The Impact of Caring and Connectedness on Adolescent Health and Well-being

by M. D. Resnick, L. J. Harris and R. W. Blum 1993

Children Youth and Family Consortium Electronic Clearinghouse. Permission is granted to create and distribute copies of this document for noncommercial purposes provided that the author and CYFCEC receive acknowledgment and this notice is included. Phone (612) 626-1212 EMAIL: cyfcec@maroon.tc.umn.edu

By M. D. Resnick (1), L. J. Harris (2) and R. W. BLUM (3)
(1) Division of Health Management and Policy, (2) National Adolescent Health Resource Centre and (3) Division of General Pediatrics and Adolescent Health, University of Minnesota, Minneapolis, Minnesota, United States
Correspondence: M. D. Resnick, University of Minnesota, Box 97 UMHC, 420 Delaware Street SE, Minneapolis, Minnesota 55455, USA.
Accepted for publication 23 July 1993.

Abstract

This study of over 36 000 7th-12th grade students focused on protective factors against the quietly disturbed and acting out behaviors, which together represent the major social morbidities of adolescence. Multivariate models developed separately for girls and boys repeatedly demonstrated the protective function of caring and connectedness in the lives of youth, particularly a sense of connectedness to family and to school. A sense of spirituality, as well as low family stress (referring to poverty, unemployment substance use and domestic violence) also functioned as protective factors. Measures of caring and connectedness surpassed demographic variables such as two parent vs. single parent family structure as protective factors against high risk behaviors.

Interventions for youth at-risk must critically examine the ways in which opportunities for a sense of belonging may be fostered, particularly among youth who do not report any significant caring relationships in their lives with adults.

Key words: adolescent health; caring; protective factors; resiliency.

Numerous reports have documented the shift from biological to social causes of morbidity and mortality among adolescents. These trends belie the traditional view that adolescence is a time of optimal health and well-being (1-7). Other trends, particularly the rising proportion of children and adolescents living in poverty, patterns of school drop out and poor school performance, disaffection, alienation, and the pervasive impact of racism and limited economic opportunities lend urgency to the need for communities to respond to the health and social needs of their youth, and for policy makers and funders to help assure a better fit between health and social programs and the youth problems they are intended to address (8-16).

In 1986, a National Invitational Conference on the Health Futures of Adolescents was convened with support from the Maternal and Child Health Bureau in Washington DC. The goal of that conference was:

to identify the major health related concerns that will face youth through the year 2000, delineate the present state of knowledge in each of the major issues, and develop a blueprint for the next 15 years for research, demonstration programming and training (17).

The report on that conference which was published in the November 1988 supplement of the Journal of Adolescent Health Care featured the recommendations of seven work groups, compiled from the 136 experts participating in the conference from the fields of public health, medicine, clinical and developmental psychology, social ecology and health education, nursing nutrition, sociology and social work. One key recommendation which cross-cut the work group reports was the need for development of comprehensive population-based studies of adolescent health problems, concerns, risk behaviors and resiliency so that prevention and intervention programs could be responsive to the most salient issues affecting youths within communities. Further, it was recommended that interventions aimed at promoting adolescent health and reducing risk behaviors respond to the clustering of those behaviors, with an understanding that social morbidities tend to co-occur, contrary to the assumptions of narrowly defined categorical programs (18). On an international scale, the same proposal was incorporated as a top priority in the recommendations of the Scientific and Technical Advisory Group to the Adolescent Health Section of the World Health Organization. This recommendation emphasized the value of developing comprehensive, population based studies of adolescent health and risk behaviors so that communities, states and nations could prioritize among adolescent health issues, identify protective factors that reduce the likelihood of health compromising behaviors, and move effectively to promote the well-being of adolescents at greatest risk (19).

In the United States, the Maternal and Child Health Bureau funded a major initiative called the Adolescent Health Database Project, to demonstrate the feasibility of establishing a statewide database of health status indicators, risk behaviors, and adolescent health concerns. This project enabled the development, testing, and widespread use of a standardized instrument (the Adolescent Health Survey) in order to develop a comprehensive adolescent health database for program and policy development, planning and research purposes. The survey has been administered to over 60 000 adolescents nationally, including statewide surveys in Minnesota, Alaska, Delaware, and a national survey of rural, reservation based American Indian youth. The purpose of this comprehensive assessment of adolescent health, risk behaviors and resiliency factors is to provide valid, timely information to key decision makers and information users, including legislators, health, social service and education professionals, youth workers, parents, and others involved with or on behalf of youth.

The primary vehicle for dissemination of information obtained from these surveys has been the development of local feedback reports to each participating school or community, as well as publications designed for professional, scholarly audiences (20-27). These have been augmented by the development of numerous specialty reports in the form of monographs that have been distributed to thousands of decision makers at the state, regional, national and international levels (28-33). Data from the Adolescent Health Survey have been utilized for advocacy, curriculum. program and policy development at the state and national levels, and for grant writing by community based agencies that have utilized the survey data to identify key health issues among adolescents, prioritize among them, and seek funding to initiate or further develop services for youth.

More recently, attention is being turned to analyses of these large scale datasets that focus on resiliency and protective factors in the lives of young people. Much of scientific inquiry in adolescent health has traditionally focused on the correlates of problemness or pathology. Here, a focus on resiliency means that inquiry is directed toward understanding success and well-being, identifying those factors that buffer against the stresses of everyday life that might otherwise result in adverse physical, social or psychological outcomes for youth (34-38). Longitudinal studies of resiliency have incorporated biological, psychological, familial, and broader social variables into models that predict who, in adolescence or adulthood will be characterized by high functional effectiveness and life satisfaction. A repeated finding in these analyses has been the centrality of caring relationships between children and adults, including relationships within and outside the family, for the development of resilient adolescents and young adults (39, 40).

This focus on caring as a protective factor presents an important direction for programs, policy and practice. The shift toward social morbidities among young people means that the major threats to their health and well-being are increasingly rooted in the organization, economics, opportunities and expectations of everyday life. This means that the search for protective factors against a variety of adverse outcomes must include an understanding of adolescents' social relationships and feelings of connections to others as they experience and live the developmental changes of their physical, social and psychological selves. Unfortunately, the presence of nurturing relationships between adults and children cannot be treated as a given. Rather, problems stemming from youth disaffection and alienation are seen as a growing by-product of post-modern society. Sociologist Phillip Slater made the observation over 20 years ago that three human desires are deeply and uniquely frustrated by Western culture (41). These include:

1. The desire for community-the wish to live in trust and fraternal co-operation with one's fellows in a fetal and visible collective entity.
   The desire for engagement-the wish to come directly to grips with social and interpersonal problems and to confront on equal terms an environment which is not composed of ego-extensions.
2. The desire for dependence—the wish to share responsibility for the control of one's impulses and the direction of one's life.
3. Reflecting these sentiments, Michael Lerner describes the misguided emphasis on individuality in Western culture, which, like Slater, he regarded as frustrating the deeply felt yearning for M. D. Resnick et al. connectedness between people, working to create meaning and happiness in the context of an interdependent community of human beings (42).

Methods

The datasource for this analysis is the Minnesota Adolescent Health Database, derived from a sample of 36,254 7th-12th grade public school students throughout Minnesota who completed the Adolescent Health Survey. This comprehensive questionnaire elicits self-report information from adolescents in the following areas: demographic and biographical data: relation ships with family, friends and other adults; school performance and conduct: personal worries and concerns; body image; help seeking and utilization of services; nutrition and eating behavior; disordered eating; sexual behaviors; sexual orientation; substance use; mental health and suicidal involvement, physical and sexual abuse; anti-social behaviors; other risk-taking behaviors.

Questionnaire development, content, and psychometric properties of scales and indices are described in other publications (20, 21, 23-28). Schools were selected through a multi-stage stratified cluster sampling design, stratified by school district size, with random sampling occurring within each stratum for each grade level. In each school, all students within designated grades were asked to complete the questionnaire within a classroom setting.

Trained survey administrators were available to answer questions, to ensure standardized test administration, and to protect confidentiality.

Scales and risk behaviour clusters

Initial work on the dataset focused on development and refinement of measures to be used in subsequent analyses related to caring and connectedness. Standard data reduction and scaling techniques were used to develop psychometrically robust indicators of key constructs including six separate measures related to caring as well as aggregated scales of caring and connectedness. These measures were tested for internal consistency, construct validity and examined for their distributional properties. Final, usable scales and indicators were created as both continuous and ordinal measures, to facilitate their use in cross-tabular form and in multivariate analyses (Harris, L., Resnick, M. D., Rosenwinkel, K., and Glum, R. W, Technical Report on the Adolescent Health Survey: Psychometric Properties of Scales and Indices. Minneapolis: University of Minnesota Adolescent Health Training Program, 1990, unpubl. data).

Drawing upon theoretical and empirical work that has identified the co-occurrence of high-risk behaviors among adolescents, composite variables were created to tap the major morbidities of adolescents as behavioral clusters (13, 43-46). Nine individual scales were subject to factor analysis, identifying two broader clusters of risk behaviors, each with factor loadings of 0.5 or greater (noted in parentheses after each variable name). These clusters included Acting Out Behaviors and Quietly Disturbed Behaviors. Acting out behaviors included polydrug use (0.737), school absenteeism (0.684), risk of unintentional injury (0.677), (e.g. drinking and driving, not wearing seat-belts, use of motorcycles or all-terrain vehicles without a helmet, driving in the back of open pick-up trucks), pregnancy risk (0.606), (including risk of either becoming pregnant or causing a pregnancy, based on a scaled combination of frequency of sexual intercourse and type and frequency of contraceptive use), and delinquency risk (0.556). Quietly disturbed behaviors included poor body image (-0.800), disordered eating (0.785), (hinging, deliberately vomiting as a strategy for weight loss, chronic dieting, fear of loss of control of eating), emotional stress (0.686), and suicidal involvement (0.514), (ideation or actual attempts). Generally, boys were more characterized by the acting out behaviors, girls by the quietly disturbed behaviors, although there was cross-over by gender, particularly among girls.

For example, of those girls who exhibited two or more quietly disturbed behaviors, 80% were at high risk for at least one acting out behavior. Of those engaging in two or more acting out behaviors, 65% also engaged in at least one quietly disturbed behavior. Some 80% of students fell into the high-risk category for at least one of the behaviors, with about 10% at high-risk for four or more behaviors.

With the completion of scale development and assessment of the clustering of high-risk behaviors, the identification of protective factors for each cluster was accomplished through the development of four discriminant function models, permitting a multivariate identification of variables that best sort or differentiate comparison groups. Protective factors were separately assessed for boys and girls against each of the two clusters of high risk behaviors. Particular attention was directed to the presence of measures of caring and connectedness as salient protective factors across each of these analyses. It was hypothesized that in comparison with demographic variables, the indicators related to caring and connectedness would emerge as more powerful protective factors against high-risk behaviors.

Results

Results are presented for the four stepwise discriminant models which describe protective factors in the order of their explanatory power. As noted in Table 1, a common set of factors in each of the four discriminant models related to various aspects of connectedness. As hypothesized, these measures were more powerful protective factors than demographic variables against both clusters of high-risk behaviors, for both boys and girls.

Among girls, the most powerful protective factor against the quietly disturbed behaviors was family connectedness, referring to adolescents who indicated they enjoyed, felt close to and cared for by family members. This variable alone explained 12.5% of variance in group classification. School connectedness, the second explanatory variable in the stepwise equation, referred to students who enjoyed school, experiencing a sense of belonging and connectedness to it (which did not always correspond with high academic performance). Family stress, the third explanatory variable, was a composite measure of parental unemployment, poverty, domestic violence, and parental substance use, with low family stress functioning as a protective factor. Spiritual connectedness referred to those students who defined themselves as spiritual or religious individuals. The last factor, age, was positively related to risk, with younger adolescents indicating less involvement in quietly disturbed behaviours than their older counterparts. Together, these five variables correctly classified 71.8% of adolescents, including close to 9 in 10 of those at low risk for quietly disturbed behaviours (Table 2).

A similar set of explainer variables functioned as protective factors against the acting out behaviors for girls, as described in Table 3. In addition, the last variable to remain in the stepwise model indicated that girls in two parent families were less likely than those in single parent families to be involved in acting out behaviors. These five variables correctly classified 71.2% of girls, including close to 8 in 10 at high risk for acting out behaviors.

Among boys, three protective factors against the quietly disturbed behaviors were identified, including family connectedness, school connectedness, and low family stress. These correctly classified 71.2% of boys overall, including 72% of those at low risk for quietly disturbed behaviors, and 64% of those at high risk.

The final discriminant model for acting out behaviors in boys included the full set of variables found in the girls' model for this behavior cluster. School connectedness was the single most powerful variable in the equation, with two parent family entering as the last explanatory variable in the model.

Discussion

Protective factors were identified for quietly disturbed and acting out behaviors, representing two clusters of health compromising behaviors that encompass the major social morbidities of adolescents. Separate analyses were conducted for boys and girls. Across all four models, measures of caring and connectedness predominated as protective factors. While family structure and socioeconomic status are both prominent in popular explanations of high-risk behaviors among youth (expressed most frequently as concern about the “breakdown” of families and family values, particularly among resource-deprived families), socioeconomic status did not remain in any of the models, meaning that adolescents could not be differentiated as low or high-risk for quietly disturbed and acting out behaviors on the basis of their families' socioeconomic status, after the preceding variables were taken into account. As far as two parent vs. single parent status was concerned, two parent family composition proved to be the weakest, though significant, explanatory variable in the models describing protective factors against the acting out behaviors. Results indicated that boys and girls from two parent families were somewhat less likely to be involved in such behaviors than their counterparts from single parent families.

The most powerful protective factors across models were family and school connectedness. What is striking about the family connectedness variable is that this factor referred to a sense of belonging and closeness to family, in whatever way family was comprised or defined by the adolescent. Thus, the centrality of families in the promotion of well-being among young people was reaffirmed, but without specifying the form or composition that families must take in order to serve this protective function. At the core of family connectedness is the adolescent's experience of being connected to at least one caring, competent adult in a loving, nurturing relationship. Similar results have been reported by investigators assessing resiliency and well-being among youth who otherwise would be expected to be at high risk for multiple adverse health and social outcomes (47-52).

School connectedness was the most salient protective factor for both boys and girls against the acting out behaviors, second in importance after family connectedness for the quietly disturbed behaviors. These findings underscore the importance of schools as a primary source of connectedness with adults and with the broader community as perceived and experienced by the adolescent.

The analysis does not reveal whether a sense of connectedness with school resides in relationships with particular teachers, coaches, or other personnel, or through a generalized feeling of belonging within the overall school environment. With school connectedness superseding family connectedness as a protective factor against the acting out behaviors, we infer that schools can and do play a vital role in reducing the likelihood of health-jeopardizing behaviors among girls and boys by providing a sense of belonging that may not also be provided by other sources such as family or peers. Supporting this inference from cross-sectional data is the fact that academic remediation is foremost among intervention strategies directed at reducing the risk of such acting out behaviors as teenage pregnancy and juvenile delinquency (10-12, 13, 54). But the findings also suggest that academic performance is but one component of this sense of connectedness, since the underlying construct was composed of both an indicator of typical school performance and attitude toward school. Thus, for students who may not be academically proficient, and for whom there may be only marginal room for improvement, it is particularly important that schools provide vehicles that promote a sense of belonging, by providing opportunities to develop and demonstrate other forms of competency, including work-study, technical skills, and involvement in visual, musical and dramatic arts. The salience of school connectedness as a protective factor against adolescent high-risk behaviors strengthens the arguments of educators, health officials, and youth advocates that there must be closer collaboration between the health and education sectors in order to promote both the well-being and educability of young people (10, 55, 56).

The importance of religious or spiritual connectedness in these multivariate assessments is consistent with other analyses that have demonstrated that adolescents who describe themselves as more religiously observant or affiliated with religious institutions are less likely to engage in high-risk behaviors than their counterparts (43). This finding suggests the important roles that parents and religious institutions can play in fostering spirituality in young people, as well as a sense of belonging and connectedness with religious institutions which can enable positive peer and adult relationships and social experiences (57).

There are important research questions that flow from these analyses. Specifically, at what point in the life trajectory of an adolescent are the protective effects of caring and connectedness simply too late? In other words, if a young person has been socialized in a climate of uncertainty, fear and disaffection, can adult connectedness, positive environmental consistency, and the nurturance of confidence and competence, which are all the hallmarks of successful interventions, overcome the impetus toward distress and self-destructive behaviors? Pediatrics as a field has always maintained that early intervention is preferable over addressing serious needs after-the-fact. In public health, primary prevention is far superior to secondary prevention, given the opportunity for early identification and intervention with people at risk. So, for health and social service professionals, educators, youth workers and parents, the critical intervention question becomes: in the absence of opportunities or initiatives for primary prevention of the social morbidities of youth, what ‘package' of interventions will be most effective beyond the point of primary prevention, when young people have already embarked on behaviors that seriously increase the likelihood of ‘rotten outcomes'? We know from the literature on program for youth at risk, that for interventions to successfully deter adolescents away from destructiveness and lowered lifelong effectiveness, the intervention must be as intense as the need itself (9). Our analyses indicate that fostering a sense of caring and connectedness between adolescents and adults should be an integral part of interventions designed to promote resiliency and protective factors, increase adolescents' competency and effective functioning, and promote a sense of meaningful place in the world.

How this should best be done for very high-risk youth populations, in our view, frames the pre eminent human services delivery questions for the 1990s. Such questions are reminiscent of the treatment-outcome models generated in pharmacy, medicine, and mental health research over the past decade: what kinds of interventions or experiences, provided to which group of adolescents, result in optimal out comes? In other words, what works best with whom?

While these analyses identify caring and connectedness as essential components of health promotion, we recognize, as Mechanic recently noted, that love alone cannot rectify a lifetime of neglect (58). Caring, while extraordinarily important in the lives of young people, is not a substitute for correcting fundamental threats to health, rooted in the economic disparities that have become increasingly manifest due to both deliberate government policies and a shifting economic infrastructure that strains the ability of families and individuals to thrive or function. In fact, the sense of disaffiliation of growing numbers of young people, noted with alarm by many commentators (18), is accentuated by trends in the economy which make it increasingly harder for families to earn sufficient income with benefits, to meet the costs of housing, food, and everyday living. Our models point to the protective function of low family stress, and each of the elements of family stress measured within this construct, including unemployment, substance abuse and domestic violence, are directly associated with economic deprivation. Once deficits in connectedness as well as heightened levels of family stress were adjusted or accounted for, socioeconomic status did not enter into the multivariate assessments of protective factors against health-compromising behaviors. But lifting families out of poverty remains a most feasible strategy for health promotion that can be addressed on the policy level. As an age group, economic hardship is most keenly felt by the young. Recently released Census Bureau data in the United States underscore the widening economic gap between old and young people spawned during the 1980s, showing that, as a group, elderly people generally held their own during the economic fluctuations of the decade, while the nation's children increasingly slipped into poverty. Of those who became poor in the decade of the 1980s, 25% were under 18 years of age, while one in 25 was age 65 or older, owing in good measure to the indexing of Social Security payments to inflation (59) and the fact that the United States made a deliberate decision to eliminate poverty in a large proportion of the elderly population through age-related entitlements. No similar commitment has been made to children and youth.

The need for redress of this economic disparity was clearly articulated in the United States by the National Commission on Children in its call for progressive economic policies that would benefit children, youth, and families (60). This bipartisan proposal represents an important step beyond the popular political rhetoric that explains poverty and its accompanying morbidities for children and youth with a conservative moral determinism that exclusively blames those at highest risk for their own predicament, with the accompanying assertion that as far as remedy is concerned, 'nothing works' (61). But the salience of caring and connectedness as protective factors against the social morbidities of adolescents also suggests that more than an economic determinism is needed to promote adolescent health and well-being.

We maintain that while reducing the prevalence of poverty must remain an enduring goal for pro-child and pro-family policy, those who craft and implement interventions to reduce the quietly disturbed and acting out behaviours must also deliberate on how and whether the elements of their interventions address the underlying need for adolescent belonging. With the urge toward connectedness representing one of our deepest human desires (42), caring as a conscious, explicit quality must pervade the people and programs that seek to optimize the life course of adolescents, particularly those at highest risk.

Acknowledgment

The authors would like to acknowledge the Lilly Endowment, Inc., the Minnesota Women's Fund and the Maternal and Child Health Bureau for their support of this research.