Externalizing Behaviors in Preschool-Aged Children With Cochlear Implants

Abstract

Purpose:

Deaf children with cochlear implants (CIs) experience challenges in early development of hearing and language skills that may place them at risk for externalizing behavior problems, such as aggression, hyperactivity–impulsivity, and oppositional behavior. This study is a longitudinal investigation of (a) between-groups differences in externalizing behavior problems between preschool-aged children with CIs and normal-hearing (NH) peers, and (b) within-group factors that may explain variability in externalizing problems within the sample of CI users.

Method:

Parents of 26 children with CIs and 30 NH peers completed externalizing behavior checklists at two visits separated by 1 year, starting at ages 3 or 4 years. Demographic/hearing history variables, language (vocabulary), nonverbal intelligence, and coping flexibility were assessed for concurrent and predictive associations with externalizing problems within the CI sample.

Results:

Results showed significantly greater externalizing behavior problems in CI users compared to NH peers at Time 1, although these differences were less pronounced 1 year later. Poorer residual hearing and better coping flexibility at Time 1 were associated with fewer externalizing behavior problems in CI users at Time 2. CI users who showed improvement in coping flexibility over the 1-year period also showed improvement in externalizing behaviors during that period. Nonverbal intelligence and language were not associated with externalizing behavior problems.

Conclusions:

Preschool-aged CI users may be at greater risk than NH peers for the early development of externalizing behavior problems. Improved coping flexibility may offer the potential for improvement in externalizing behavior problems for young CI users.

Deaf children who receive cochlear implants (CIs) experience an early period of auditory and spoken language deprivation, followed by improvement in auditory, speech recognition, and spoken language functioning after implantation. However, even after implantation, the auditory signals provided by the CI are fragile and underspecified, and some components of speech recognition (e.g., speech in noise) continue to be challenging for children with CIs. The complicated history of sensory and linguistic experiences for children with CIs, combined with ongoing challenges of language development and the underspecified signals provided by the device, may have downstream effects on other domains of everyday functioning that are influenced by auditory and language skills.

Psychosocial functioning is an important functional domain that may be influenced by these downstream effects (Wong et al., 2017). Psychosocial functioning encompasses psychological adjustment (behavioral–emotional adaptation and well-being) and social adaptation (forming, maintaining, and using relationships with other people; Wong et al., 2017). Psychological adjustment falls into two higher order symptom domains: internalizing symptoms are characterized by internal distress, such as anxiety, depression, and withdrawal, whereas externalizing symptoms are characterized by outwardly directed disruptive behaviors, such as aggression, hyperactivity, impulsivity, and intrusive behaviors (Achenbach & Edelbrock, 1978).

Psychological adjustment symptom domains can be distinguished from social adaptation (Achenbach & Rescorla, 2001; Verona et al., 2011). Social adaptation consists of relationships with family, friends, and a partner. Social adaptation also includes social skills, behaviors, and achievements such as leadership, empathy, perspective taking, and occupational–economic status (Achenbach & Rescorla, 2001, 2003; Reynolds & Kamphaus, 2015).

The two elements of psychosocial functioning—psychological adjustment and social adaptation—are core components of quality of life. As defined by the World Health Organization, quality of life encompasses psychological health, quality of social relationships, physical health, functional independence, spirituality/personal beliefs, and environmental quality (e.g., safety, resources; World Health Organization, 2012). Health-related quality of life refers to quality of life in response to health conditions or treatments and emphasizes psychological, social, and physical functioning (Kaplan & Hays, 2022; McRackan et al., 2019). Thus, achieving optimal psychosocial functioning is crucial for attaining quality-of-life goals following cochlear implantation.

Deficits or delays in language skills increase the risk for problems in psychological adjustment and social adaptation. Language is used to form appraisals of stress and distress, which underlie psychological symptoms such as depression and anxiety (Beck, 2021; Lazarus & Folkman, 1984). Language is also a primary tool for coping behaviors such as self-talk, information seeking, and seeking social support, which act to mitigate the adverse effects of stress on psychological adjustment (Lazarus & Folkman, 1984). Internalized language provides directions for self-control (Alderson-Day & Fernyhough, 2015; Winsler et al., 2009) and representation of internal plans to guide future behaviors (Alderson-Day & Fernyhough, 2015), which allow for emotional control in challenging situations. Language also serves as a core component of social interaction and communication (Leigh et al., 2015), and the quality of language development is a strong influence on social functioning in young children (Huttunen & Välimaa, 2010). Thus, because language is used extensively to manage psychological adjustment and social adaptation, children with language delays or differences are at greater risk for the development of problems in psychosocial functioning (Barker et al., 2009; Bat-Chava et al., 2005; Castellanos et al., 2020; Huttunen & Välimaa, 2010).

The potential downstream effects of language delays on psychosocial functioning have led to a growing body of research investigating psychosocial outcomes in children with CIs (Bigler et al., 2019; Castellanos et al., 2020; Warner-Czyz et al., 2022). Studies in this area have addressed two significant research topics (Kronenberger et al., 2024): Between-groups studies have investigated whether psychosocial outcomes differ in children with CIs compared to normal-hearing (NH) peers (Hoffman et al., 2016); thus, these studies attempt to understand and explain variability that may exist between groups of children with CIs and those with NH. Within-group studies (Leigh et al., 2015) investigate factors that are associated with variability in psychosocial outcomes within the population of children with CIs; these studies attempt to explain the large amount of variability in psychosocial outcomes that exists within the group of CI users.

For prelingually deaf children implanted at young ages, the earliest accomplishments and challenges in psychosocial development occur during the preschool-age period when children are approximately 3–6 years old. At these ages, children encounter the first expectations and demands to use self-directed strategies to tolerate frustration, follow rules, and control behavior (Studts et al., 2022). Leigh et al. (2015) investigated 3-year-old deaf and hard of hearing (DHH) children with hearing aids (HAs) or CIs as a part of the large population–based Longitudinal Outcomes of Children With Hearing Impairment (LOCHI) study (Ching et al., 2013). In the DHH sample, scores on a standardized measure of social functioning fell at the low end of the average range compared to norms. Better language and the absence of additional disabilities predicted better social adjustment. In a sample of 5-year-old DHH children obtained longitudinally from the LOCHI study, Wong et al. (2017) found that mean scores on a global measure of behavioral and emotional problems fell within 1 SD of the normative mean. However, mean sample scores on a measure of social functioning were more than 1 SD below the normative mean, suggesting greater challenges in social than in behavioral–emotional functioning. More behavioral problems were found in children with severe-to-profound hearing loss who used HAs compared to those who used CIs. Among those using CIs, additional disabilities, poorer language, and poorer functional auditory outcomes were associated with poorer psychosocial outcomes.

Studies of quality of life in preschool-aged children with CIs have also investigated components of psychosocial functioning. Warner-Czyz et al. (2009) found that preschool-aged children with CIs rated their overall quality of life (consisting of physical, social, and psychological–emotional functioning) similarly to their NH peers. Children's ratings of their quality of life were higher than ratings provided by parents. In a follow-up study incorporating preschool and older children with CIs, younger children rated their quality of life more positively than older children (Warner-Czyz et al., 2011). A qualitative study of 3- to 8-year-old children produced similar results: Preschool-aged children with CIs reported increased self-confidence, and their communication skills after implantation improved in a way that benefitted psychosocial development (Huttunen & Välimaa, 2010). These quality-of-life studies suggest that preschool-aged children with CIs rate their own psychosocial functioning similar to or better than that of NH peers.

Two longitudinal studies investigated psychosocial functioning in preschool-aged children with CIs by focusing specifically on parent-rated behavior problems related to executive functioning delays, such as attention problems and impulsivity. Kronenberger et al. (2020) found greater impulsivity and more attention problems in preschool-aged children with CIs compared to NH peers, but children with CIs showed large and consistent improvement in these problem areas between ages 3 and 8 years. Better language predicted fewer attention problems longitudinally. Jamsek et al. (2022) identified significant differences in executive functioning behavior problems between preschool-aged (3–8 years of age) DHH children and NH peers at baseline measurement. However, the DHH group showed significant improvement in executive functioning over a 1-year period and did not differ from NH peers 1 year later. Similar to the findings of Kronenberger et al., better language skills predicted fewer attention problems 1 year later.

Between-groups studies combining preschool-aged and older children and adolescents have yielded conflicting results when comparing DHH and NH samples. Some studies find no differences in psychosocial functioning between DHH and NH children or find stronger psychosocial functioning in DHH children (Ketelaar et al., 2013; Sarant et al., 2018), while others identify greater psychosocial problems in the DHH group (Van Eldik et al., 2004). Within-group studies of language communication skills in DHH samples (e.g., language comprehension, expressive language, and speech production, as compared to office-based vocabulary or language knowledge tests) have been consistently associated with psychosocial outcomes in children with CIs (Freeman, 2018; Hintermair, 2006; Leigh et al., 2015; Wong et al., 2017). Milder hearing loss (Patrick et al., 2018) and bilateral CIs (Sarant et al., 2018) have also been associated with fewer psychosocial problems.

In summary, between-groups studies of psychosocial outcomes in preschool-aged children with CIs have produced varied results, with some finding no group differences and others finding poorer psychosocial functioning in the DHH sample. This inconsistency in findings may be due to differences in the methodology and domain of psychosocial functioning that has been studied. Self-reports of broad psychosocial or quality-of-life functioning often find no differences between samples of CI and NH preschool-aged children (Warner-Czyz et al., 2009, 2011), whereas ratings of social functioning provided by parents more often find differences (Leigh et al., 2015; Wong et al., 2017). Additionally, psychosocial behavior problems associated with executive functioning delays such as inattention and hyperactivity–impulsivity are typically more prevalent in CI users than NH peers during preschool ages (Jamsek et al., 2022; Kronenberger et al., 2020).

In addition to between-groups findings, significant variability in psychosocial outcomes has been consistently reported within DHH samples, indicating that within-group factors (e.g., explaining within-group variability) are at least as important as between-groups factors in understanding the process of psychosocial adjustment after implantation. Within-group investigations of preschool-aged DHH children suggest that better language skills, better functional auditory outcomes, and the absence of additional disabilities are usually associated with more positive psychosocial outcomes.

The extant literature investigating psychosocial outcomes in preschool-aged CI users has been limited by some methodological factors. Some studies combine CI users with other DHH children or combine preschool-aged children with older children. Other studies are based only on cross-sectional data and are not able to evaluate the progression of psychosocial functioning or factors that predict changes in psychosocial functioning over time. A wide variety of measures have been used, assessing diverse subdomains of psychosocial functioning: Attention problems, social adjustment, hyperactivity–impulsivity, and global psychological functioning have been investigated in different studies. Thus, while existing research provides valuable insights into psychosocial adjustment during preschool ages after implantation, there is a pressing need for longitudinal investigations focusing on specific, well-defined psychosocial domains that are particularly relevant for preschool-aged children.

Externalizing behaviors are among the most common psychological adjustment problems shown by preschool-aged children (American Psychiatric Association, 2013; Bitsko et al., 2022) and include aggression, hyperactivity–impulsivity, oppositional-defiant behavior, and severe rule-breaking/conduct problems (Achenbach & Edelbrock, 1978; Reynolds & Kamphaus, 2015; Thompson et al., 1992; Witkiewitz et al., 2013). The presence of significant externalizing behavior problems during early childhood is immensely stressful to parents and is associated with long-term risks including academic underachievement, social rejection, and trouble with authorities (Barkley, 2014; Bierman et al., 2013; Coie et al., 1991; Dodge et al., 2015). Despite their prevalence and adverse effects on child and family well-being, externalizing behaviors respond well to behavioral treatments, and well-established evidence-based treatments exist for externalizing behaviors (Barkley, 1997; Studts et al., 2022).

Because of age and limited experience with spoken language, preschool-aged children with CIs may be at risk for problems with externalizing behaviors (Kronenberger et al., 2014; Studts et al., 2022). Greater listening effort and consequent fatigue from CI use may also deplete limited cognitive resources used for coping and behavioral control. Delays in language development and in the use of language for social–emotional reasoning are associated with the development of externalizing behavior problems at preschool and early school ages (Barker et al., 2009; Dodge et al., 2015). As a result, it has been hypothesized that children with CIs may be at elevated risk for externalizing behavior problems (Studts et al., 2022). Some studies have supported this hypothesis in older children with CIs, especially when language delays are present (Barker et al., 2009; Castellanos et al., 2018a). However, no study to date has investigated externalizing behavior problems in a sample of preschool-aged CI users using longitudinal methods and measures specifically designed for the assessment of externalizing behaviors.

There is a significant need to understand between-groups differences in externalizing behaviors, in order to establish if and when CI users are at elevated risk for externalizing behaviors compared to NH peers. Externalizing behaviors often emerge at preschool ages, and interventions to address these behaviors may be more effective at younger ages, when parents can be involved in the implementation of behavior strategies and when behaviors are less well established (Studts et al., 2022). Preschool-aged children with CIs experience significant demands for language development and adaptation to the CI, with language showing rapid improvement during preschool ages (Kronenberger et al., 2020; Niparko et al., 2010).

A need also exists to understand within-group factors contributing to variability in externalizing behaviors within the population of preschool-aged CI users. Although prior work suggests that better language skills are associated with positive psychosocial outcomes within older samples of children with CIs, there has been little research investigating language influences or other potential within-group factors on externalizing behaviors during the preschool period of development. Research with samples of young children experiencing stresses from physical and developmental conditions suggests that one of the largest influences on psychosocial outcomes, including externalizing behaviors, is coping behavior (Kronenberger et al., 1996; Spirito et al., 1988; Thompson & Gustafson, 1996). Coping skills and behaviors may influence externalizing behavioral outcomes after implantation because they are used to mitigate the frustrations and stresses of adjusting to the challenges of hearing loss, cochlear implantation, and language delays (Castellanos et al., 2020). However, no research to date has investigated the within-group effects of coping behaviors on psychological adjustment in children with CIs.

A robust literature in NH children, adolescents, and adults across the life span indicates that coping behaviors in response to stressors have a significant effect on short- and long-term psychological adjustment (Lazarus & Folkman, 1984), including externalizing behavior problems (Kronenberger et al., 1996). Coping behaviors, defined as cognitive or behavioral responses to mitigate the potential adverse effects of stresses (Lazarus & Folkman, 1984; Thompson & Gustafson, 1996), are particularly important for adjustment in response to the stressors of physical conditions (Thompson & Gustafson, 1996). When individuals deploy coping strategies that effectively target stresses by solving problems or improving emotional reactions, the adverse impact of stresses on psychosocial functioning is minimized, producing positive outcomes (Lazarus & Folkman, 1984; Thompson & Gustafson, 1996). For young children with CIs, the use of coping skills to regulate emotion and flexibly adjust behavior (“coping flexibility”) in the face of stresses may be particularly important for psychosocial adjustment, given the risk of delays in executive functioning and self-regulation in this population (Castellanos et al., 2018a, 2020; Kronenberger et al., 2020).

In this study, we investigated between-groups and within-group variability in externalizing behavior problems in preschool-aged children with CIs compared to NH peers at two study visits 1 year apart. For between-groups comparisons, we hypothesized that CI users would experience greater levels of externalizing behavior problems than NH peers at both time points because language delays are particularly pronounced in CI users during preschool ages (Niparko et al., 2010). Within the CI sample, we expected that better language and coping flexibility would be associated with fewer externalizing behavior problems concurrently and prospectively. We assessed language using a test of vocabulary because of the young age of study participants (and consequent need for an engaging and brief measure of language knowledge). Finally, we expected that improvements in language skills and coping flexibility during the 1-year period would be associated with decreases in externalizing behaviors over that same time period, demonstrating a dynamic, predictive association between language, coping, and externalizing behavioral problems in preschool-aged CI users.

Method

Participants

Participants for the present study were 56 children who were 3 or 4 years of age at the time of their first study visit (see Table 1): 26 children who were deaf at birth and received CIs prior to age 37 months and 30 children who had NH. Inclusion criteria for both samples included home environment in which English was the primary language spoken; absence of developmental, cognitive, or neurological diagnoses; and nonverbal intellectual ability (measured by the Differential Ability Scales–Second Edition [DAS-II; Elliott, 2007] Picture Similarities subtest) greater than 2 SD below the normative mean. Additional inclusion criteria for the sample of children with CIs were (a) cochlear implantation at age 36 months or younger, (b) use of CIs for at least 6 months, and (c) enrollment in education or rehabilitation interventions focused on the development of listening and spoken language skills. Additional inclusion criteria for the sample of NH peers were (a) NH and language by parent report and (b) hearing screening within normal range, based on each ear individually tested at frequencies of 500, 1000, and 2000 Hz in dB HL using Telephonics TDH-50P headphones in an Acoustic Systems RE-243 soundbooth.

Table 1.

Sample description.

Variable	Cochlear implant (N = 26)		Normal hearing (N = 30)		t
	M (SD)	Range	M (SD)	Range
Chronological agea	3.74 (0.51)	3.07–4.98	3.90 (0.55)	3.12–4.99	1.12
Income levelb	6.80 (3.19)	1–10	7.57 (2.11)	1–10	1.07
Age of onset of deafnessc	0 (0)	0–0	NA	NA
Duration of deafnessc	18.28 (7.70)	7.98–36.60	NA	NA
Best ear preimplant PTAd	99.80 (12.12)	75–118.43	NA	NA
Age at implantationc	18.28 (7.70)	7.98–36.60	NA	NA
Duration of CI usea	2.21 (0.85)	0.53–4.31	NA	NA
DAS-II Picture Similaritiesf	53.88 (8.86)	42–81	61.53 (13.08)	41–90	2.45*
PPVT-4g	83.72 (18.62)	45–117	120.03 (10.10)	98–139	9.19***
BRIEF-P Shiftf	49.46 (9.38)	40–74	48.80 (8.68)	38–67	0.27
Time between visitsa	1.01 (0.23)	0.50–1.44	0.98 (0.25)	0.61–1.81	0.38
Visit 1 age (3 years/4 years)	20/6		19/11
Sex, n (female/male)	11/15		15/15
Race/ethnicity
Asian	0		0
Black/African American	5		1
White	23		30
Hispanic/Latino	1		1
Communication modee
Total communication	2		NA
Oral communication	24		NA

Note. Values are at Visit 1. N = 26 for CI sample and 30 for NH sample, except N = 25 for CI sample Income and PPVT-4 values, and N = 24 for CI sample DAS-II values. NA = not applicable; PTA = pure-tone average; CI = cochlear implant; DAS-II = Differential Abilities Scales–Second Edition; PPVT-4 = Peabody Picture Vocabulary Test–Fourth Edition; BRIEF-P = Behavior Rating Inventory of Executive Function–Preschool Version; Time between visits = time (years) between Visit 1 and Visit 2. p value for Visit 1 age and sex obtained from Fisher's exact tests, two-sided = .39 and .60, respectively.

^aIn years.

^bOn a scale of 1 (under $5,500/year) to 10 ($95,000+/year).

^cIn months.

^dPTA for frequencies 500, 1000, and 2000 Hz in dB HL.

^eCommunication mode used in educational setting (from Geers & Brenner, 2003).

^fT scores.

^gStandard scores.

^*p < .05.

^***p < .001.

Procedure

Written consent from parents was obtained prior to initiation of study procedures, which were approved by the local institutional review board (Study No. 1011003908). Children completed a battery of speech, language, and neurocognitive abilities, while parents completed questionnaires regarding psychosocial adjustment, adaptive behavior, learning, and executive functioning. At Visit 1, 51 mothers, four fathers, and one grandmother (who was the mother figure and primary caregiver) completed questionnaires. Questionnaires were completed by the same parent respondents at Visit 2, with the exception of two fathers who completed questionnaires for participants whose questionnaires had been completed by mothers at Visit 1. Demographic, medical, and hearing history information was obtained from interviews with parents and review of medical records. Tests were administered by American Speech-Language-Hearing Association–certified speech-language pathologists using standard instructions. Verbal directions, visual demonstrations, and practice were used to ensure that children fully understood the tests prior to administration of scored items.

Study participants were evaluated annually as a part of a 5-year longitudinal project (Kronenberger et al., 2020); the first two study visits were used in the current investigation. For the current study, evaluations were conducted at each year of age starting at age 3 or 4 years, with no fewer than 6 months and no longer than 23 months between visits (target interval of 12 months). Results from externalizing behavior questionnaires completed by caregivers during the first 2 years of the study are reported here, in order to investigate externalizing behaviors in preschool-aged children.

Measures

Demographics and Hearing History

Chronological age at each visit, sex, time between evaluations, and income level were assessed for all participants. Income level was rated on a 1–10 scale, with anchors ranging from $5,500/year to > $95,000/year (one family of a CI user opted not to answer this item). For participants in the CI sample, age at implantation, duration of CI use, communication mode in educational/intervention settings, and best-ear preimplant pure-tone average (PTA; available for N = 23 CI users) for 500, 1000, and 2000 Hz in dB HL were obtained from chart review and parent interview. Communication mode was coded as total communication or oral communication based on descriptions provided by Geers and Brenner (2003); 92% of participants were in oral communication settings. Because all participants were deaf at birth, duration of deafness prior to implantation was equal to age at implantation. Because of insufficient variability or redundancy with other variables, communication mode, age of onset of deafness, and duration of deafness prior to implantation were dropped from further analysis.

Externalizing Behavior Measures

Two parent-completed behavior checklists were used to assess externalizing behaviors: The Conduct–Hyperactive–Attention Problem–Oppositional Symptom (CHAOS) scale (Levy et al., 2017) is a 22-item measure of disruptive behavior disorder symptoms based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnostic criteria for the disruptive behavior disorders: attention-deficit/hyperactivity disorder (ADHD) predominantly inattentive type, ADHD predominantly hyperactive–impulsive type, oppositional-defiant disorder, and conduct disorder. A panel of experts identified the five most prototypical DSM-IV symptoms for each of those four diagnoses, resulting in a 20-item parent-completed disruptive behavior questionnaire with four subscales (Attention Problems, Hyperactivity–Impulsivity, Oppositional Behavior, and Conduct Problems) and a total score (Levy et al., 2017). No CHAOS scale items involve audition.

CHAOS subscale scores are supported by factor analysis and have excellent internal consistency (median alpha value across the four subscales of 0.88), parent–parent interrater reliability (median interrater correlation of .58), and test–retest reliability (median test–retest correlation of .78; Levy et al., 2017). CHAOS scores correlate strongly (r = .80–.90) with behavior ratings based on corresponding DSM-IV symptoms, demonstrating scale validity (Levy et al., 2017). The CHAOS has been used to measure psychosocial functioning in school-aged children and adolescents (age 6–21 years) with CIs in prior work (Castellanos et al., 2018a). CHAOS scores are the sum of item scores, which are rated on a 0–3 scale of frequency and intensity based on behavior during the past month. Based on a large referred clinical sample of 205 children, the average parent-reported total CHAOS raw score was 23.86 for children with ADHD diagnoses alone and 24.26 for children with oppositional-defiant disorder diagnoses alone (Levy et al., 2017). Thus, a total CHAOS raw score of 24 is typical of children with clinically significant externalizing problems and can be used as a clinical cutoff score. In the present sample, the internal consistency values of the four CHAOS subscales and total scores were alpha = .88, .83, .85, .90, and .93, respectively.

The Outburst Monitoring Scale (OMS) is a 20-item measure of frequency and intensity of specific types of aggressive outbursts in children and adolescents (Kronenberger et al., 2007). The first 16 items of the OMS correspond directly to the aggressive behavior categories and levels of the Overt Aggression Scale (Yudofsky et al., 1986) and Modified Overt Aggression Scale (Donovan et al., 2000). No OMS scale items involve audition. Four types of aggressive outbursts (verbal aggression, aggression toward property, aggression toward self, and physical aggression) are each assessed by four items representing differing levels of severity, ranging from mild to severe outbursts. Each item is rated on a scale from 0 (not at all) to 4 (three or more times in a typical day) based on behavior during a typical week in the past year. Subscale scores for Verbal, Property, Self, and Physical Aggression, as well as a total score across all 16 items, are derived by summing item scores. The OMS has demonstrated excellent reliability and validity, including support as a sensitive measure of change in outbursts over time (Kronenberger et al., 2007). OMS scores indicate not only the severity of aggressive outbursts in general but also severity for different types of outbursts. In a nonclinical sample, mean OMS total score was found to be 0.8, with SD = 1.2; therefore, OMS scores of 4 or greater (more than 2 SD above the mean) may be considered to be abnormally elevated (Kronenberger et al., 2007). In the present sample, OMS subscale and total scores demonstrated excellent internal consistency, with alpha for subscale and total scores of .86, .93, .79, .86, and .93, respectively.

Nonverbal Intelligence

Norm-based T scores from the Picture Similarities subtest of the DAS-II (Elliott, 2007) were used to assess nonverbal/fluid intellectual ability. The DAS-II Picture Similarities subtest asks children to match a picture on a card with a conceptually similar picture from a row of four pictures (e.g., the picture of a table would be selected to match with a picture of a chair, since both are furniture), without requiring verbal labeling or verbal identification of the concept. Psychometric analyses with children as young as 2.5 years of age support the reliability and validity of the Picture Similarities subtest and indicate that Picture Similarities scores load on a nonverbal/fluid reasoning factor that is distinct from DAS-II verbal subtests (Elliott, 2007).

Vocabulary

The Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4) is a measure of one-word receptive vocabulary for ages 2.5–90 years (Dunn & Dunn, 2007). From a set of four pictures, children select the one picture that matches a vocabulary word spoken by the examiner. The PPVT-4 is one of the most widely used and validated language measures in children and adolescents and provides standard scores based on a national norm sample.

Coping Flexibility

The Shift subscale of the Behavior Rating Inventory of Executive Function–Preschool Version (BRIEF-P; Gioia et al., 2003) was used to measure coping flexibility under stress. Although the BRIEF-P is intended as measure of executive functioning, the Shift subscale of the BRIEF-P assesses adaptive cognitive and behavioral flexibility when confronted by emotionally challenging settings, which overlaps with the concept of coping with stress (Lazarus & Folkman, 1984). The content of Shift items involves control and adjustment of behavioral and emotional responses in new, challenging, unexpected, unfamiliar, or stimulating activities and settings; higher scores indicate more difficulties with flexible coping in those settings (Gioia et al., 2003). Thus, BRIEF-P Shift subscale T scores were used in this study as a measure of coping flexibility under stress.

Statistical Analysis

Total scores on the CHAOS and OMS were used as the primary measures of externalizing behavior problems in this study because of their strong internal consistency and breadth of coverage of externalizing behaviors. Analyses using CHAOS and OMS subscale scores were conducted to better understand in detail the results of total scores and are considered to be exploratory. The exploratory analyses are provided for additional information only and are not part of the primary study analyses. Visit 1 and Visit 2 scores for each measure were analyzed separately, and a change score was calculated by subtracting the Visit 2 score from the Visit 1 score.

In order to address between-groups research questions (comparison of externalizing behaviors of preschool-aged children with CIs and NH children over a 1-year period), we conducted 2 (Hearing Group – CI/NH) × 2 (Time – Visit 1/Visit 2) repeated-measures analyses of variance (ANOVAs) using CHAOS and OMS total scores as dependent variables, with a two-tailed p value of .05 to evaluate statistically significant differences. Statistically significant findings were then further investigated with t tests comparing the CI and NH samples on CHAOS and OMS total and subscale scores at Visit 1 and Visit 2. Correlations were calculated between Visit 1 and Visit 2 total scores of the CHAOS and OMS in order to assess stability of parent-rated externalizing behaviors over a 1-year period. Similar correlations for CHAOS and OMS subscales are reported to provide additional information about stability/change of externalizing behaviors over time.

In order to address within-group research questions (associations of demographic, hearing history, neurocognitive, language, and coping factors with externalizing behaviors over a 1-year period in preschool-aged children with CIs), correlations were calculated in the CI sample between Visit 1 demographic/hearing history variables (age, family income, best preimplant PTA, age at implantation, and duration of implant use), nonverbal intellectual ability (DAS-II Picture Similarities score), vocabulary (PPVT-4 score), and coping flexibility (BRIEF-P Shift score) scores and Visits 1 and 2 externalizing behavior scores (CHAOS and OMS total scores). These correlations investigated the extent to which Visit 1 within-group factors—especially language and coping flexibility—were associated concurrently (correlations with Visit 1 CHAOS and OMS scores) and predictively (correlations with Visit 2 CHAOS and OMS scores) with externalizing behaviors in young children with CIs. Furthermore, correlations between Visit 1–Visit 2 change scores on measures of externalizing behaviors (CHAOS and OMS total scores) with Visit 1–Visit 2 change scores on measures of nonverbal intellectual ability (DAS-II Picture Similarities score), vocabulary (PPVT-4 score), and coping flexibility (BRIEF-P Shift score) were used to investigate predictors of change in externalizing outcomes over a 1-year period.

Results

Sample Description

Children with CIs did not differ from NH peers in chronological age, family income, or time between study visits (see Table 1). Most participants in both samples were 3 years of age at study entry, and time between study visits averaged 1 year. The NH sample scored significantly higher (reflecting stronger abilities) than the CI sample on nonverbal intelligence (DAS-II Picture Similarities) and vocabulary (PPVT-4), although scores for the CI sample were slightly above the DAS-II Picture Similarities normative mean T score of 50 (see Table 1). The samples did not differ in coping flexibility (BRIEF-P Shift) scores. The CI sample was implanted at 18 months of age on average and had used their CIs for slightly more than 2 years (see Table 1).

Externalizing Behavior Scores in Children With CIs or NH

At Visit 1, the mean CHAOS total raw scores in the CI and NH samples (see Table 2) fell below the clinical cutoff value of 24 for the CHAOS scale obtained from clinically referred, diagnosed samples with ADHD or oppositional-defiant disorder (Levy et al., 2017). In the CI sample, 11 participants (42.3%) scored above the CHAOS clinical cutoff value of 24, compared to four participants (13.3%) in the NH sample at Visit 1 (Fisher's exact p value = .018). Mean CHAOS total raw scores in the CI and NH samples for Visit 2 also fell below CHAOS total score clinical cutoff values. Percentages of the CI and NH samples falling above CHAOS clinical cutoff values were 30.8% (N = 8) and 10% (N = 3), respectively (Fisher's exact p value = .09).

Table 2.

Conduct–Hyperactive–Attention Problem–Oppositional Symptom (CHAOS) and Outburst Monitoring Scale (OMS) externalizing scores at Visits 1 and 2.

Measure	Cochlear implant (CI)		Normal hearing (NH)		CI vs. NH
					Visit 1	Visit 2
	Visit 1	Visit 2	Visit 1	Visit 2	t	t
CHAOS (N)	26	26	30	30
Total	21.86 (13.37)	18.08 (11.96)	14.38 (8.08)	12.37 (9.05)	2.57*	2.03*
Attention	6.92 (3.74)	5.42 (3.98)	3.93 (2.91)	2.80 (2.59)	3.36***	2.96**
Hyperactive	7.86 (4.72)	7.08 (4.45)	6.50 (3.41)	5.97 (4.03)	1.24	0.98
Oppositional	5.15 (3.94)	4.46 (3.54)	3.48 (2.58)	3.13 (3.22)	1.90a	1.47
Conduct	1.92 (3.47)	1.12 (3.04)	0.47 (0.86)	0.47 (1.07)	2.22*	1.10
OMS (N)	26	26	29	30
Total	6.54 (10.40)	5.22 (9.58)	1.41 (1.47)	2.07 (2.48)	2.63*	1.74a
Verbal	2.35 (3.74)	2.03 (3.88)	0.52 (0.83)	1.03 (1.40)	2.57*	1.31
Property	2.27 (3.95)	2.00 (3.67)	0.72 (0.96)	0.73 (0.94)	2.05*	1.82*
Self	0.65 (1.79)	0.46 (0.99)	0.10 (0.41)	0.13 (0.43)	1.61	1.65
Physical	1.27 (2.85)	0.73 (1.66)	0.07 (0.37)	0.17 (0.46)	2.25*	1.78a

Note. CHAOS and OMS scores are raw (sum of item) scores. Attention = Attention Problems; Hyperactive = Hyperactivity–Impulsivity; Oppositional = Oppositional Behavior; Conduct = Conduct Problems.

^ap < .10.

^*p < .05.

^**p < .01.

^***p < .001.

Results of a 2 (Group – CI/NH) × 2 (Time – Visit 1/Visit 2) repeated-measures ANOVA for CHAOS total score were significant for main effects of Group, F(1, 54) = 6.23, p = .016, and Time, F(1, 54) = 7.06, p = .010. The Group × Time interaction was nonsignificant, F(1, 54) = 0.68, p = .415. A repeated-measures t test for CHAOS total score from Visit 1 to Visit 2 in the combined CI and NH samples revealed a significant decline in scores (fewer externalizing behaviors) over time, t(55) = 2.61, p < .02. Within each sample, CHAOS score declined significantly for the CI sample between Visit 1 and Visit 2, t(25) = 2.36, p < .03, but did not change significantly for the NH sample between Visit 1 and Visit 2, t(29) = 1.36, p = .18. Additional follow-up t tests indicated that the CI sample scored higher than the NH group (indicating greater externalizing behaviors) at both Visit 1 and Visit 2 (see Table 2 and Figure 1).

Figure 1.

Conduct–Hyperactive–Attention Problem–Oppositional Symptom total raw scores in samples of preschool-aged children with cochlear implants (CI) or normal hearing (NH) at two visits (Visit 1 and Visit 2) separated by a 1-year interval. Whiskers are ± 1 standard error.

At Visit 1, the mean OMS total raw score in the CI sample (see Table 2) was greater than the clinical cutoff value of 4 (Kronenberger et al., 2007). The mean NH sample OMS total raw score fell below the OMS clinical cutoff value. In the CI sample at Visit 1, 11 participants (42.3%) scored above the OMS clinical cutoff value, compared to three participants (10%) in the NH sample (Fisher's exact p value = .012). At Visit 2, the percentages of the CI and NH samples falling above OMS clinical cutoff values were 34.6% (N = 9) and 16.7% (N = 5), respectively (Fisher's exact p value = .14).

For the OMS total score, the main effect for group was statistically significant, F(1, 53) = 5.21, p = .026, but effects for time, F(1, 53) = 0.29, p = .592, and Group × Time, F(1, 53) = 2.97, p = .091, did not reach significance. Follow-up t tests indicated that the CI sample had greater numbers of outbursts than the NH sample at Visit 1 but not at Visit 2 (see Table 2). Inspection of the pattern of scores across the hearing group samples revealed much larger variability in the CI sample than in the NH sample (see Figure 2); furthermore, the CI sample showed a numeric decline in outbursts, whereas the NH sample showed a slight increase in outbursts, across the 1-year study period (see Figure 2). Nevertheless, comparisons of Visit 1 and Visit 2 OMS total scores within each of the hearing samples were nonsignificant for the CI sample, t(25) = 1.19, p = .25, and for the NH sample, t(28) = 1.46, p = .16.

Figure 2.

Outburst Monitoring Scale total raw scores in samples of preschool-aged children with cochlear implants (CI) or normal hearing (NH) at two visits (Visit 1 and Visit 2) separated by a 1-year interval. Whiskers are ± 1 standard error.

In order to better understand findings for total scores of the CHAOS and OMS, exploratory analyses were conducted using subscale scores. The CI sample reported more externalizing problems on the CHAOS Attention Problems subscale at Visit 1 and Visit 2 (see Table 2). For Visit 1 only, the CI sample scored higher than the NH sample on CHAOS Conduct Problems, OMS Verbal, OMS Property, and OMS Physical, reflecting more problems in these externalizing subdomains. In general, differences between the CI and NH samples were smaller at Visit 2 than at Visit 1, mirroring findings with the total scores (see Table 2).

Correlations between Visit 1 and Visit 2 total scores of the CHAOS and OMS revealed a high degree of consistency over time in the CI sample and a lower degree of consistency in the NH sample, particularly for outbursts on the OMS (see Table 3). For the CI sample, large and statistically significant correlations were found between all CHAOS subscales and for all OMS subscales except Self, which was infrequently endorsed by both samples (see Table 2). For the NH sample, large and statistically significant correlations were found for CHAOS Attention Problems, Hyperactivity–Impulsivity, and Oppositional Behavior subscales, as well as for the OMS Verbal subscale (see Table 3). The other CHAOS and OMS subscales were infrequently endorsed in the NH sample (see Table 2).

Table 3.

Conduct–Hyperactive–Attention Problem–Oppositional Symptom (CHAOS) and Outburst Monitoring Scale (OMS) externalizing score correlations between Visits 1 and 2.

Measure	Cochlear implant	Normal hearing
CHAOS
Total	.80***	.56***
Attention Problems	.58**	.73***
Hyperactivity–Impulsivity	.68***	.52**
Oppositional Behavior	.75***	.51**
Conduct Problems	.84***	.24
OMS
Total	.84***	.27
Verbal	.86**	.53**
Property	.74***	.16
Self	.28	−.08
Physical	.87***	−.07

Note. Values are Pearson correlation coefficients between Visit 1 and Visit 2 CHAOS and OMS total and subscale scores.

^**p < .01.

^***p < .001.

Association of Demographic and Hearing History Variables With Externalizing Behavior Scores in CI Users

In the sample of preschool CI users, higher family income was associated with lower externalizing scores on the CHAOS and OMS at Visit 1 (see Table 4). Poorer preimplant residual hearing was associated with fewer externalizing problems in the form of lower CHAOS total scores at Visits 1 and 2, as well as lower OMS scores at Visit 2. Chronological age, age at implantation, and duration of use were not associated with CHAOS or OMS scores at any visits.

Table 4.

Correlations between externalizing scores and Visit 1 demographics and hearing history (cochlear implant sample).

Measure	Visit 1
	Age	Income	Best PTA	Age at implant	Duration of use
CHAOS total
Visit 1	−.31	−.41*	−.49*	.16	−.30
Visit 2	−.27	−.27	−.51*	.19	−.30
OMS total
Visit 1	−.34a	−.47*	−.37a	−.02	−.19
Visit 2	−.27	−.20	−.43*	−.03	−.14

Note. Values are Pearson correlation coefficients between CHAOS and OMS total scores and chronological age, family income, best ear preimplant pure-tone average (PTA), age at implantation, and duration of implant use. CHAOS = Conduct–Hyperactive–Attention Problem–Oppositional Symptom; OMS = Outburst Monitoring Scale.

^ap < .10.

^*p < .05.

Association of Nonverbal Intelligence, Language, and Coping Flexibility With Externalizing Behavior Scores

Problems in coping flexibility (higher BRIEF-P Shift scores) at Visit 1 were strongly and significantly associated with more externalizing problems (higher CHAOS and OMS total scores) at Visit 1 and Visit 2 (see Table 5). Nonverbal intelligence and vocabulary were not significantly correlated with CHAOS or OMS scores at either visit. Analyses of correlations between change scores on measures of externalizing behaviors, nonverbal IQ, vocabulary, and coping flexibility indicated that greater improvement in coping flexibility was associated with greater improvement in CHAOS total externalizing behaviors over time (see Table 6).

Table 5.

Correlations between externalizing scores and Visit 1 nonverbal intelligence, vocabulary, and coping flexibility (cochlear implant sample).

Measure	Visit 1
	Nonverbal intelligence (DAS-II PS)	Vocabulary (PPVT-4)	Coping flexibility (BRIEF-P Shift)
CHAOS total
Visit 1	−.15	−.01	.59**
Visit 2	−.11	−.09	.41**
OMS total
Visit 1	−.35a	−.07	.64***
Visit 2	−.30	.16	.60***

Note. Values are Pearson correlation coefficients between CHAOS and OMS total scores and DAS-II PS T score, PPVT-4 standard score, and BRIEF-P Shift T score. DAS-II PS = Differential Abilities Scales–Second Edition Picture Similarities; PPVT-4 = Peabody Picture Vocabulary Test–Fourth Edition standard score; BRIEF-P = Behavior Rating Inventory of Executive Function–Preschool Version Shift T score; CHAOS = Conduct–Hyperactive–Attention Problem–Oppositional Symptom; OMS = Outburst Monitoring Scale.

^ap < .10.

^**p < .01.

^***p < .001.

Table 6.

Correlations between 1-year change in externalizing scores and nonverbal intelligence, vocabulary, and coping flexibility (cochlear implant sample).

Measure	Visit 1 – Visit 2 change
	Nonverbal intelligence (DAS-II PS)	Vocabulary (PPVT-4)	Coping flexibility (BRIEF-P Shift)
Visit 1 – Visit 2 change
CHAOS total	−.16	.30	.55**
OMS total	−.14	.19	−.11

Note. Values are Pearson correlation coefficients between change (calculated by subtracting Visit 2 score from Visit 1 score) in CHAOS and OMS total scores and DAS-II PS T score, PPVT-4 standard score, and BRIEF-P Shift T score. DAS-II PS = Differential Abilities Scales–Second Edition Picture Similarities; PPVT-4 = Peabody Picture Vocabulary Test–Fourth Edition standard score; BRIEF-P = Behavior Rating Inventory of Executive Function–Preschool Version Shift T score; CHAOS = Conduct–Hyperactive–Attention Problem–Oppositional Symptom; OMS = Outburst Monitoring Scale.

^**p < .01.

Discussion

In this study, we investigated two broad research questions: (a) between-groups differences comparing parent-reported externalizing behavior problems in preschool-aged children with CIs and NH peers and (b) within-group factors associated with variability in parent-reported externalizing behavior problems in preschool-aged children with CIs. Externalizing behaviors and within-group factors were assessed longitudinally at two visits separated by 1 year, allowing for investigation of predictive associations, stability, and change in externalizing behavior outcomes. Study results demonstrated higher levels of externalizing behavior problems in CI users compared to NH peers at Visit 1, although these differences were less pronounced 1 year later. Externalizing behaviors were stable over the 1-year study period, especially in the CI sample, as demonstrated by statistically significant Visit 1–Visit 2 correlations and nonsignificant changes in mean scores (although the scores showed numeric declines, especially in the CI sample). Variability in externalizing behaviors was large in the CI sample, exceeding that of the NH sample for outbursts on the OMS. Within-group analyses of the CI sample demonstrated strong associations between greater Visit 1 coping flexibility and fewer concurrent (Visit 1) and future (Visit 2) externalizing behavior problems. Furthermore, improvement in coping flexibility over the 1-year study period was associated with a decrease in externalizing behavior problems during that time period. Contrary to expectations, neither vocabulary nor nonverbal intelligence was associated with externalizing behaviors in the CI sample.

Prior investigations of psychological adjustment in preschool-aged children with CIs have used global, broadband measures of behavior problems that incorporate externalizing (acting out, disruptive) and internalizing (internal emotional distress, anxiety, depression) symptoms. These measures include the Kiddy KINDL (Warner-Czyz et al., 2009) and the Strengths and Difficulties Questionnaire (Wong et al., 2017). Other studies (Jamsek et al., 2022; Kronenberger et al., 2020) have investigated behavioral outcomes in preschool-aged CI users with measures assessing behavioral subdomains of executive functioning such as inhibition and working memory using the BRIEF (Gioia et al., 2003) or Learning, Executive, and Attention Functioning scale (Castellanos et al., 2018b; Kronenberger et al., 2018). The current investigation differs from prior work in its use of measures exclusively focused on externalizing behaviors using two different assessment frameworks: the CHAOS and the OMS. The CHAOS scale is based on behavioral symptoms of disruptive behavioral disorders as defined in the DSM (Levy et al., 2017) and therefore reflects the most common diagnostic manifestations of externalizing behaviors in the form of hyperactivity, attention problems, oppositional-defiant behavior, and conduct/delinquent behavior. The OMS, on the other hand, assesses specific types of behavioral outbursts consisting of verbal aggression, physical aggression, aggression to property, and aggression to self.

Between-groups analyses demonstrated greater amounts of total externalizing behaviors in the sample of preschool-aged CI users compared to NH peers at Time 1 (aged 3–4 years), as measured by diagnostic behavioral criteria (CHAOS) and aggressive outbursts (OMS). Additionally, the CI sample showed greater variability in total externalizing behaviors compared to the NH sample. Further analyses of subscales suggested that differences in attention and conduct problems on the CHAOS scale contributed most significantly to CHAOS total score differences between CI and NH samples at Time 1. Because CHAOS Attention Problems subscale items reflect deficits in attention and organization, the group differences in attention problems are consistent with prior work showing executive functioning delays in preschool-aged CI users (Beer et al., 2014; Jamsek et al., 2022; Kronenberger et al., 2020). CHAOS Conduct Problems subscale items assess angry-aggressive behaviors such as bullying, threatening, fighting, and destroying things (Levy et al., 2017); differences in CHAOS Conduct Problems subscale scores therefore indicate greater angry aggressive behaviors reflected in outbursts in preschoolers with CIs. On the OMS, CI users were rated as showing greater amounts of outbursts related to verbal aggression, physical aggression, and aggression to objects than NH peers at Time 1, which is consistent with the CHAOS Conduct Problems differences in aggression/outbursts between groups. Thus, CHAOS and OMS total and subscale scores indicate differences in cognitive (attention) and behavioral (anger/aggression) self-control in preschool-aged CI users compared to NH peers, which are manifest most significantly in behavioral outbursts.

The between-groups differences in externalizing behaviors at Time 1 were less pronounced 1 year later at Time 2, although significant Time 1–Time 2 correlations indicate that the relative amount of externalizing behaviors across participants in the samples was consistent over time. While CHAOS total score differences between CI and NH samples remained significant at Time 2, only the Attention Problems subscale score was significantly poorer in CI users than in NH peers at Time 2. Furthermore, differences in OMS total scores were no longer significant between the samples at Time 2. These findings suggest the possibility of fewer externalizing behavioral differences—especially in outburst-related behavioral problems—between CI users and NH peers with older age and development. On the other hand, none of the Group × Time interactions were significant in ANOVAs for CHAOS or OMS total scores, and a main effect of Time was found only for CHAOS total score, which declined between Visit 1 and Visit 2. In a previous study, Kronenberger et al. found that inhibition and working memory skills improved significantly over time between ages 3 and 7 years in CI users but not in NH peers (Kronenberger et al., 2020). Thus, findings of the present study are consistent with prior work indicating improvement in externalizing and related (e.g., executive functioning) behaviors during preschool ages in CI users, but more research is needed to establish whether this improvement occurs at a greater rate than that of NH peers. Future investigations of changes in psychological adjustment in preschool-aged CI users longitudinally during development are strongly recommended.

Within-group analyses showed few associations between demographics/hearing history and externalizing behavior problems in the preschool-aged CI sample. Chronological age, age at implantation, and duration of CI use were unrelated to externalizing behaviors, and income showed only a concurrent (not predictive) association with CHAOS and OMS scores at Time 1. Surprisingly, poorer residual hearing prior to implantation was associated with fewer externalizing problems as measured by both the CHAOS and the OMS. This finding is unexpected and contrasts with research suggesting that better functional auditory outcomes are associated with more positive psychosocial outcomes (Huttunen & Välimaa, 2010; Wong et al., 2017) in preschool-aged DHH children. However, studies showing associations between better auditory functioning and positive psychosocial outcomes assess functional auditory outcomes after hearing intervention (Huttunen & Välimaa, 2010; Wong et al., 2017). It is possible that children with poorer residual hearing before intervention receive more intensive interventions and more attention to overall psychosocial functioning because of their more severe hearing losses, resulting in improved psychosocial outcomes. Alternatively, this may be a chance finding, given that none of the children in the CI sample had usable residual hearing prior to implantation and that all children with CIs had identifiable hearing loss at birth. Additional study of this finding is warranted.

Within-group correlations of coping flexibility with externalizing behavior outcomes consistently demonstrated that greater coping flexibility was associated with fewer externalizing behavior problems, with correlation effect sizes in the medium-to-large range (.41–.64). These associations were not only concurrent but also predictive: Greater coping flexibility at Time 1 predicted fewer externalizing problems on both the CHAOS and OMS scales at Time 2, 1 year later. Furthermore, improvement in coping flexibility during the 1-year period between visits was associated with improvement in externalizing behavior problems on the CHAOS scale during the same period. This suggests that not only static (measured at one time) but dynamic (changes over time) coping flexibility is predictive of externalizing behavioral outcomes.

The positive association between better coping flexibility and better behavioral outcomes in preschool-aged CI users is consistent with decades of research demonstrating that coping behaviors are significant determinants of psychosocial outcomes in children and adolescents, especially when coping with the stresses of physical conditions (Thompson & Gustafson, 1996). Coping flexibility was assessed in the present study using the BRIEF-P Shift subscale and includes item content such as “has trouble changing activities” and “takes a long time to feel comfortable in new places or situations.” This ability to cope by regulating emotion and by flexibility adjusting behavior in the face of change or stress is likely to be very important for children with CIs, who face significant needs for auditory, social, learning, communication, and behavioral adaptation after receiving an implant. Stress/challenge in and of itself does not cause a negative psychosocial outcome; rather, the adaptation of the individual in the form of engaging and coping with stress is the primary determinant of outcome under stress (Kronenberger et al., 1996, 1998; Lazarus & Folkman, 1984). This process may explain the significant variability in externalizing behavioral outcomes and the association between coping flexibility and externalizing behavioral outcomes in preschool-aged CI users. Specifically, the coping response of the child may be more important in determining psychosocial outcome than the stress of implantation or language challenges, which affect all preschool-aged CI users. Thus, while additional research is needed to better understand this process of stress, coping, and adaptation in CI users, well-established stress-and-coping models may offer valuable insight into psychosocial adjustment following cochlear implantation (Lazarus & Folkman, 1984; Thompson & Gustafson, 1996).

Contrary to expectations, neither vocabulary nor nonverbal intellectual ability was associated with externalizing behavior problems in preschool-aged CI users. No participant in the current study had a nonverbal IQ less than 1 SD below the mean, and average nonverbal IQ values for both the CI and NH samples were higher than the normative mean. Thus, within this range of average or better intellectual ability, nonverbal IQ was not associated with externalizing behavior problems in the current sample. However, research that included CI users with intellectual and other disabilities has found an association of multiple other disabilities and psychosocial problems (Leigh et al., 2015; Wong et al., 2017). Hence, intellectual ability in the very low ranges may be more predictive of adverse psychosocial outcomes, whereas differences in intellectual ability within the average and above-average ranges may be less relevant for explaining individual differences in psychosocial outcomes.

The absence of a language-externalizing behavior association is contrary to findings from other studies demonstrating associations between language development and psychosocial outcomes in CI users (Leigh et al., 2015; Wong et al., 2017). This discrepancy in findings may be a result of language assessment strategies in different studies. In the present study, vocabulary was used as an index of language skills because of the young age of the children and the time constraints of the study visit. Other studies finding associations between language and psychosocial outcomes have focused on the communication components of language or on higher order language characteristics such as speech intelligibility, comprehension, expression, or discourse (Barker et al., 2009; Freeman et al., 2017; Kronenberger et al., 2024; Leigh et al., 2015; Studts et al., 2022; Wong et al., 2017). The ability to use language for purposes of comprehension and concept formation is much more likely than vocabulary size to be associated with effective coping, because coping thoughts/behaviors involve processing and execution of conceptual thinking. These discrepancies highlight the importance of the type of language skill assessed and suggest that language knowledge is less important than language application for psychosocial adjustment in CI users (Kronenberger et al., 2024).

The results of this study should be interpreted in the context of several methodological characteristics and limitations. First, although longitudinal methods were used to investigate predictive associations, causal conclusions cannot definitively be drawn from the current data because variables were not experimentally manipulated. On the other hand, predictive associations were consistent with expectations of coping affecting future externalizing behavior outcomes. Second, measures of externalizing behaviors and coping behaviors were based on parent report, which may be affected by rater bias, lack of awareness, or personality factors such as general negativity or positivity. Use of a multirater approach and behavioral coding measures of stress, coping, and psychosocial adjustment are recommended in future studies (Holmbeck et al., 2002). Third, the samples differed in nonverbal intelligence and vocabulary, which could partially explain between-groups differences in externalizing problems. However, because these factors were not associated with within-group variability in the CI sample, they are unlikely to be the primary factor explaining between-groups differences. Fourth, we used only one measure of coping, the BRIEF-P Shift subscale, as a measure of coping flexibility. Coping is a complex multifactorial construct, and investigating with more detailed measures of coping (Kronenberger et al., 1996, 1998) is recommended in future studies. Investigation of more specific coping behaviors, such as information seeking, using social support, and avoidance, may also suggest specific intervention strategies to improve psychosocial outcomes by improving positive coping strategies and reducing negative ones. Furthermore, investigation of stressors that might deplete limited cognitive resources, such as listening fatigue, is recommended in future studies of stress, coping, and psychosocial outcomes after cochlear implantation. Finally, because of limited sample sizes, we were able to investigate only two primary externalizing outcome variables (CHAOS and OMS total scores) without using corrections for statistical tests, in order to balance between alpha and beta error. Results with CHAOS and OMS subscales are exploratory and should be interpreted with caution. Future studies with larger sample sizes will allow for investigation of additional variables using greater statistical power.

In summary, study results indicate that preschool-aged children with CIs are at greater risk for externalizing behavior problems compared to NH peers and that variability in externalizing behavior problem outcomes within the group of CI users is explained by individual differences in coping flexibility. Clinically, early assessment using measures specifically designed for the detection of externalizing problems may help to identify at-risk children at very young ages, for purposes of prevention or clinical referral programs. Programs designed to improve coping flexibility may reduce the severity or persistence of externalizing behavior problems. Future research is recommended to better understand the stress, appraisal, and coping processes involved in the psychosocial adjustment of preschoolers following cochlear implantation.

Data Availability Statement

Data are available to individuals within the scientific community upon request from the authors.

Funding Statement

This work was supported by the National Institute on Deafness and Other Communication Disorders Grants R01DC009581 (awarded to D.B.P.), R01DC015257 (awarded to W.G.K. and D.B.P.), and R01DC021339 (awarded to I.C.).