Chronic Health Conditions and Adolescents’ Social Connectedness

Abstract

This study investigated associations between chronic developmental/behavioral and physical health conditions and social connectedness of adolescents using rich population-based data from a national U.S. birth cohort study. Potentially disabling health conditions were reported by caregivers and categorized by our team as developmental/behavioral or physical. Social connectedness was assessed using a validated scale that measured adolescents’ reports of positive social connectedness across relevant contexts (family, friends, school). Of the 3,207 adolescents included, over 1/3 had at least one chronic health condition. Unadjusted and adjusted linear and logistic regression models of associations between the presence of chronic health conditions (any developmental/behavioral health condition and any physical health condition, compared to no conditions) and adolescents’ social connectedness outcomes were estimated. Compared to those with no chronic health conditions, adolescents with developmental/behavioral health conditions had lower odds of high positive social connectedness scores (AOR 0.80; CI: 0.67 - 0.94), having friends they really care about (AOR 0.76; CI: 0.61 - 0.94), having people who care (AOR: 0.65; CI: 0.50 - 0.84), and having people with whom to share good news (AOR: 0.77; CI: 0.63 - 0.94). Adolescents with chronic physical health conditions had lower odds of reporting having people who care about them (AOR: 0.72; CI: 0.55 - 0.94). The findings point to the need for interventions designed to foster the development of positive interpersonal relationships, reduce loneliness, and increase positive social identity among adolescents with chronic health conditions, particularly those with developmental/behavioral health conditions.

INTRODUCTION

Social connectedness, defined as positive psychological bonds with others and a feeling of belonging (Diendorfer et al., 2021), is important for adolescents’ optimal development, health, and long-term well-being (Chu et al., 2010; Diendorfer et al., 2021; Jose et al., 2012) and may be particularly important for those with chronic health conditions or disabilities. Substantial evidence points to links between positive social connections, especially relationships with friends, and better physical and mental health throughout the lifespan and protection against early mortality (Ehsan et al., 2019; Ho, 2016; Holt-Lunstad et al., 2010; Leon, 2005). Social disconnectedness has been linked to the development of chronic diseases, such as heart disease and diabetes, and has been identified as a key social determinant of health (Haslam et al., 2018). Positive social connections in family, peer, school, religious, and neighborhood settings have been associated with life satisfaction, self-esteem, mastery, and coping, and lower perceived stress and depressive symptoms (Rose et al., 2019).

Social connections of adolescents can be with friends, family members, teachers, or others and are associated with positive outcomes into adulthood (Bond et al., 2007; Steiner et al., 2019). Having more areas of connectedness (e.g., across family, peers, and school) has been associated with more favorable outcomes (Jose et al., 2012; Libbey et al., 2002), but strong social connections in even one area can be beneficial; e.g., secure friendship attachments during adolescence have been associated with fewer depressive symptoms in adulthood (Cook et al., 2016) and positive social connections with friends have been associated with increased treatment adherence to medical regimens, which can be vital for adolescents with chronic health conditions (La Greca et al., 2002). Supportive family connections have favorable effects on youths’ mental health (Gunn et al., 2018; Libbey et al., 2002), life satisfaction, confidence, positive affect, and aspirations (Jose et al., 2012), and strong parent-child bonds have been associated with academic achievement and psychological well-being in adulthood (Roksa & Kinsley, 2019; Stafford et al., 2016). Teacher support has been associated with youths’ positive self-concept, conduct, and psychological and social adjustment (Chu et al., 2010), and positive interactions with teachers have been associated with better mood, less substance use, and increased likelihood of completing school (Bond et al., 2007). Overall, there is abundant evidence that social connections that provide social and emotional support have positive effects on adolescent physical and mental health trajectories and academic achievement.

The mechanisms by which social connections appear to lead to better outcomes are by decreasing the burden of cumulative stress and adversity and increasing positive social identity through group membership and a sense of belonging (Haslam et al., 2018; Larrabee Sonderlund et al., 2019). The former is particularly important for those in disadvantaged groups (Haslam et al., 2018), and adolescents with disabilities are more likely than their non-disabled peers to have experienced adverse life events (Berg et al., 2019). Adolescents with disabilities are also more likely to experience negative social status in peer groups, which compromises individuals’ positive social identity through stigmatization or feeling a lack of belonging (Ferrie et al., 2020; Maxey & Beckert, 2017) and may limit opportunities for social interactions or lead to avoidance of informal social situations for fear of peer rejection (McNicholas et al., 2020). There is some evidence that youth with disabilities or health conditions that are not visible—such as anxiety, depression, and ADHD—are treated with even less acceptance than disabilities that are visible such as conditions requiring a wheelchair or walker (Davis, 2005; Hurley-Hanson et al., 2020; Werner, 2015). Youth with disabilities may also have fewer opportunities for social interactions—for example, if they have physical conditions that do not allow them to participate in sports or need to spend time in physical therapy or other treatments for their conditions. Finally, they may need family support for care and functioning which can potentially add to parental stress and reduce the quality of family relationships (Cousino & Hazen, 2013; F. Craig et al., 2016).

A recent review of developmental issues relating to adolescents with disabilities highlighted the salience of relationships with family and friends/peers, which are important for all adolescents as they begin the transition to adulthood (Maxey & Beckert, 2017). These relationships are widely studied in typically developing adolescents but much less so in the context of adolescent disabilities or chronic health conditions, which often confer additional challenges in forming and maintaining relationships (Maxey & Beckert, 2017). That is, youth with disabilities while needing strong social connectedness like all youth, may be at increased risk of having difficulties in multiple settings—informal social circles, school, and their households.

Between 2 and 24% of children in the United States are considered to have a disability, depending on the definition used (Byers et al., 2018). The lowest estimate (2%) is based on having “a medically determinable physical or mental impairment, which results in marked and severe functional limitations, and which can be expected to result in death, or which has lasted or can be expected to last for a continuous period of not less than 12 months” (Byers et al., 2018, p. xi) and household income low enough to qualify for the Supplemental Social Insurance program. The highest estimate (23.5%) pertains to children with special health care needs (Byers et al., 2018), which government agencies define as children “who have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally” (Byers et al., 2018; U.S. Department of Health and Human Services, n.d.). Special healthcare needs can reflect a wide array of physical and mental health conditions, many of which can be disabling (Byers et.al, 2018). Approximately 1 in 4 children ages 12 to 17 were identified in the National Children’s study as having special healthcare needs in 2019 (U.S. Department of Health and Human Services, n.d.).

Activity limitations due to chronic health conditions among U.S. children is the only measure of child disability for which national trends can be documented, and this measure more than tripled between 1960 and 2009. This trend largely reflects increases in cognitive and behavioral disorders, which now represent the dominant sources of child disability in the United States (Halfon et al., 2012). It is thought to reflect some combination of increases in the prevalence of cognitive and behavioral disorders as well as changes over time in the definitions of disability and rates of diagnosis. Activity limitations due to chronic conditions increase with age (Halfon et al., 2012), perhaps because some diagnoses are not made until children are older. These trends and patterns suggest that studies of determinants, correlates, or sequelae of child disability should focus on children of similar ages who were born in approximately the same year and, ideally, should consider cognitive and behavioral disorders separately from physical disabilities.

Studies have indicated lower levels of social competence, social functioning, and social connectedness in youth with chronic illnesses, such as cancer, obesity, and autism, but systematic reviews and meta-analyses concluded that the quality of most studies on the topic was mediocre at best (Diendorfer et al., 2021; Martinez et al., 2011; Pinquart & Teubert, 2012). More recent studies based on a national urban birth cohort found that youth with chronic and potentially disabling health conditions felt less connected to school, had more trouble getting along with teachers and peers, and had lower rates of participation in school and community activities compared to youth with no chronic conditions (James et al., 2021; Noonan et al., 2020). Another study of more than 1,600 high school students with a wide range of learning, emotional, and physical disabilities also found lower levels of school connectedness among these youth with disabilities (Forber-Pratt et al., 2020). Others found that students with various types of disabilities or chronic health conditions were at increased risk of experiencing peer victimization, including relational aggression (Forber-Pratt et al., 2020; James et al., 2021). Less representative studies found associations between child disability and having fewer friendships, more difficulty forming friendships, and lower quality friendships (Helms et al., 2015; Salas et al., 2018; Sedgewick et al., 2019).

Overall, most existing studies of adolescent disability or chronic health conditions and social connectedness focused on singular or specific types of conditions (e.g., autism, traumatic brain injury, or cystic fibrosis), used small convenience samples that were not representative of most youth with disabilities or with chronic health conditions, did not compare youth with disabilities to those with no disabilities, or relied on parental reports of adolescents’ social connections. In addition, few studies considered emotional/behavioral disorders beyond autism spectrum disorder.

In this study, we use rich population-based data from a national U.S. birth cohort study to comprehensively investigate associations between chronic health conditions that are potentially disabling and youths’ reports about their social connections at age 15 using a validated scale that measures positive social connectedness across relevant contexts (family, friends, and school). We explicitly recognize that not all social connectedness is positive; for example, peers can encourage substance use (Lee et al., 2017) and social media has been shown to exacerbate bullying behavior (W. Craig et al., 2020). However, we focus specifically on measures of positive social connectedness. We hypothesized that children with chronic health conditions would report less positive social connectedness than their peers with no chronic health conditions. The developmental pediatrician on our team categorized caregiver-reported potentially disabling health conditions as developmental/behavioral or physical. The importance of children’s social connectedness, which is essential for optimal development, came into the spotlight during the recent COVID-19 pandemic, which created an environment of high uncertainty and substantially reduced opportunities for social interactions (Lund et al., 2020). Youth with disabilities or potentially disabling health conditions may have been particularly vulnerable for the reasons elucidated earlier and their already high rates of isolation and loneliness (Maes et al., 2017). The findings from this study of adolescents in the pre-pandemic era has important implications for targeting interventions going forward.

METHODS

We conducted a retrospective study of associations between chronic developmental/ behavioral and physical health conditions and social connectedness of 15-year-old adolescents in a U.S. birth cohort using t-tests and linear and logistic regression analysis.

Data and Study Sample

We used data from the Future of Families and Child Wellbeing study (FFCWB), a national birth cohort study that randomly sampled births in large U.S. cities between 1998 and 2000. By design, about 3/4 of interviewed mothers were unmarried. Face-to-face interviews were conducted with 4,898 mothers while still in the hospital after giving birth (Data and Documentation | Fragile Families and Child Wellbeing Study, n.d.; Reichman et al., 2001). The baseline response rate was 86% among eligible mothers. Follow-up interviews were conducted 1, 3, 5, 9, and 15 years later. At 9 and 15 years, the child’s primary caregiver (in the vast majority of cases, the mother) provided information on the child’s health. At 15 years, the adolescents reported on their own relationships and social interactions.

Of the 4,898 children at baseline (birth), 3,429 adolescents completed the 15-year survey and had an associated completed primary caregiver survey. Of those, 222 had missing information on key analysis variables, leaving 3,207 adolescents in the analysis sample. Comparisons (at baseline) between the 3,207 cases in our sample to the 1,691 cases not in our sample indicated that children who were not in the sample were more likely to be born to mothers who had lower income and education, more likely to have mothers who were Hispanic, more likely to be immigrants and unmarried, and more likely to have been low birth weight (<2500 grams).

Exposure: Chronic Health Conditions

We considered developmental/behavioral health and physical health conditions using the mother’s or primary caregivers’ reports when the children were 9 and 15 years old (Appendix Table 1). We considered all disabilities reported by the primary caregiver when the child was age 15 and coded adolescents with attention deficit disorder/attention deficit hyperactivity disorder (ADD/ADHD), autism, developmental delay, seizure, epilepsy, depression, or anxiety as having a developmental/behavioral health condition. In addition, adolescents were coded as having a developmental/behavioral health condition if it was reported that the child had Down syndrome or cerebral palsy at year 9 (the last time these conditions were explicitly assessed).

Similarly, the developmental pediatrician on our team considered all disabilities reported by the primary caregiver when the child was age 15 and coded adolescents who had problems with limbs, heart disease, diabetes, high blood pressure, or anemia as having physical health conditions. We also coded adolescents as having a physical health condition if their primary caregiver reported that their activities at home, school, or work were limited because of allergies, digestive problems, headaches, ear infections, stuttering, or breathing difficulties. Primary caregivers were asked when the child was 15 if the child had any of these conditions, and if so which ones, and then whether the child had activity limitations from any of the 6 conditions. If a mother reported that her child had activity limitations from any of the 6 conditions, but the only condition of the 6 that she reported was stuttering, we did not consider that child to have a physical health condition based on question about activity limitations (this applied to only 5 children). Since some questions about physical health were not asked at age 15, adolescents were coded as having a physical health condition if any of the following conditions were reported about them at age 9: sickle cell anemia, blindness (partial or full), or deafness (partial or full).

Outcomes

The main outcomes in this study were based on the Connectedness sub-scale of the EPOCH Measure of Adolescent Well-Being, which was validated in 10 studies of youth aged 10-18; Cronbach’s alpha was between .77 and .81 (Kern et al., 2016) and embedded in the adolescent 15-year FFCWB survey. The acronym EPOCH stands for five aspects of positive functioning: Engagement, Perseverance, Optimism, Connectedness, and Happiness; connectedness refers to the extent to which an individual “has satisfying relationships with others; believ(es) that one is cared for, loved, esteemed, and valued; and provid(es) friendship or support to others” (Kern et al., 2016, p. 587).

Each adolescent was asked to consider and report on how much the following four statements, all referring to the past four weeks, pertained to them: “I have friends that I really care about;” “There are people in my life who really care about me;” “When something good happens to me, I have people who I like to share the good news with;” “When I have a problem, I have someone who will be there for me.” For each question, the adolescent’s response was assigned 1, 2, 3, or 4 points (1 = strongly agree, 2 = somewhat agree, 3 = somewhat disagree, and 4 = strongly disagree). As recommended in the FFCWB documentation (User’s Guide for the Fragile Families and Child Wellbeing Study Public Data, Year 15, 2021), we reverse-coded the four response options (assigning points as follows: 1 = strongly disagree, 2 = somewhat disagree, 3 = somewhat agree, and 4 = strongly agree) and created a score that was the individual’s average number of points across the four questions. Thus, the possible range of scores was 1 point (strongly disagreed with all four statements) to 4 points (strongly agree with all 4 statements), with a higher score indicating feeling more connected. We also constructed binary measures using the overall score as well as each of the four questions that equaled 1 if the individual’s score was 4 and 0 otherwise. We assessed the sensitivity of the findings to alternative binary cutoffs.

In additional models, we considered alternative relevant outcomes using two questions from the Social Skills Rating System (SSRS), which was modified for the FFCWB adolescent survey to capture self-reports rather than parent or teacher reports. The SSRS includes seven social skills subdomains: Communication, Cooperation, Engagement, Assertion, Responsibility, Empathy, and Self-Control (Gresham et al., 2011). We considered 2 questions (from the Assertion domain) that involve social connections. Specifically, adolescents were asked whether each of the following statements was not true, sometimes true, or often true: “I make friends easily” and “I am liked by others.” For each, we constructed a binary outcome that was assigned a value of 1 if the adolescent’s response was “often true” and 0 otherwise. Overall reliability estimates for subscales for the student SSRS forms indicate a median alpha level of .71 (Crosby, 2011).

Covariates

All potentially time-varying covariates were from the 15-year survey. In adjusted models, we controlled for the adolescent’s gender, age in years at the time of the 15-year primary caregiver interview, and race-ethnicity; the primary caregiver’s age, education, employment status, and marital status; whether the primary caregiver was the child’s biological mother; and the number of children and poverty status of the household.

Analysis

First, we compared the outcomes and sample characteristics of adolescents with no chronic health conditions to those with any developmental/behavioral or physical health condition, any developmental/behavioral health condition, and any physical health condition. Statistically significant differences between adolescents with any developmental/behavioral or physical health condition and those with no chronic health conditions were ascertained using 2-tailed t tests for comparisons of means, with p =. 05 as the threshold.

Second, we estimated unadjusted and adjusted linear and logistic regression models of associations between the presence of chronic health conditions (any developmental/behavioral health condition and any physical health condition, compared to no conditions) and adolescents’ social connectedness outcomes. The health condition variables were not mutually exclusive, so having one type of condition did not preclude having the other type. Estimates for the continuous outcomes are presented as Ordinary Least Squares (OLS) coefficients and 95% confidence intervals (CIs) and for the binary outcomes as odds ratios (OR) and 95% CIs from logistic regression models.

Third, we estimated supplementary models that alternatively included an indicator for having more than one health condition (of any type), included an indicator for having at least one of both types of conditions, included measures of specific types of health conditions, and used alternative cutoffs for the binary outcomes. We also estimated models separately by the adolescent’s gender. All analyses were conducted using Stata Version 15.0 statistical software. The authors’ Institutional Review Boards determined this study to be exempt.

RESULTS

Main Analyses

In our sample, 34.9% (1,119/3,207) of adolescents had a chronic developmental/behavioral or physical health condition (Table 1). Estimates from the National Survey of Children’s Health indicate that about 28% of children aged 12 to 17 in the United States in 2017 had chronic physical, developmental, behavioral or emotional conditions (National Survey of Children’s Health, 2017).

Table 1:

Adolescent Connectedness Outcomes and Sample Characteristics by Presence of Adolescent Health Conditions

		Full Sample	No chronic health conditions	Any developmental or behavioral or physical health condition(s)
		N=3,207	N=2,088	N=1,119
Adolescent Outcomes
Positive adolescent functioning (EPOCH Connectedness Scale)
	Mean score, which ranged from 1 to 4 (s.d.)*	3.77 (0.36)	3.79 (0.34)	3.75 (0.39)
	Score = 4 (vs. < 4)*	0.56	0.58	0.54
Components of EPOCH Connectedness Scale (binary indicator: strongly agree, vs. agree, disagree, or strongly disagree)
	Has friends he/she cares about	0.82	0.83	0.80
	Has people who care about him/her*	0.89	0.91	0.86
	Has people to share good news with*	0.79	0.81	0.77
	Has someone who will be there when there is a problem	0.75	0.76	0.74
Components of Social Skills Rating System (SSRS) Assertion Sub-Scale (binary indicator: often true, vs. sometimes true or not true)
	Makes friends easily*	0.60	0.63	0.55
	Is liked by others*	0.64	0.67	0.60
Adolescent characteristics
Male*		0.51	0.48	0.57
Age, mean in years (s.d.)		15.57 (0.76)	15.57 (0.76)	15.57 (0.74)
Race-ethnicity
	Non-Hispanic white*	0.17	0.15	0.21
	Non-Hispanic black*	0.47	0.48	0.43
	Hispanic	0.24	0.24	0.22
	Other race-ethnicity	0.07	0.07	0.08
	Missing race*	0.05	0.04	0.07
Primary caregiver characteristics
Age, mean in years (s.d)		40.98 (6.22)	41.11 (6.28)	40.74 (6.10)
Education
	< High school graduate	0.17	0.18	0.16
	High school graduate	0.19	0.19	0.20
	Some college	0.44	0.44	0.46
	College graduate	0.19	0.20	0.19
Employed*		0.72	0.73	0.69
Married*		0.39	0.41	0.36
Household characteristics
Adolescent lives with biological mother		0.93	0.92	0.93
# children in household, mean (s.d.)		2.63 (1.55)	2.58 (1.55)	2.63 (1.55)
Household income, % of federal poverty line
	< 100%*	0.34	0.35	0.34
	100-199%*	0.25	0.25	0.25
	> 200%*	0.41	0.40	0.41

Notes: All figures are proportions unless indicated otherwise.

S.d. = standard deviation.

All time-varying characteristics were assessed when the child was 15 years old. See Appendix Table 1 for information about coding of health conditions.

^*Statistically significant differences between the 2,088 adolescents who had no health conditions and the 1,119 adolescents who had any developmental/behavioral or physical health condition(s) using 2-tailed t tests for comparisons of means, with p = .05 as the threshold for statistical significance.

In our sample, 755 adolescents (23.5%) had developmental/behavioral health conditions, 567 (17.7%) had physical health conditions, and 203 (6.3%) had both (Appendix Table 1). Almost half of adolescents (45.4%) with developmental/behavioral conditions had depression or anxiety (343/755) and about 2/3 (507/755 = 67.2%) had ADD/ADHD. Almost half the adolescents with physical health conditions had at least one of the conditions for which activity-limiting conditions were assessed (262/567 = 46.2%). Of the adolescents with any health condition, 34% had more than one (not shown).

Overall, adolescents without any chronic health conditions were significantly more likely to report having positive social connections (Table 1). Compared to those with any chronic health conditions, adolescents with no conditions were significantly more likely to have people who care about them and with whom to share good news. They were also more likely to make friends easily and report being liked by others. Adolescents with any health conditions were more likely to be male and white compared to those with no conditions. The primary caregivers of adolescents with any health conditions were less likely to be employed, less likely to be married, and more likely to reside in households with incomes below the poverty level compared to adolescents with no chronic health conditions.

Adjusted estimates of associations between adolescents’ health conditions and their Connectedness scale values indicate that having a developmental/behavioral health condition was associated with significantly lower scores (OLS coefficient: −.05; CI: −.08 to −.02) and lower odds of scoring an average of 4 (highest level of connectedness) on that scale (Adjusted Odds Ratio (AOR) 0.80; CI: 0.67 - 0.94) and the following components: having friends they really care about (AOR 0.76; CI: 0.61 - 0.94), having people who care about them (AOR: 0.65; CI: 0.50 - 0.84), and having people with whom to share good news (AOR: 0.77; CI: 0.63 - 0.94) (Table 2, with full adjusted logit results in Appendix Table 2). Adjusted estimates also indicate that having a chronic physical health condition was associated with significantly lower odds that the adolescents reported having people who care about them (AOR: 0.72; CI: 0.55 - 0.94).

Table 2:

Adjusted and Unadjusted Regression Estimates of Associations Between Adolescent Health Conditions and Adolescent Connectedness Outcomes Based on EPOCH Connectedness Scale (N=3,207)

	Positive Adolescent Functioning: EPOCH Connectedness Scale		Components of EPOCH Connectedness Scale (strongly agree, vs. agree, disagree, or strongly disagree)
	Average score (ranges from 1 to 4)	Score = 4 (vs. < 4)	Has friends he/she cares about	Has people who care about him/her	Has people to share good news with	Has someone who will be there when there is a problem
	OLS coefficient [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]
PANEL A Unadjusted
Any developmental or behavioral health condition(s)	−0.05 [−0.08 - −0.02]	0.78 [0.66 - 0.92]	0.74 [0.60 - 0.91]	0.67 [0.52 - 0.86]	0.74 [0.61 - 0.90]	0.99 [0.82 - 1.20]
Any physical health condition(s]	−0.02 [−0.05 - 0.02]	0.95 [0.79 - 1.14]	1.00 [0.79 - 1.27]	0.69 [0.53 - 0.90]	0.89 [0.71 - 1.11]	0.90 [0.73 - 1.11]
PANEL B Adjusted
Any developmental or behavioral health condition(s)	−0.05 [−0.08 - −0.02]	0.80 [0.67 - 0.94]	0.76 [0.61 - 0.94]	0.65 [0.50 - 0.84]	0.77 [0.63 - 0.94]	0.98 [0.81 - 1.19]
Any physical health condition(s)	−0.02 [−0.05 - 0.02]	0.96 [0.80 - 1.16]	1.03 [0.81 - 1.30]	0.72 [0.55 - 0.94]	0.89 [0.72 - 1.12]	0.91 [0.74 - 1.13]

Notes: In each panel, the columns present estimates from a single regression model containing the two chronic health condition measures.

OLS = Ordinary Least Squares. CI = confidence interval. OR = odds ratio.

Adjusted models control for all adolescent, primary caregiver, and household characteristics in Table 1. See Appendix Table 1 for information about coding of health conditions. See Appendix 2 for full adjusted logit regression results.

In analyses using the outcomes from the SSRS, adjusted models indicated that adolescents with developmental/behavioral health conditions had lower odds of making friends easily (AOR: 0.71; CI: 0.59 - 0.84) and of reporting that they were liked by others (AOR: 0.70; CI: 0.59 - 0.83) compared to adolescents with no developmental/behavioral health conditions (Table 3 and Appendix Table 2). There were no significant differences in outcomes between adolescents with any physical health condition(s) and those with no physical health conditions.

Table 3:

Adjusted and Unadjusted Logistic Regression Estimates of Associations Between Adolescent Health Conditions and Components of SSRS Assertion Sub-Scale (N=3,207)

	Makes friends easily (often true vs. sometimes true or not true)	Is liked by others (often true vs. sometimes true or not true)
	OR [95% CI]	OR [95% CI]
PANEL A Unadjusted
Any developmental or behavioral health condition(s)	0.73 [0.62 - 0.86]	0.76 [0.64 - 0.90]
Any physical health condition(s]	0.85 [0.70 - 1.02]	0.83 [0.68 - 1.00]
PANEL B Adjusted
Any developmental or behavioral health condition(s)	0.71 [0.59 - 0.84]	0.70 [0.59 - 0.83]
Any physical health condition(s)	0.87 [0.72 - 1.05]	0.86 [0.71 - 1.04]

Notes: In each panel, the columns present estimates from a single regression model containing the two chronic health condition measures.

OR = odds ratio; CI = confidence interval.

Adjusted models control for all adolescent, primary caregiver, and household characteristics in Table 1. See Appendix Table 1 for information about coding of health conditions. See Appendix 2 for full adjusted regression results.

Adolescents who were male had significantly lower odds than their female counterparts of scoring a 4 on the Connectedness scale (highest level of connectedness), as well as having friends they cared about and having people with whom to share good news (Appendix Table 2). They also had higher odds of making friends easily and being liked by others (Appendix Table 2). Adolescents who were non-Hispanic black had significantly lower odds of having friends they cared about (Appendix Table 2) and higher odds of making friends easily compared to those who were non-Hispanic white (Appendix Table 2).

Supplementary Analyses (not shown)

Analyses that included a measure of having more than one chronic health condition (any type) in addition to the indicators for each type of condition revealed that having more than one condition significantly decreased the odds of making friends easily (AOR: 0.69; CI: 0.51 - 0.94) and being liked by others (AOR: 0.70; CI: 0.52 - 0.95). In models that added an indicator for having both types of conditions (at least one of each type) in addition to the separate indicators for each type, the estimates for having both types of conditions were never statistically significant, i.e., having both types of conditions was not associated with any of the outcomes above and beyond the associations for having each type of condition.

Models that included indicators for depression or anxiety and ADD/ADHD (the most common developmental/behavioral conditions) as well as for any other type of developmental/behavioral condition in addition to the indicator for any physical health condition showed that: (1) Depression or anxiety was associated with significantly lower scores on the Connectedness scale as well as significantly lower odds of having a Connectedness score of 4 (AOR: 0.77; CI: 0.60- 0.98), having people who care (AOR: 0.50; CI: 0.36 - 0.70), making friends easily (AOR: 0.63; CI: 0.50 - 0.81), and being liked by others (AOR: 0.70; CI: 0.55 - 0.89). (2) ADD/ADHD was associated with significantly lower odds of reporting being liked by others (AOR: 0.72; CI: 0.59 - 0.89) and having friends they care about (AOR: 0.69; CI: 0.54 - 0.80).

Models that included an indicator for the most commonly-reported physical health condition—activity limitations from allergies, digestive problems, headaches, ear infections, or breathing difficulties—in addition to separate indicators for any other type of physical health condition and any developmental/behavioral condition indicated that the measure of activity limitations was associated with significantly lower odds of making friends easily (AOR: 0.74; CI: 0.57- 0.96) and having people who care (AOR: 0.56; CI: 0.40 - 0.79).

Several models using alternative binary outcome measures were estimated. For the Connectedness scale, adjusted estimates indicated that having a developmental/behavioral health condition was significantly associated with lower odds of scoring at least 3.75 (AOR 0.71; CI: 0.58 - 0.86) and of scoring at least 3.50 (AOR 0.69; CI: 0.54 - 0.88). In models that used alternative binary outcomes that equaled 1 if the adolescent’s response was “strongly agree” or “somewhat agree” and 0 otherwise for estimating each of the components of the Connectedness scale (“I have friends that I really care about,” “There are people in my life who really care about me,” “When something good happens to me, I have people who I like to share the good news with,” “When I have a problem, I have someone who will be there for me”), adolescents with a developmental/behavioral condition had lower odds of having people who care about them (AOR: 0.58; CI: 0.42 - 0.79) and having people to share good news with (AOR: 0.51; CI: 0.31 - 0.85) compared to those with no chronic developmental/behavioral conditions. Models using alternative binary outcomes for the relevant components of the SRSS Assertion sub-scale (“I make friends easily” and “I am liked by others”) that equaled 1 if the adolescent’s response was “often true” or “sometimes true” and 0 otherwise indicated that adolescents with a developmental/behavioral health condition had lower odds of making friends easily (AOR: 0.48; CI: 0.36 - 0.65) and being liked by others (AOR: 0.43; CI: 0.26 - 0.70) compared to adolescents with no chronic health conditions. Having a chronic physical health condition was not significantly associated with any outcome in this entire set of specifications using alternative binary outcomes.

DISCUSSION

In a national population-based urban birth cohort, chronic developmental/behavioral health conditions among 15-year-olds were strongly associated with lower social connectedness overall, and specifically with reports of not having friends they care about, not having anyone with whom to share good news, not able to make friends easily, and not being liked by others. In contrast, in the same cohort, chronic physical health conditions were not associated with any of these outcomes, but both chronic developmental/behavioral conditions and chronic physical conditions were associated with youth reports of not having people who cared about them. These are markedly concerning findings given the importance of positive social connections and belonging for all adolescents, particularly those with disabling conditions, and the implications for their overall well-being.

The greater salience of developmental/behavioral conditions than of physical health conditions for forming friendships and positive social connections may be related to deficits in social skills (Diendorfer et al., 2021; Craig et al., 2016), or fewer opportunities for interaction. There is evidence that social skill interventions can be helpful in improving social inclusion and connection among adolescents with intellectual disability and autism spectrum disorder (Diendorfer et al., 2021; Louw et al., 2020). However, it is not clear whether these improvements in social interactions would apply to youth with other developmental disabilities or whether they would lead to the development of friendships and other social connections. Participation in special education classrooms is associated with having fewer friends (Schoop-Kasteler & Müller, 2020), and the majority of Individuals with Disabilities Education Act (IDEA) services provided in U.S. schools in the 2017-2018 school year were students with developmental/behavioral conditions (NCES, 2020). As such, students with developmental/behavioral conditions may be more likely than those with physical health conditions to be in special education classrooms (NCES, 2020). However, even when included in general classrooms, many youth with disabilities—especially disabilities that are more severe—tend to be pulled out of class, arrive late, or leave early and thus have less exposure to other students and missed opportunities for peer and teacher interactions (Feldman et al., 2016). In addition, previous research has found that adolescents with developmental/behavioral health conditions participate in fewer school and community activities (Noonan et al., 2020), which leads to fewer interactions with peers, teachers, and coaches. Overall, the opportunities for peer interaction available to youth with disabling health conditions appear to play a stronger role than social skills or classroom type alone, although robust and conclusive research in this area is lacking. In addition, as noted previously, less visible disabilities including many behavioral health conditions, are characterized by behaviors that interfere with social skills and peer connections that may contribute to the different findings for the two broad disability types. To our knowledge, this distinction has not been examined in prior studies of adolescents’ interpersonal connections and is an important contribution to this body of research.

We found that both adolescents with developmental/behavioral conditions and those with physical health conditions had lower odds of reporting that they had people who cared about them. Adolescents answering this question may have been considering their peers, family members, teachers, school aides or other staff, members of their faith community, or other people in their lives in their responses, but we are not able to distinguish between those. Adolescence is a time of heightened need for independence and autonomy, but some adolescents with chronic health conditions may need care and support beyond the level required by their peers who do not have health conditions; as such, in some cases, adolescents with chronic health conditions may be sensitive to stresses that family members or aides exhibit when providing care for them. Some adolescents with physical health conditions may require demanding treatment regimens, which have been associated with parental stress (Cousino & Hazen, 2013). While youth with developmental/behavioral health conditions may not require physical assistance, the emotional and behavioral support required to manage limitations in social and cognitive skills and problem behaviors that accompany some conditions have been associated with parental stress (Craig et al., 2016). Receipt of support could be seen as intrusive and possibly cause stress and lower self-esteem; i.e., the actual support one receives may be less effective than the perception of that support (Chu et al., 2010). In other words, it is important that adolescents are able to balance support needs with age-appropriate autonomy. Relatedly, there was no difference between youth with chronic developmental/behavioral conditions and those with physical conditions (in both cases, compared to youth without such conditions) in reporting having someone available if they needed them. In our study, youth with and without chronic health conditions reported having someone available if they need them, but those with chronic health conditions were less likely to report having someone who cares about them. A person being available is quite a different experience than having someone provide support with emotional care, and this could indicate that youth with chronic conditions may have aides, school staff, or family available to support their care or educational needs, but do not feel the emotional connection that comes with reciprocal relationships or enough opportunities for social connection outside of their specific needs. This is an important finding that deserves additional attention in future research and practice.

The findings from this study build and expand upon prior research finding associations between specific health conditions and social connections (Diendorfer et al., 2021; Martinez et al., 2011; Pinquart & Teubert, 2012; Maes et al., 2017; Salas et al., 2018; Sedgewick et al., 2019; Helms et al., 2015) by considering chronic health conditions more broadly and at the population level and classifying them as developmental/behavioral or physical, but also considering specific conditions that were prevalent in our population-based sample. This way, we were able to include conditions that may be relatively rare or less severe than those that have been previously studied. We found much stronger negative associations with social connectedness-related outcomes for developmental/behavioral versus physical health conditions, strong negative associations for depression or anxiety and ADHD in particular, and some negative associations for activity limitations owing to allergies, digestive problems, headaches, ear infections, or breathing difficulties. Robust links between child disability or chronic health conditions and social connectedness have not previously been established at the population level.

Our findings point to the importance of ensuring that youth with developmental/behavioral health conditions have opportunities for social inclusion and participation in school and community activities, as well as the development of positive interpersonal relationships (Louw et al., 2020). Our findings indicating deficits in social connectedness among adolescents with developmental/behavioral conditions, many of which may be less visible and less accommodated than physical conditions, suggest a need for positive identity development and decreased social isolation for those youth. For example, youth with mental illness may experience stigma-related difficulties connected to their need for social identity and connectedness, such as peers who reject them, not feeling like they fit in, feeling shame, attempting to hide their condition, and avoiding help-seeking (Ferrie et al., 2020). This lack of social connectedness can then worsen mental health outcomes (Ferrie et al., 2020) and potentially physical health and educational outcomes. Thus, targeted interventions to reduce loneliness and increase positive identity are key for adolescents with developmental/behavioral conditions.

Youth with chronic developmental/behavioral conditions may need some support from adults such as family members or teachers to facilitate social interactions. Positive youth development programs that include sustained adult-youth connections have shown psychosocial and health-related benefits for adolescents with chronic conditions (Maslow & Chung, 2013). Structured and organized interventions with opportunities for interpersonal relationships and community participation show the most favorable outcomes for friendship development for youth with disabilities (Louw et al., 2020). Interventions that specifically target loneliness in adolescents have shown effectiveness (Eccles & Qualter, 2021), as has Groups 4 Health, a novel intervention demonstrating decreased loneliness and depression among adolescents (Cruwys et al., 2022; Haslam et al., 2018). Despite some safety concerns and barriers to accessing technology, the internet and social media can potentially foster exchanges with peers, leading to positive social interactions and the development of social connectedness (Diendorfer et al., 2021; Louw et al., 2020).

Interventions that increase school and family connectedness are also important for youth with disabilities given the findings of this study and previous research on chronic health conditions or disability of youth and school disconnectedness (James et al., 2021) and family stress (Cousino & Hazen, 2013; Craig et al., 2016). This could include reducing peer victimization and increasing time with teachers for support and connection (James et al., 2021) as well as greater resources and support for parents to manage problematic behaviors and cope with stressors (Craig et al., 2016). Family teamwork for spreading out caregiving duties and adding support, such as aides, to assist with caregiving needs when possible could also reduce parental stress (Cousino & Hazen, 2013). All of these interventions can enhance positive social identity and a sense of belonging, which are core components of social connectedness and are associated with enhanced physical and mental health outcomes in later years (Ehsan et al., 2019; Ho, 2016; Leon, 2005). Interventions that aim to build positive identity, provide social and psychological support, and buffer stress are vital to changing these youths’ trajectories.

Strengths of this study include the focus on a population-based national birth cohort; distinguishing between youth with and without chronic health conditions and with different types of conditions; controlling for a rich set of adolescent, primary caregiver, and household characteristics; measuring social connectedness using a validated scale not previously used with this population; and utilizing youth reports of relationships and friendships. Limitations are that the findings may not be generalizable to all populations and settings, the associations may be confounded by unobserved factors, and the chronic health conditions included and categorized together as developmental/behavioral or physical are heterogeneous and do not include all disability types.

CONCLUSION

Strong social connectedness during the teenage years across multiple relationship contexts can have long-term consequences for future health and well-being. Compared to their peers with no chronic health conditions, adolescents with chronic developmental/behavioral health conditions had lower odds of reporting having friends, having someone with whom to share good news, being able to make friends easily, and being liked by others, while adolescents with chronic developmental/behavioral and physical health conditions both had lower odds of reporting they had people who care about them. These findings point to a need for targeted interventions to provide opportunities for positive social interactions for youth with chronic health conditions or disabilities and ensure that families have adequate support when caring for youth with chronic health conditions.

PUBLIC POLICY RELEVANCE STATEMENT.

In this study of U.S. urban youth, adolescents with chronic developmental/behavioral health conditions had lower odds of positive social interactions compared to their peers with no chronic health conditions. The findings point to the need for interventions designed to foster the development of positive interpersonal relationships, reduce loneliness, and increase positive social identity among adolescents with chronic health conditions, particularly those with developmental/behavioral health conditions.

ABBREVIATIONS

ADD/ADHD

Attention Deficit Disorder/Attention Deficit Hyperactivity Disorder

AOR

Adjusted Odds Ratio

CI

Confidence Interval

FFCWB

Future of Families & Child Wellbeing Study (formerly Fragile Families & Child Wellbeing Study)

OLS

Ordinary Least Squares

Appendix Table 1:

Chronic Developmental or Behavioral and Physical Health Conditions in the Fragile Families and Child Wellbeing Study Analysis Sample (3,207)

		Ages condition assessed		# coded positive
		9 years	15 years	(based on latest time point condition was assessed)
Developmental or behavioral conditions
	ADD/ADHD	X	X	507
	Autism	X	X	60
	Down syndrome	X		4
	Developmental delay	X	X	9
	Seizure and epilepsy	X	X	52
	Cerebral palsy	X		5
	Depression or anxiety		X	343
Any developmental or behavioral condition(s)				755
Physical health conditions
	Sickle cell anemia	X		30
	Blind (full or partial)	X		28
	Deaf (full or partial)	X		25
	Problem with limbs	X	X	104
	Heart disease	X	X	87
	Diabetes	X	X	29
	High blood pressure		X	17
	Anemia	X	X	83
	Usual activities limited at home, school, or work because of allergies, digestive problems, headaches, ear infections, stuttering, or breathing difficulties		X	262
Any physical health condition(s)				567
Both developmental or behavioral and physical health conditions				203

Note: The conditions were assessed from reports by mothers about the types of health conditions the child had, if any. The surveys included specific response choices and those were not necessarily medical terms (e.g., “problem with limbs”). The developmental pediatrician on our team coded the response choices into conditions that are chronic (or likely to be chronic based on the information available) and those that are not, and then categorized the conditions that were chronic (or likely to be chronic) as developmental/behavioral or physical. Some of the conditions coded as behavioral/developmental are developmental disabilities with associated physical impairments (e.g., Down Syndrome and cerebral palsy). This table includes all chronic (or likely to be chronic) conditions reported by the mothers for the children in our sample.

Appendix Table 2:

Full Set of Logistic Regression Estimates of Associations Between Adolescent Health Conditions and Friendships/Informal Social Interactions (N = 3,207)

		Connected Scale Score = 4 (vs. < 4)	Teen has friends they care about	Teen has people who care about him/her	Teen has people to share good news with	Teen has someone who will be there when there is a problem	Teen Makes Friends Easily	Teen is liked by others
		OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]
Any developmental or behavioral health condition(s)		0.80*** [0.67 - 0.94]	0.76** [0.61 - 0.94]	0.65*** [0.50 - 0.84]	0.77** [0.63 - 0.94]	0.98 [0.81 - 1.19]	0.71*** [0.59 - 0.84]	0.70*** [0.59 - 0.83]
Any physical health condition(s)		0.96 [0.80 - 1.16]	1.03 [0.81 - 1.30]	0.72** [0.55 - 0.94]	0.89 [0.72 - 1.12]	0.91 [0.74 - 1.13]	0.87 [0.72 - 1.05]	0.86 [0.71 - 1.04]
Teen characteristics
Male		0.84** [0.73 - 0.96]	0.68*** [0.56 - 0.82]	1.20 [0.95 - 1.50]	0.73*** [0.61 - 0.87]	1.06 [0.90 - 1.25]	1.56*** [1.35 - 1.81]	1.50*** [1.29 - 1.74]
Age (years)		1.02 [0.92 - 1.12]	0.99 [0.88 - 1.11]	0.90 [0.78 - 1.04]	1.02 [0.91 - 1.15]	1.00 [0.89 - 1.11]	0.99 [0.90 - 1.09]	1.08 [0.98 - 1.19]
Race-ethnicity
	Non-Hispanic black	0.85 [0.68 - 1.06]	0.58*** [0.43 - 0.80]	0.99 [0.69 - 1.43]	1.01 [0.77 - 1.31]	0.98 [0.76 - 1.26]	1.46*** [1.17 - 1.83]	0.98 [0.78 - 1.24]
	Hispanic	0.87 [0.68 - 1.10]	0.77 [0.55 - 1.07]	1.01 [0.68 - 1.49]	1.06 [0.78 - 1.42]	0.90 [0.68 - 1.18]	1.26* [0.99 - 1.61]	0.91 [0.71 - 1.17]
	Non-Hispanic other	0.74* [0.54 - 1.01]	0.80 [0.52 - 1.24]	0.63* [0.40 - 1.01]	0.83 [0.57 - 1.19]	0.81 [0.57 - 1.16]	1.39** [1.01 - 1.91]	0.86 [0.62 - 1.19]
	Missing race	0.94 [0.65 - 1.36]	0.58** [0.36 - 0.92]	1.13 [0.63 - 2.04]	1.54* [0.94 - 2.52]	1.01 [0.66 - 1.54]	1.00 [0.69 - 1.44]	0.58*** [0.40 - 0.84]
Primary caregiver characteristics
Age (years)		1.00 [0.99 - 1.02]	1.01 [0.99 - 1.02]	0.99 [0.97 - 1.01]	1.00 [0.98 - 1.01]	1.00 [0.99 - 1.01]	0.99 [0.98 - 1.01]	1.00 [0.98 - 1.01]
Education
	High school graduate	1.07 [0.84 - 1.36]	1.23 [0.90 - 1.66]	1.30 [0.90 - 1.87]	1.36** [1.01 - 1.83]	1.05 [0.80 - 1.37]	1.14 [0.89 - 1.45]	1.38** [1.08 - 1.76]
	Some college	1.12 [0.90 - 1.38]	1.07 [0.82 - 1.40]	1.24 [0.91 - 1.70]	1.11 [0.87 - 1.43]	1.13 [0.89 - 1.43]	1.03 [0.83 - 1.27]	1.30** [1.05 - 1.61]
	College graduate	1.16 [0.89 - 1.52]	1.39* [0.97 - 1.98]	1.30 [0.86 - 1.96]	1.24 [0.89 - 1.72]	1.00 [0.73 - 1.35]	1.22 [0.93 - 1.60]	1.62*** [1.23 - 2.15]
Employed		1.08 [0.91 - 1.28]	0.97 [0.78 - 1.20]	1.27* [0.98-1.641	1.09 [0.88 - 1.34]	1.17 [0.96 - 1.41]	1.06 [0.89 - 1.26]	1.02 [0.85 - 1.21]
Married		1.02 [0.87 - 1.20]	0.89 [0.72 - 1.09]	1.07 [0.83 - 1.38]	1.05 [0.86 - 1.27]	1.11 [0.93 - 1.34]	1.14 [0.97 - 1.34]	1.01 [0.85 - 1.19]
Household characteristics
Lives with biological mother		0.97 [0.73 - 1.29]	0.87 [0.59 - 1.27]	1.08 [0.69 - 1.68]	0.94 [0.66 - 1.33]	0.92 [0.66 - 1.27]	0.81 [0.61 - 1.08]	0.78 [0.58 - 1.05]
# children		1.01 [0.96 - 1.07]	1.05 [0.98 - 1.12]	0.96 [0.89 - 1.03]	1.01 [0.95 - 1.07]	1.02 [0.96 - 1.08]	1.01 [0.96 - 1.06]	1.01 [0.95 - 1.06]
Income, % of federal poverty line
	< 100%	0.80** [0.65 - 1.00]	0 64*** [0.49 - 0.84]	0.87 [0.62 - 1.23]	0.91 [0.70 - 1.18]	1.01 [0.79 - 1.29]	0.89 [0.72 - 1.11]	0.83* [0.66 - 1.04]
	100-199%	0.92 [0.76 - 1.11]	0.84 [0.65 - 1.08]	0.85 [0.62 - 1.16]	0.92 [0.73 - 1.16]	0.96 [0.77 - 1.19]	0.87 [0.71 - 1.05]	0.74*** [0.61 - 0.91]

Notes: Each column present results from a separate regression model that corresponds to those in Panel B of Table 2.

CI = confidence interval.

Reference categories: Male (for teen’s gender), non-Hispanic white (for teen’s race-ethnicity), less than high school (for primary caregiver’s education), and >200% of the federal poverty line (for household income).