.09	-
5.	Use of inhaled corticosteroids	−.15	.05	−.01	.27*	-
6.	Use of beta agonists	−.03	.13	−.15	.28*	.23*	-
7.	Home environment control	−.14	−.07	−.07	−.08	.04	−.03	-
8.	Exposure to second-hand smoke	.16*	.02	−.02	−.14	−.04	.01	−.18*	-
9.	Family size	−.18*	−.04	.13	.01	.14	.14	−.06	−.02	-
10.	Birth order	−.08	−.02	.01	−.16*	−.03	−.14	.13	.19*	−.18*	-
11.	SES	.03	.05	.45*	−.12	.00	−.29*	.02	−.05	.24*	−.03	-
12.	Family obligation	−.25*	.16*	−.03	−.03	.07	.05	.09	−.06	.21*	.10	−.05	-
13.	Airway inflammation	.17*	.02	−.20*	−.10	−.12	.08	.03	−.02	−.12	−.03	−.17	.07	-
14.	Parent-reported asthma control	.11	−.08	.30*	−.37*	−.19*	−.54*	.06	−.06	.00	.18*	.31*	−.13	−.06	-
15.	Youth-reported asthma control	.05	−.19*	.03	−.16	−.09	−.43	.10	−.07	−.11	.09	.12	−.18	.02	.49

Note.

^*p<.05. Gender was effect-coded such that male was coded −1 and female was coded 1. Race/Ethnicity was effect-coded such that European American was coded 1 and non-European American was coded −1. Birth order was effect-coded such that first born was coded 1 and later born was coded −1.

SES, Family Obligations, and Asthma Outcomes.

Results from all regression models are displayed in Table 3.

Table 3.

Hierarchical Regression Analyses of SES and Family Obligation Predicting Asthma Outcomes (N = 171).

	Airway Inflammation	Parent-Reported Asthma Control	Youth-Reported Asthma Control
Step 1
Age	β = .16, p = .03	β = .14, p = .07	β = .08, p = .28
Gender	β = .00, p = .98	β = −.10, p = .18	β = −.20, p = .01
Race/Ethnicity	β = −.19, p = .01	β = .31, p < .001	β = .03, p = .70
Step 2
SES	β = −.11, p = .20	β = .22, p = .01	β = .14, p = .10
Family obligation	β = .12, p = .14	β = −.06, p = .43	β = −.14, p = .08
Step 3
SES × Family obligation	β = −.17, p = .02	β = .15, p = .04	β = .08, p = .29

Note. Gender was effect-coded such that male was coded −1 and female was coded 1. Race/Ethnicity was effect-coded such that European American was coded 1 and non-European American was coded −1.

Airway Inflammation.

There was a significant interaction between SES and family obligation predicting airway inflammation, β = −.17, p = .02, with no main effects of either SES or family obligation. Specifically, greater family obligation was linked with greater airway inflammation only among youth from lower SES backgrounds, β = .26, p = .010, but not among youth from higher SES backgrounds, β = −.08, p = .482 (see Figure 1, Panel A).

Figure 1.

Interaction between SES (socioeconomic status) and family obligation predicting airway inflammation (Panel A) and parent-reported asthma control (Panel B), controlling for demographic variables. Airway inflammation was measured with FeNO (fractional exhaled nitric oxide) in ppb. The standardized regression coefficients (βs) and p-value notations refer to the estimated regression lines for the links between family obligation and asthma outcome at ±1 SD of SES. Low and high family obligation also refers to ±1 SD. Error bars reflect standard errors of asthma outcome scores at low SES (−1 SD) and low family obligation (−1 SD), at low SES and high family obligation, at high SES and low family obligation, as well as at high SES and high family obligation.

Parent-Reported Asthma Control.

There was a significant interaction between SES and family obligation predicting parent-reported asthma control with a main effect of SES, but no main effect of family obligation. The main effect of SES was such that higher SES was linked with greater parent-reported asthma control. The interaction effect, β = .15, p = .04, was such that among youth from lower SES backgrounds, the association between family obligation and parent-reported asthma control was β = −.19, p = .055, whereas among youth from higher SES backgrounds, the association between family obligation and parent-reported asthma control was, β = .11, p = .310 (see Figure 1, Panel B).

Youth-Reported Asthma Control.

There was no significant interaction between SES and family obligation predicting youth-reported asthma control. There were no main effects for SES and for family obligation. Although the interaction was not statistically significant, tests of simple slopes indicated that, similar to the above outcomes, greater family obligation was linked with worse youth-reported asthma control among youth from lower SES backgrounds, β = −.21, p = .041, but not among youth from higher SES backgrounds, β = −.05, p = .693.

Secondary Analyses.

Were results robust when controlling for potential confounds?

Medical covariates.

Because airway inflammation (FeNO) may be affected by medical covariates (Rodway, Choi, Hoffman, & Sethi, 2009; Taylor, Pijnenburg, Smith, & De Jongste, 2006), we tested whether the SES by family obligation interaction on airway inflammation was robust when controlling for asthma severity and medication use additional to demographic variables. Results remained the same, interaction β = −.17, p = .020, suggesting that the SES by family obligation on airway inflammation was independent of these medical covariates. We did not include asthma severity and medication use in analyses of asthma control because of the overlap in these variables – that is, medication use is part of the assessment of asthma control. In addition, because this was a sample of youth with pre-existing asthma, asthma severity was determined through a combination of symptoms and medications, as recommended by the National Asthma Education and Prevention Program/Expert Panel Report 2 guidelines (Bacharier et al., 2004), which has overlap with the asthma control variable.

Environmental covariates.

We further tested whether the SES by family obligation interactions were robust when controlling for home environment control and exposure to second-hand smoke in addition to the demographic variables. Results remained the same, interaction βs > .15, ps < .043, suggesting that the SES by family obligation interactions were not driven by environmental covariates.

Family structure covariates.

Because the oldest child of the household may have more family obligations and because youth from larger families may have more family obligations, we tested whether the SES by family obligation interactions were robust when controlling for birth order and family size in addition to the demographic variables. Results remained the same, interaction βs > .15, ps < .047, suggesting that the SES by family obligation interactions were independent of birth order and family size.

Were results consistent across different types of family obligation behaviors?

Family obligations theoretically encompass two dimensions: helping the family and spending time with family members (Fuligni et al., 1999). It is possible that these two dimensions of family obligations may have differential correlates with health (e.g., regular engagement of helping behaviors may be more stressful and demanding than spending time with family members). Therefore, we tested whether similar interaction effects were apparent if family obligation was operationalized by focusing only on behaviors related to helping the family, α = .72, or behaviors related to spending time with the family, α = .81. Patterns of results largely remained the same, suggesting that the SES by family obligation interactions were not driven by any one subscale: for helping behaviors, the interaction regression weight for airway inflammation was β = −.17, p = .027, the interaction regression weight for parent-reported asthma control was β = .14, p = .061; for spending time with family, the interaction regression weight for parent-reported asthma control was β = .15, p = .043 and the interaction regression weight for airway inflammation was β = −.10, p = .175.

Were results consistent across different ages?

Our sample allowed for examination of whether similar interaction effects were apparent across younger and older youths. It is possible that the interaction effects are different for youths of different ages (e.g., older youths may have more competing demands from school and work, making family obligation behaviors more stressful for them than for younger youths). Therefore, we tested the three-way interaction, examining whether age moderated the SES by family obligation interactions on asthma outcomes. There was no evidence that the SES by family obligation interactions were different across ages (βs for three-way interaction terms < .12, ps > .146).

Discussion

The current study suggests that fulfilling family obligations in certain family contexts is associated with asthma outcomes in youth. Specifically, we found that engaging in family obligations was associated with greater airway inflammation and poorer parent-reported asthma control among youth from lower SES backgrounds, but not among youth from higher SES backgrounds. This study builds on previous research documenting that family obligations can be detrimental to physical health (Fuligni et al., 2009) by examining this relationship in the context of a chronic illness and demonstrating that family obligations have relevance for disease outcomes. Our findings are consistent with previous research that found links between helping the family and worse health-related outcomes among youths who experienced more family conflicts (Telzer et al., 2014) and experienced greater demands from the family (Levine et al., 2017), both of which tend to be more common among youths from low SES families.

A novel aspect of this study is that we examined an understudied type of family interaction that is typically performed to aid in family life (e.g., youth running errands for the family) and found that this type of aid may come with a cost for youth from low SES backgrounds. One difference of this finding compared to other studies of family relationships and asthma is that the effect of family obligations is seen only in lower SES youth, whereas many studies of family relationships and asthma find main effects (Kaugars, Klinnert, & Bender, 2004; Wood, Miller, & Lehman, 2015), with a few exceptions (Mangan, Wittich, & Gerald, 2007). As such, this study’s results highlight the importance of examining the context in which family behaviors take place.

Other researchers have suggested that because families from lower SES backgrounds often have fewer resources, youth may have to engage in family obligations more frequently than youth from high SES families (Burton, 2007), which may negatively impact health. However, in this sample, SES was not associated with family obligation frequency (r = −.05), suggesting that SES-moderations were not due to SES-related differences in frequency, but instead may be due to differences in perceived meaning and experience of family obligations. For example, because alternative options for accomplishing the behaviors assessed in the obligation measure may not be readily available among lower SES families (Evans et al., 2005), youths may perceive family obligation tasks as mandatory duties that could lead to serious negative consequences if they fail to fulfill them, making the experience more burdensome and stressful (Gallant, Smale, & Arai, 2017).

In turn, more stressful psychological experiences have been linked with worse asthma outcomes among youth with asthma (Chen & Miller, 2007; Chen, Schreier, & Chan, 2012; Wright, Rodriguez, & Cohen, 1998). This may be because psychological stress impacts asthma outcomes via biological pathways. For example, psychological stress is associated with greater production of cytokines in response to in vivo allergen challenge that are theorized to be instrumental in airway inflammation (Liu et al., 2002) and in vitro mitogen stimulation (Chen, Fisher, Bacharier, & Strunk, 2003; Chen et al., 2006). Second, psychological stress may also impact asthma outcomes via behavioral pathways, such as self-management of asthma. For example, previous research has linked psychological stress with difficulty in managing asthma (Shalowitz, Berry, Quinn, & Wolf, 2001) as well as with lower adherence to medical treatments (Cluley & Cochrane, 2001).

Although there was a significant interaction between SES and family obligations for parent-reported asthma control, there was no such interaction for youth-reported asthma control. The correlation between parent-reported and youth-reported asthma control suggests moderate, correlation (r = .49). Perhaps the parent-reports captured a different perspective than the youth-reports. Because youth age was as low as eight years old, the youth-reports may have more error and inaccuracies in reporting frequencies, such as missed school days due to asthma, compared to parent-reports.

Exploratory analyses revealed no strong evidence that the SES by family obligation interactions were driven more by helping the family versus by spending time with family. Helping behaviors are more physically-demanding; however, spending time with family may take away time from other important activities for youth, such as socializing with friends and schoolwork, and thus both may be associated with negative asthma outcomes in lower SES youth. Exploratory analyses also revealed no evidence that age moderated the SES by family obligation interactions, suggesting that the SES by family obligation interactions were relatively similar across younger and older youths. While younger (vs. older) youths from low SES families may find obligation tasks more demanding (e.g., it may be more difficult for a younger youth to perform housework), older (vs. younger) youths may be put in charge of more difficult tasks (e.g., babysitting younger siblings, cooking for family), and thus may find them equally demanding. However, results for the three-way interactions should be considered tentative and should be followed-up with a larger sample.

This study highlights the role that socioeconomic contexts play in the link between family obligations and health. However, another influential context that may also shape these associations is race/ethnicity. Studies have found established racial/ethnic differences in how much youth value familism and engage in family obligation behaviors. For example, studies have found that because of cultural traditions and immigrant status, youth from Asian, Latino/Hispanic, and African American backgrounds, compared to youth from European background, typically place stronger importance on the idea of familism and communalism (Fuligni et al., 1999; Landale, Oropesa, & Bradatan, 2006; Schwartz et al., 2010) and spend more time fulfilling family obligations (Fuligni et al., 1999; Hooper et al., 2012; Khafi, Yates, & Luthar, 2014; Knight et al., 2002; Telzer & Fuligni, 2009). Furthermore, the concept of family may be different across different races/ethnicities. Supporting this notion, some studies have shown that extended family integration, such as living with and having close contact with extended family, is more common among Latino/Hispanic American and African American families than European American families (Brown, Cohon, & Wheeler, 2002; Landale et al., 2006; Sarkisian, Gerena, & Gerstel, 2006). As such, because of differences in familism values, family obligation behaviors, and conceptions of what constitutes a family, the SES by family obligation interactions may be different within different racial/ethnic groups. For example, because of stronger values on familism, Asian, Latino/Hispanic, or African American youths from lower SES background may reap more meaning and purpose through engaging in family obligation behaviors, which may in turn be associated with better health outcomes and buffer against the effects of low SES. Furthermore, although these racial/ethnic minority groups may have similar emphasis on familism, it is important to acknowledge that each racial/ethnic group has unique practices, values, ideals that may further nuance the SES by family obligation interactions. Unfortunately, because of our small sample sizes of Asian Americans (N=10), Latino/Hispanic Americans (N=13), and African Americans (N=40), this study was unable to test the interactions within each racial/ethnic group. However, these analyses are important future directions as they may provide a more refined understanding of how family relationships situated within the broader socioeconomic and racial/ethnic context contribute to childhood asthma disparities.

The present study has several limitations. First, this study relied on cross-sectional data; thus, although we propose that engaging in family obligation could negatively influence asthma outcomes among youth from low SES background, neither causality nor directionality can be determined from these correlational results. Future research will benefit from utilizing longitudinal designs to test the prospective links between family obligation and asthma outcomes in youths from low and high SES backgrounds. Second, although there are racial/ethnic differences in family obligation norms between youth from European backgrounds and those from Asian, Latin American, and African backgrounds (Fuligni, Tseng, & Lam, 1999; Hooper, Wallace, Doehler, & Dantzler, 2012), this study was unfortunately not able to examine the role of race/ethnicity in our findings. Our sample sizes of Asian Americans (N=10), Latino/Hispanic Americans (N=13), and African Americans (N=40) were not large enough to have adequate statistical power to examine the SES by family obligation interactions within any of these racial/ethnic backgrounds. We recognize that findings might be different within certain ethnic groups and thus that our results might be limited in their generalizability. Our findings represent results across a diverse sample of youth (youth of European, Asian, Latino/Hispanic, and African backgrounds), but it is possible that these results mask other interesting patterns that may emerge only within certain ethnic groups. Furthermore, because race/ethnicity has substantial overlap with SES in the United States, the SES by family obligation findings may be confounded by race/ethnicity. Future research will benefit from recruiting a large sample of youths from specific racial/ethnic backgrounds, with varying degree of SES within each, and examining the relationships among SES, family obligations, and health within specific racial/ethnic groups. Third, while the Asthma Control Test™ is an established clinical measure of asthma control (Nathan et al., 2004), it is still based on subjective perceptions. As indicated by the moderate correlation between the parent- and youth-reported asthma control scores, multiple factors likely contributed to how both parents and youths perceived and reported on asthma control. It is possible that parents’ perceptions of their child’s asthma control influenced their expectations about the amount of family obligation behaviors youth could engage in, or that youth’s symptom-reporting tendencies affected the amount of obligatory family behaviors they take on. It is also possible that parents’ knowledge about children’s symptoms and reporting of control may be systematically different by SES (e.g., parents from low SES families may not have as much knowledge about their child’s symptoms). These possibilities cannot be ruled out in a cross-sectional study, but the consistency in patterns across both biological and symptom measures lends credibility to our proposed explanations for the study findings. Fourth, we were unable to test potential mechanisms for the SES by family obligations interactions. For example, the link between family obligation and asthma outcomes among youths from low SES backgrounds may be explained by the increased demands (perceived or objective) or environmental hazards that often accompany daily life in lower SES households. Future research will benefit from formal mediated moderation analyses of these and other psychological processes (e.g., feelings and meaning derived from fulfilling family obligation behaviors) and behavioral processes (e.g., asthma management strategies) that may account for the moderation. Fifth, this study used a single assessment of family obligation. Future research may consider a daily diary approach to measure the actual amount of time spent each day on family obligation behaviors (Fuligni et al., 2009; Telzer & Fuligni, 2009). This would eliminate the need to recall obligation behaviors over time and eliminate potential scale-related inaccuracies in reporting (e.g., youths from lower, vs. higher, SES background may have different definitions for the scale labels “almost never” or “almost always”). In addition, other potential mechanistic processes could also be measured through diary formats. Finally, utilizing a diary approach could also increase the ecological validity of the findings by targeting real-world processes as they happen in youths’ everyday lives.

The present study has several implications. First, most previous research on family obligations has focused on psychosocial outcomes (East et al., 2006; Kuperminc et al., 2013; Telzer & Fuligni, 2009), and thus the present research builds on these studies by suggesting implications of obligations for physical health as well. Second, this research adds to the emerging literature that studies family-level, instead of individual-level, psychosocial factors in health disparities (Brody, Miller, Yu, Beach, & Chen, 2016; Chen, Cohen, & Miller, 2010; Repetti et al., 2002). Specifically, this study expands on previous research that has examined family relationships by demonstrating specifically that family time and family routines that come in the form of obligations can be detrimental for asthma among youths from low SES backgrounds. Third, this study may help increase awareness that while engaging in family obligations is considered normative, certain youths may face health costs of these obligations as a result. For example, family therapies for asthma typically focus on changing dysfunctional family relationships (Panton & Barley, 2000), but our findings suggest that even the daily routines in families’ lives may have implications for youth asthma, particularly among those from lower SES homes. Thus, it may be valuable for service providers to consider youths’ engagement in these day-to-day family experiences, particularly among low-SES youths with asthma.

Conclusions

Providing help to one’s family and spending time with family are common components of youth’s family lives. The present study adds to nascent research on how these common behaviors may be related to physical health among youths with asthma from families of lower (vs. higher) SES backgrounds. These findings demonstrate that, among youth with asthma, greater engagement of family obligation behaviors is linked to poorer asthma outcomes among lower SES youth. Findings from this research build onto health disparities research by considering a novel family-level psychosocial factor that may have implications for our understanding of how family relationships contribute to childhood asthma disparities.