The overarching goal of the Arizona Twin Project addresses a central question of developmental and clinical psychology — namely, how resilience (the capacity to bounce back following adversity) develops and affects the impact of early risk on child physical health and common mental health disorders such as anxiety, depression, conduct problems, and attention deficit hyperactivity disorder. The need to understand these relations is underscored by findings that 9–15% of toddlers and preschoolers exhibit clinical or sub-clinical emotional or behavioral problems (Briggs-Gowan et al., Reference Briggs-Gowan, Carter, Skuban and Horowitz2001; Campbell, Reference Campbell1995; Egger & Arnold, Reference Egger and Angold2006) that interfere with daily family life, and tend to remain stable across childhood (Briggs-Gowan et al., Reference Briggs-Gowan, Carter, Bosson-Heenan, Guyer and Horwitz2006; Campbell, Reference Campbell1995; Keenan et al., Reference Keenan, Shaw, Delliquadri, Giovanelli and Walsh1998). There is a critical need for scientifically valid studies that can inform parents, policy makers, and practitioners about causal processes in child health and psychopathology. By focusing on components of early resilience within a representative twin study design, we can elucidate processes by which children bounce back after adversity and grow to be healthy and competent individuals.
The Arizona Twin Project is a longitudinal study focusing on the impact of the early environment (prenatal and infancy periods) on developing resilience across childhood. The primary goal of this research is to understand genetic and environmental influences on early resilience and physical and mental health, as well as the association between resilience and health problems and/or competencies. Early health problems considered in the Arizona Twin Project include dysregulated mood, sleep, and eating, as well as aggression, inattention, and hyperactivity. Aspects of early competency include positive mood, attentional focusing, inhibitory control, prosocial behaviors, empathy, compliance, imitative play, and motivation. Elucidating the important components of early resilience and how relations between resilience and health are developed and maintained by genes and environments informs our understanding of etiology, and aids in identifying specific genes and environments linked to problems and competencies.
Risk factor research has dominated efforts to predict and ultimately prevent child mental health problems across both phenotypic and genetically informed studies, and this work has been invaluable in identifying key sources of vulnerability. Nevertheless, there are levels of disorder that cannot be accounted for in the accumulation of risk indices. For example, genetically influenced traits, such as self-regulation, confer health advantage after accounting for multivariate risk ratios (Moffitt et al., Reference Moffitt, Arsenault, Belsky, Dickson, Hancox, Harrington and Caspi2011), and positive environmental influences, such as social support or positive family climate, may be instrumental in offsetting genetic or environmental risk for disorder (Silk et al., Reference Silk, Vanderbilt-Adriance, Shaw, Forbes, Whalen, Ryan and Dahl2007). Thus, resilience theory leads researchers to develop indices of positive adaptation that complement the identification of risk factors, and emphasizes that to best predict health, both protective and risk factors must be considered. To this end, the Arizona Twin Project seeks to identify components of resilience across person, family, and cultural levels of analysis, and to determine their association with children's health.
Although numerous studies have established associations between risk factors and child mental health, research that considers resilience from a genetically informed perspective is still relatively rare (Lemery-Chalfant, Reference Lemery-Chalfant, Reich, Zautra and Hall2010). As an early twin study assessing multiple aspects of the prenatal and early postnatal family and social environments, the Arizona Twin Project is in a position to examine both genetic and environmental contributions to the development of resilience across domains, including physiological health and regulation (e.g., sleep and eating), attention, emotional reactivity and regulation, and early pro-social behavior and empathy. In addition, we test models examining the complex interplay between heritability and environments, such as moderation by relevant environments (e.g., obstetrical complications, parenting warmth, cultural ecological context).
Another key aspect of the Arizona Twin Project is our focus on the development of early self-regulation. Self-regulation is a component of resilience that likely mediates or moderates the relations between risk and optimal functioning across a variety of domains. After reviewing a substantial body of literature linking regulation with developmental outcomes, the National Academy of Science committee report, From Neurons to Neighborhoods, noted that ‘the growth of self-regulation is a cornerstone of early childhood development that cuts across all domains of behavior’ (Shonkoff & Phillips, Reference Shonkoff and Phillips2000, p. 3). Despite the importance of self-regulation to health and well-being, research devoted to understanding how self-regulation enhances health in the context of adversity is limited.
Given the importance of a representative sample for estimating heritability and genetic and environmental covariance, we recruited from state birth records, rather than volunteer organizations. As far as we know, this project is the only early childhood twin study in the United States with a large percentage of Mexican Americans (25%), the largest and the most rapidly growing ethnic group in the United States. This demographic affords the opportunity to consider the impact of acculturation, and the interaction of culture and genetics.
We began the longitudinal study in infancy because mounting evidence suggests the prenatal and early postnatal environment have instrumental roles in establishing biological set points of recovery from adversity or stress (Fox et al., Reference Fox, Levitt and Nelson2010), and we hope to elucidate the relevant environments. The next longitudinal follow-up was at 30 months, because 30 months is an important age for the development of early resilience. A follow-up assessment at 30 months allows us to examine whether twins at risk at 12 months of age due to prematurity or developmental immaturity have now matured past these infancy issues, and to examine environmental factors that predict this recovery. Furthermore, 30-month-olds, unlike 12-month-olds, are able to use verbal communication to regulate their emotions and behavior and enhance their social relationships, two key components of resilience. Thirty months is also a time of behavioral and emotional transition, as children are in the process of developing a sense of self and a relationship with the world around them, gaining understanding of the standards, rules, and goals of their society, and developing self-conscious, secondary emotions that are central to resilience, such as empathy, pride, and shame. Thirty-month-olds are also just beginning to use more complex forms of play (e.g., cooperative play, dramatic play). Lastly, the toddler years are understudied, with infancy researchers focusing on under-24 months of age, and preschool researchers beginning at 36 months of age.
Twin Panel Recruitment
The families with twins were recruited from birth records through collaboration with the Arizona State Department of Health Services. Because Arizona is a closed-records state, we were not able to contact families directly because we did not have access to identifying information until the families provided it to us. Instead, the Division of Public Health Services, Office of Vital Records mailed letters (in English and Spanish) to a random sample of mothers over the age of 18 years who had given birth to live twins in an Arizona hospital between July 2007 and July 2008. Recruitment letters were mailed two months prior to the twins’ 12-month birthday, including a postage-paid return letter where families could indicate their interest in participating. Follow-up letters were mailed one month after the first letter if we did not receive a response. Informed consent was obtained prior to all interviews, and participants were compensated a total of US$40 at each measurement occasion, as well as an additional US$10 for completing follow-up demographic and zygosity questionnaires.
As part of the initial assessment, we recruited and assessed a sample of 582 twins. As twins were turning 30 months of age, we re-contacted parents through telephone, e-mail, and mail to request continued participation in the study. Between assessments, we maintained contact with families and promoted interest by mailing birthday cards to twins and project newsletters three to four times a year to parents.
Five hundred and eighty-two twins (26% monozygotic (MZ) twins, 36% same-sex dizygotic (DZ), 38% opposite-sex DZ, making up 291 pairs) and their mothers participated when the twins were 12 months of age, and 520 (89%) also participated in the longitudinal follow-up assessment at 30 months. Of the 62 twins (31 families) who did not participate at 30 months, only four families declined to participate, while others did not participate because they were too busy or were traveling during the target window. Twin participants are 25% Hispanic, 66% Caucasian, 5% Asian American, and 4% African American. Income ranged from less than US$20k to over US$100k, with a median of US$60k–80k. Parental education ranged from less than a high school degree to a professional degree, with mean education of a college degree.
The Zygosity Questionnaire for Young Twins (Goldsmith, Reference Goldsmith1991) was used to determine whether twins were identical or fraternal. This questionnaire yields over 95% agreement with zygosity determined via genotyping (Forget-Dubois et al., Reference Forget-Dubois, Perusse, Turecki, Girard, Billette, Rouleau and Tremblay2003; Price et al., Reference Price, Freeman, Craig, Petrill, Ebersole and Plomin2000). Zygosity was further verified with infant birth medical records.
History of Data Collection: Prenatal, 12-Month, and 30-Month Assessments
The pregnancy and birth records assessment consisted of three scales of obstetrical complications and neonatal risk factors, as well as maternal report of perceived prenatal stress. The Obstetrical Complications Scale (OCS; adapted from Littman & Parmalee, Reference Littman and Parmalee1974) was used to measure a wide range of specific prenatal and obstetrical risks (e.g., history of stillbirth, prenatal substance use), labor and delivery information, and other maternal characteristics (e.g., maternal parity). We obtained information on chorionicity and twin-to-twin transfusion syndrome when possible. The Neonatal Complications Scale (NCS; adapted from Littman & Parmalee, Reference Littman and Parmalee1974) was used to examine specific measures of infant health (e.g., birth weight, Apgar scores) and neonatal risks and complications (e.g., admission to the NICU), and The Neonatal Morbidity Scale (NMS; Minde et al., Reference Minde, Whitelaw, Brown and Fitzhardinge1983) evaluated the presence and severity of health complications, including bradycardia, tachypnea, and whether twins were prohibited from feeding. These scales were completed using medical records collected from birth hospitals, and maternal perceived prenatal stress was assessed during the 12-month interview.
The 12-month assessment consisted of two, hour-long telephone or online interviews with the twins’ primary caregiver. One doctoral student oversaw a team of five to eight undergraduate interviewers who were extensively trained in interviewing methodology. Bilingual research assistants also translated and back-translated the interviews into Spanish, and conducted interviews with Spanish-speaking participants.
As part of the first 12-month interview, primary caregivers answered detailed demographic and zygosity questionnaires, including questions assessing not only demographic characteristics, such as income, education, and racial or ethnic background, but also hours and days per week of in-home and out-of-home childcare, number of other adults in the home, and the country of origin of twins’ primary caregiver and grandparents. Primary caregivers also reported on their twins’ functioning in core developmental areas (e.g., motor, language, social), multiple dimensions of temperament, and twins’ developmental competencies (e.g., social approach, imitative play) and internalizing, externalizing, and attentional problem behaviors.
The second 12-month interview assessed twins’ regulation concerning sleep, eating, and sensory sensitivity, as well as several characteristics of the home and family environment, including chaos, emotional expressivity, and parenting-related stress and social and family support. In addition, multiple cognitive and behavioral aspects of parenting were assessed during the second interview, including emotional availability, parental warmth and control, overprotective and hostile behavioral tendencies, self-efficacy related to parenting, and bedtime routines. Finally, we assessed parental depression, and positive aspects of caregiver personality that serve as resilience factors: effortful control, empathy and perspective taking, self-compassion, optimism, personal mastery, and hope.
The 30-month follow-up assessment consisted of two, hour-long telephone or online interviews with the twins’ primary caregiver when the twins were 30 months of age (adjusted for gestational age). Interviews were scheduled two weeks in advance and took place approximately a week apart. In addition, for families whose twins had indeterminate or missing zygosity questionnaires at 12 months, we scheduled zygosity interviews of approximately 15–20 min in length. Measures of child development and home and family environment carried over from the 12-month to the 30-month assessment include the assessments of chaos, social and family support, emotional availability, parenting daily hassles, parental depression, as well as twins’ developmental maturity, problem behaviors, and competencies. New measures of twins’ development introduced at 30 months include assessing the twins’ general physical health, communication, and sense of self, positive affect, and additional dimensions of regulatory temperament. In addition, new measures of the home environment include parents’ use of punitive and authoritative discipline, quality of stimulation and support available to twins in the home, and the number of hours and minutes per day twins spend being read to or looking at books, watching television, and being exposed to background television.
Table 1 lists completed assessments at 12 and 30 months. Demographic variables, such as ethnicity and number of adults and children in the home, are assessed at each occasion. All measures have been used in previous research and have acceptable levels of reliability and validity with Caucasian samples. Whenever possible, we used measures that have also been used with Hispanic samples.
TABLE 1 Summary of Measures in the Arizona Twin Project
Preliminary Findings
We have begun to analyze data and below we highlight findings in three key domains: problem behaviors, health behaviors, and prenatal-birth environments.
Parenting and Infant Problem Behaviors
Positive parent personality may increase emotional availability (related to infant attachment security), which is protective for children's problem behaviors. We examined emotional availability as a mediator between maternal positive personality and infant problem behaviors (O'Brien et al., Reference O'Brien, Kao, Valiente and Lemery-Chalfant2011). Using multilevel modeling to account for twin dependence, the mediated effect was significant, with positive personality relating negatively to infant problem behaviors, but not when accounting for emotional availability. Positive personality related positively to emotional availability, and emotional availability related negatively to infant problem behaviors. Furthermore, infant problem behaviors were moderately heritable (h2 = 0.36), whereas maternal emotional availability was largely environmental (h2 = 0.10). Emotionally available parents communicate an atmosphere of warmth and acceptance toward their children (Biringen & Robinson, Reference Biringen and Robinson1991), a process through which parent positive personality may protect children from developing problem behaviors.
Heritability of Toddlers’ Behavior Problems Varies by Early Parental Warmth
A lack of parental warmth and harsh verbal and physical punishment have been associated with childhood problem behaviors (McKee et al., Reference McKee, Roland, Coffelt, Olson, Forehand, Massari and Zens2007). Using Purcell's (Reference Purcell2002) moderated ACE model, parental warmth in infancy significantly moderated the A and E paths such that heritability increased as parental warmth increased (Swann et al., Reference Swann, Araica, O'Brien, Kao, Valiente and Lemery-Chalfant2012). Thus, problem behaviors are more heritable under optimal warm parenting conditions, whereas the environment has a larger impact when caregivers are more rejecting of their young children.
Protecting Infant Sleep
An estimated 25–41% of children aged 1 to 5 years do not get enough sleep (Lam et al., Reference Lam, Hiscock and Wake2003), which is a major concern because early sleep difficulties are a strong predictor of later sleep problems, as well as behavioral and emotional problems (Gregory & O'Conner, Reference Gregory and O'Connor2002; Morrell & Steele, Reference Morrell and Steele2003; Wolke et al., Reference Wolke, Meyer, Ohrt and Riegel1995). Our goal was to examine three possible protective factors for infant sleep: parental optimism, parenting cognitions, and bedtime routines (Drake et al., Reference Drake, O'Brien, McDonald, Valiente and Lemery-Chalfant2010). First, infant sleep dysregulation was only modestly heritable (h2 = 0.22). Next, we used multilevel modeling to test process models: both parenting cognitions and bedtime behaviors mediated the relation between parent optimism and infant sleep dysregulation. Optimistic parents may have infants who sleep better because of their positive cognitions and bedtime routines.
Temperament Predicts Infant Feeding Difficulties
Difficulties with feeding are one of the most common problems that occur in pediatrics, and in addition to being concurrently associated with higher behavior problems (Budd et al., Reference Budd, McGraw, Farbisz, Murphy, Hawkins, Heilman and Werle1992), severe or untreated feeding problems may lead to malnourishment or micronutrient deficiency (Chatoor, Reference Chatoor2002), which are in turn risk factors for faltering growth, and cognitive problems (Branca & Ferrari, Reference Branca and Ferrari2002; Reif et al., Reference Reif, Beler, Villa and Spirer1995). Using multilevel modeling, we found that proneness to anger, rate of recovery from distress, and duration of orienting all predicted feeding difficulties (Jeon et al., Reference Jeon, O'Brien, McDonald, Valiente and Lemery-Chalfant2010). Furthermore, heritability of eating difficulties was non-significant, indicating that the associations with temperament were likely environmental in nature.
Heritability of Infant Outcomes Varies by Obstetrical and Birth Conditions
We examined whether or not prenatal and obstetrical environmental conditions moderated the heritability of infant outcomes (McDonald, Reference McDonald2011). Maternal perceived prenatal stress and obstetrical complications (coded from medical records) moderated the heritability of infant developmental maturity and competence, while gestational age moderated the heritability of infant developmental maturity, dysregulation, and competence. Thus, obstetrical and birth conditions may be important nonlinear influences on infant outcomes.
These preliminary findings suggest the importance of early environmental conditions for infant health, development, and adjustment, both directly and as moderators of the expression of heritable traits. However, many models central to the Arizona Twin Project's primary aims have yet to be tested with the existing data, that is, components of resilience as moderators of the relations between environmental risks and negative outcomes. In addition, we have yet to consider self-regulation as a potential mediator or moderator of early environmental conditions and children's competency, or as a predictor of resilience under conditions of environmental risk. Testing these models, and examining predictors of recovery from health or behavioral problems between 12 and 30 months, both represent immediate next steps for the Arizona Twin Project.
Future Directions
Currently we are seeking funding for a third longitudinal assessment of the sample focused on the social environments, behaviors, and physiological processes that influence sleep and other health behaviors. Because of the widespread prevalence of sleep problems (25–40% for all ages, with 6% qualifying for a diagnosis of insomnia; Barclay et al., Reference Barclay, Eley, Rijsdijk and Gregory2011; Ferber, Reference Ferber, Ferber and Kryger1995; Sadeh et al., Reference Sadeh, Raviv and Gruber2000), and early sleep problems strongly predicting later sleep problems and disorders (Gregory & O'Connor, Reference Gregory and O'Connor2002), it is important to explicate their etiology and identify targets of preventive interventions. The new assessment includes day-to-day objective measures of sleep, artificial light exposure, physical activity, and diurnal cortisol, as well as diary assessment of mood, diet, and health behaviors to elucidate genetic and environmental mediation of dynamic bidirectional processes that characterize relations between the social environment and sleep. By focusing on the early social environment within a longitudinal twin study design, the proposed study elucidates social processes by which children develop healthy sleep habits that serve to promote physical and mental health.
We plan to add additional participants to the Arizona Twin Project by recruiting additional twin pairs through birth records from the same birth cohorts. There is also the opportunity to add additional birth cohorts to the sample as opportunities arise.
Conclusion
The Arizona Twin Project is a new twin panel that will likely increase in size and research focus. Twins were recruited through birth records, with significant representation of Mexican American, as well as European American participants. Multiple aspects of the prenatal, birth, infancy, and toddler environments have been assessed, creating an opportunity to study both genetic and environmental contributions to the development of resilience, and physical and mental health. In addition, our extensive phenotyping is ideal for addressing gene–environment interplay, such as gene–environment correlation, and moderation of heritability.
Acknowledgments
This research was supported by grants from the Institute for Mental Health Research and Arizona State University's Challenged Child Project and T. Denny Sanford School of Social and Family Dynamics.
