Pediatric obesity is one of the most significant preventable public health problems in the United States [1
]. It has been consistently associated with poor physical (e.g., type 2 diabetes, cardiovascular disease) and psychological outcomes (e.g., mood disorders) later in life [2
]. Thirty-five percent of youth aged 2–19 years old are currently classified as overweight or obese (Body Mass Index (BMI) for age and gender ≥ 85th percentile), with data from the 2015–2016 cycle of the National Health and Nutrition Examination Survey (NHANES) highlighting persisting health disparities for Hispanic youth [5
]. Specifically, Hispanics have the highest overall rates of pediatric overweight and obesity (46%) compared to non-Hispanic black (38%), white (30%), and Asian (23%) youth [5
]. These rates correspond to low levels of physical activity and poor dietary quality [6
]. Developing, evaluating, and disseminating evidence-based obesity prevention interventions for Hispanics, estimated to comprise 28% of the U.S. population by the year 2060 [9
], is thus critical to the future health of the nation.
The family may be the most fundamental social system influencing child health and development [10
], and is especially relevant for Hispanics due to familismo
(i.e., a Hispanic cultural value placing a strong emphasis on family loyalty and commitment) [12
]. A recent systematic review of obesity-related randomized controlled trials conducted with Hispanic youth aged 5–19 years old identified 11 interventions published between 2010–2015, the majority of which (72%) included some sort of family-based component [13
]. Although family-based pediatric weight loss studies have been found to successfully reduce youth BMI [14
], stronger effects are often detected in those with higher versus lower rates of attendance [15
]. Notably, poor attendance in both clinical and research settings has been cited as one of the most common challenges of pediatric obesity interventions, with the highest rates of attrition observed among low income and ethnic minority populations [18
]. Understanding participant attendance patterns, as well as factors predicting those patterns, is therefore necessary to implement strategies that will maximize the public health impact of family-based obesity prevention programs, particularly among Hispanic youth at greatest risk for being overweight and obese.
1.1. Intervention Attendance: Rates versus Patterns
Most studies investigating intervention attendance (and predictors thereof) in the field of pediatric obesity have operationalized it as a sum (total number of sessions attended) or a rate (percentage of total sessions attended), e.g., References [23
]. While informative, such approaches are limited in that they do not provide important temporal information regarding participant attendance patterns and trajectories over the course of an intervention. As noted by Mauricio et al. [25
], understanding attendance trajectories and profiling families based on these patterns may help interventionists identify when and for whom declines in attendance are likely to occur.
Examples of studies that capture attendance more dynamically as a pattern over time can be found in the broader family-based prevention science literature [25
]. For instance, a study of attendance patterns among mothers and fathers enrolled in a 10-session parenting-focused preventive intervention for divorced/separated parents used a latent class growth analysis to identify four classes of attendance: non-attenders, early dropouts, declining attenders, and sustained attenders [25
]. This study found that mothers who were early dropouts were more likely to be Hispanic and report higher parental conflict than those who were sustained attenders [25
]. Similar classes of attendance have been observed in other family-based preventive interventions targeting youth internalizing, externalizing, and risky behaviors (e.g., substance use, sexual risk-taking) among exclusively Hispanic samples [26
]. To our knowledge, however, no prior obesity-related intervention studies targeting Hispanic youth have examined participant attendance patterns.
1.2. Predictors of Participant Attendance
Understanding what predicts participant attendance in family-based interventions is complex, as attendance may be influenced by a broad range of variables, including individual characteristics of both the target youth and their parents, family dynamics, and the broader cultural variables, e.g., References [23
]. Our approach to examining predictors of attendance was therefore grounded in the eco-developmental framework [32
]. The eco-developmental framework is a contextual schema that organizes risk and protective factors across multiple ecological systems. It draws from social ecological theory [33
], which emphasizes that human behavior is influenced by factors ranging from those most proximal to the individual (i.e., microsystems, or contexts in which the individual participates directly) through to those more distal to the individual (i.e., macrosystem, or the broader cultural and societal context). In this study, we examined predictors of attendance patterns in a family-based obesity prevention intervention across three levels: individual (adolescent and parent), family, and the broader cultural context.
1.2.1. Individual-Level Predictors of Attendance
Strong evidence suggests sociodemographic factors impact attendance, with higher levels of parental education, greater household income, and partnered (vs single) parenting consistently associated with higher attendance and retention in family-based obesity interventions [18
]. Families with girls versus boys have also been shown to be more likely to initiate and complete obesity-related programs [35
]. Other individual-level variables, such as the health status and lifestyle behaviors of both parents and youth at baseline may also be associated with attendance, though relatively little research has examined these associations. In a study that investigated demographic and psychosocial predictors of attrition in a family-based weight control trial among adolescents aged 13–16 years old, for example, parent BMI was the most robust predictor of attrition [37
]. Specifically, families with a parent who had a higher versus lower BMI were 4.6 times more likely to drop out of the intervention. Another study found that the child’s healthy eating at baseline was positively associated with attendance in a family-based lifestyle intervention for overweight children [24
]. Together, these studies suggest that families of adolescent girls who have more resources/support and are physically “healthier” at intervention onset may be more likely to have sustained attendance.
1.2.2. Family-Level Predictors of Attendance
Other key factors potentially influencing attendance in family-based pediatric obesity interventions may be related to the functioning of the family unit. For example, Williams and colleagues [18
] found that those classified as program non-completers and partial completers had higher levels of family dysfunction (e.g., family members avoiding each other when at home) than did those classified as program completers. Due to the limited number of studies examining the influence of family variables on attendance in obesity-related studies, we drew from the broader family-based prevention science literature. Family-level variables, including poorer family communication, lower family organization (i.e., the degree of order in the home), and higher levels of conflict, have been previously associated with an increased likelihood of intervention dropout [29
], poorer initial engagement (defined as attending one of the first three sessions) [38
], and lower program participation and retention [31
]. These or similar family variables, including low use of effective parenting skills, however, have also been found to positively predict attendance in family-based substance use and problem behavior preventive interventions due to potentially being associated with a greater intervention need [30
]. Overall, more research that examines the effects of family-level variables on attendance in obesity prevention interventions is needed.
1.2.3. Cultural Level Predictors of Attendance
The family-based prevention literature additionally points to acculturation as an important cultural factor to consider in interventions targeting Hispanic immigrant families. Acculturation, or the processes of immigrants adapting to their host country’s culture, may influence a family’s willingness to participate in an intervention program [44
]. Less acculturated individuals, or those who identify more strongly with their culture of origin and less strongly with their host culture, tend to participate in culturally syntonic family-based interventions at higher rates [26
]. In a recent study of attendance patterns among Mexican-Americans participating in a mental health and substance use prevention program, for example, those with lower versus higher levels of acculturation were more likely to belong to the sustained attendance group [28
]. Although studies with more diverse Hispanic samples have found similar associations [30
], these studies are not as well-represented in the literature as are those with exclusively Mexican-American samples.
1.3. Current Study and Hypotheses
In the context of a culturally relevant family-based obesity prevention intervention for Hispanic adolescents known as Familias Unidas for Health and Wellness [45
], the aims of the current study were to: (1) identify subgroups of participant attendance patterns, and (2) examine the role of individual (i.e., adolescent gender, parent education, annual household income, adolescent and parent weight status and healthy lifestyle behaviors), family dynamics (i.e., positive parenting, parent-adolescent communication), and cultural variables (i.e., acculturation, Hispanicism, Americanism) in distinguishing attendance patterns. We hypothesized that multiple attendance patterns would emerge from our data and that families with the following characteristics at baseline would be most likely to consistently attend intervention sessions: female adolescent gender, higher parent education, higher annual household income, lower parent and adolescent weight status, healthier adolescent and parent lifestyle behaviors, more positive parenting, better parent-adolescent communication, and less acculturation.
This study examined participant attendance patterns as well as individual, family dynamics, and cultural predictors of these patterns among 140 Hispanic families enrolled in a 12-session family-based obesity prevention intervention. Results indicated the presence of three distinct attendance patterns: consistently low (average of <1 session attended), moderate and decreasing (average of 4.5 of 12 sessions attended), and consistently high (average of 11 of 12 sessions attended), with the overwhelming majority of participants in the consistently high attendance group. Parents in the consistently high attendance group reported lower Americanism than parents in either the moderate and decreasing or consistently low attendance groups. For adolescents, those in the consistently high attendance group reported lower Hispanicism than those in either the moderate and decreasing or consistently low attendance groups. No other baseline variables, including adolescent gender, parent education, annual household income, adolescent and parent weight status, adolescent and parent healthy lifestyle behaviors (i.e., physical activity, fruit and vegetable intake, added sugar intake), positive parenting, and parent-adolescent communication, significantly discriminated between attendance groups. Results suggested that sustained attendance in the Familias Unidas for Health and Wellness intervention may be driven by Hispanic parents’ desire to better understand their host culture, connect with other culturally similar parents, and reconnect their adolescents with their heritage culture.
To our knowledge, this is the first study in the field of obesity prevention to examine attendance as a pattern, rather than as a sum or a rate, e.g., References [23
]. Visually examining the attendance patterns presented in Figure 1
reinforced the importance of attendance at the first few sessions, particularly among those in the moderate and decreasing attendance group, whose attendance to group sessions was modest from the outset but then steeply declined following the second group session. This type of information is inherently missing from studies that present attendance as a sum or a rate and may be used to determine optimal times for encouraging or incentivizing participants to continue attending sessions.
Notably, the majority of participants randomized to the intervention condition (72%) belonged to the consistently high attendance group, suggesting that once most families began the intervention, they continued returning to sessions. A recent review of barriers and facilitators to initial and continued attendance in community-based lifestyle programs for overweight and obese youth concluded that a family-centered approach (where both parents and youth attend sessions), practical sessions (e.g., hands on cooking activities and physical activity), and social interaction and support were important factors influencing continued attendance [67
]. All of these characteristics are central to the present intervention and were previously endorsed by those participating in the initial pilot/feasibility study as important features of the program [45
]. Specifically, parents described feeling engaged by intervention content and social connections, reported enjoying hands-on nutrition training activities, perceived improvements in family cohesion, and described the intervention social climate as being very positive.
Beyond identifying patterns of attendance, we examined a series of eco-developmental variables at baseline that might distinguish between the attendance groups. Unlike previous obesity-related studies [24
], sociodemographic variables, namely parent education, annual household income, and marital status did not significantly predict participants’ attendance group. This finding suggested that in a culturally-specific intervention for Hispanics like Familias Unidas for Health and Wellness, socioeconomic status (SES) may not be as important for attendance as it is in interventions that are not culturally syntonic or that do not target specific ethnic groups. Although not significant, it is noteworthy that those in the consistently low attendance group reported the lowest income and highest levels of Hispanicism. Because we recruited participants from two somewhat distinct areas (a middle class, more established area and a lower-income area with a higher concentration of recently-arrived immigrants), the consistently low attendance group may have been comprised primarily of recently-arrived immigrants whose ability to dedicate time to a family-based intervention may be more limited.
Like SES variables, neither weight status nor healthy lifestyle behaviors at baseline for either parents or adolescents distinguished between the attendance groups, despite indicating a high need for the intervention. Specifically, all adolescents and the majority of parents (86%) were in an unhealthy weight range. Adolescents self-reported engaging in 60 min/day of physical activity only about half the week and consuming less than 2 cups/day of fruits and vegetables. Similarly, parents self-reported consuming high amounts of added sugars (14.7 teaspoons/day). These findings suggested that when recruiting for an intervention such as Familias unidas for Health and Wellness, focusing on intervention health benefits and participants’ potential need may not be as important as focusing on culture. Alternatively, many of the non-significant findings in the current study, particularly those related to participants’ weight status and healthy lifestyle behaviors, may be explained by the largely homogenous sample that resulted from study recruitment and selection procedures (e.g., enrolling adolescents in the overweight or obese weight range only).
Acculturation variables, specifically parents’ Americanism and adolescents’ Hispanicism, were the only variables to distinguish among the three attendance patterns, highlighting the importance of taking acculturation into account as it relates to Hispanics’ participation in preventive interventions, e.g., References [26
]. Acculturation is comprised of two dimensions (receiving culture acquisition, heritage culture retention), each of which occur in three domains (practices, values, identifications) [68
]. Findings from the current study indicated that parents with lower receiving culture acquisition (i.e., lower Americanism) and adolescents with lower heritage culture retention (i.e., lower Hispanicism) were the most likely to belong to the consistently high attendance group. Importantly, 90% of parents and approximately 40% of adolescents enrolled in the intervention were immigrants from Latin American countries including Cuba, Nicaragua, Honduras, and Venezuela. Previous research suggests that culturally relevant interventions may present immigrant parents with opportunities to better understand numerous aspects of their host culture (e.g., parenting in the U.S., education and the school system) and simultaneously forge new social connections with other culturally similar adults [69
]. It is therefore possible that parents viewed the intervention as an opportunity to connect with other similar adults, better understand American practices and values, and reconnect their children with their heritage culture. Similarly, it is possible that adolescents viewed the intervention as an opportunity to connect with other similar adolescents and for parents to understand their perspective.
This study has several limitations. Although the sample of Hispanics is diverse and represents immigrants from numerous Latin American countries, all of them were recruited from a single geographical location (Miami, FL, USA), limiting the generalizability of findings. Measures of parent and adolescent healthy lifestyle behaviors, including physical activity, fruit and vegetable intake, and added sugar intake, were self-reported and thus subject to bias. Beyond being self-reported, our measure of dietary intake may have failed to accurately capture the healthfulness of many foods (e.g., cereals with varying grams of sugar). In addition, alpha reliability for the positive parenting measure in this sample was low. Finally, while attendance measured an important aspect of intervention participation, it did not capture how actively participants engaged with session content. A growing body of literature is examining predictors of intervention “responsiveness”, or “dynamic engagement”, which includes constructs such as attendance, satisfaction, completion of outside practice assignments, and active participation in sessions [40
This study also has various notable strengths. It adds to a growing body of literature on attendance and retention within pediatric obesity interventions and highlights how a repeated measures latent class analysis can be used to conceptualize attendance over the course of an intervention. It also uses a large sample of Hispanic families participating in an ongoing efficacy trial to understand attendance in a population for whom previous studies have reported problems with intervention attendance, retention, and drop out. Finally, the study assessed relevant predictors of attendance patterns in both parents and children rather than just one or the other.