- Top of page
- Key Practitioner Message:
Background: Relatively few treatment studies address mental health issues in very young children. This study examined the effectiveness of a treatment program for toddlers whose behaviour problems were further complicated by living in poverty.
Method: An evidence-based treatment program was adapted for use in the homes of 102 toddlers for an average of 12 weekly sessions.
Results: Significant improvements were found for the children’s behaviour problems and their compliance to parent requests.
Discussion: The inherent challenges in working with at-risk families and the challenges in delivering mental health services for very young children living in poverty are discussed.
Challenging behaviours are common during the toddler and preschool years and for some children, they may escalate into severe tantrums, self-injury, aggression, destructiveness, oppositional behaviour and hyperactivity. Campbell (1995) reported that between 10-15% of young children have mild to moderate behaviour problems that may persist well into the child’s formal school years (Campbell, 1997), increase in severity (Hofstra, Van der Ende, & Verhulst, 2002), adversely affect the children’s relationship with caregivers (Greene & Doyle, 1999), and interfere with their development of social (Mendez, Fantuzzo, & Cicchetti, 2002) and communication skills (Sigafoos, 2000). Young children’s behaviour problems also have been associated with higher levels of parental stress (Eyberg, Boggs, & Rodriguez, 1992).
There are a number of factors that contribute to the onset, escalation and persistence of behaviour problems in young children. In their review, Huaqing Qi and Kaiser (2003) reported that preschool children from low-income families had a significantly higher incidence of behaviour problems (31%) than those in the general population. To better understand this relationship of behaviour problems and poverty, Fox, Platz, and Bentley (1995) found that younger, single, less-educated mothers from lower-income levels tended to use more frequent verbal and corporal punishment when parenting their young children who also had more behaviour problems. Brenner and Fox (1998) reported that the best predictor of behaviour problems in young children was parental use of verbal and corporal punishment. This relationship between behaviour problems in children and harsh parenting practices was also reported in other research (Baker & Heller, 1996; Nix et al., 1999); however, the direction of this relationship is not clear. That is, do behaviour problems elicit harsh parenting practices or do harsh parenting practices result in behaviour problems? In addition to negative parenting practices, a young child with developmental delays is at an increased risk for having behaviour problems (Feldman et al., 2000; Roberts et al., 2003). Baker et al. (2002) found that children with delays were three to four times more likely to obtain clinically significant scores on a child behaviour scale than their non-disabled peers. Einfeld and Tonge (1996) reported that 41% of children with intellectual delays had severe behaviour and emotional problems.
A number of evidence-based parent management programs have emerged to address the mental health needs of young children using developmentally-appropriate treatment strategies (Eyberg, Boggs, & Algina, 1995; Fox & Nicholson, 2003; Sanders, Mazzucchelli, & Studman, 2003; Webster-Stratton, 2001). However, relatively few of these programs address very young children living in poverty. In one study including 882 children in Head Start programs, Reid, Webster-Stratton, and Baydar (2004) reported significant improvements in the children’s behaviour problems following caregiver participation in a group-based, parenting program. Nicholson, Brenner, and Fox (1999) provided a 10-week, group-based program of parent management training in community-based, nonprofit agencies for 143 children and their mothers. Results showed that parents were more nurturing and used less corporal and verbal punishment with their young children; also children’s problem behaviours reduced significantly. In a controlled study of low-income parents who reported frequent pre-treatment use of corporal and verbal punishment with their young children, Nicholson et al. (2002) reported significant post treatment reductions in these negative parenting practices, parental anger, parent distress, and child behaviour problems.
Fewer studies addressed children with developmental delays. In a U.S. Department of Education survey (2002) of children enrolled in early intervention programs, less than 3% received any mental health services at all. Roberts et al. (2006) implemented a parenting program in a clinic setting for families with preschoolers who were mildly delayed with behaviour problems. The outcomes included improvements in the children’s behaviour and parent-child interactions and reduced parental stress. Similarly, group adaptations of Parent-Child Interaction Therapy (Eyberg et al., 1995) and the Incredible Years Parent Training series (Webster-Stratton, 2001) for parents of young children with developmental delays demonstrated improvements in child compliance, behaviour problems, and parenting practices (Bagner & Eyberg, 2007; McIntyre, 2008). The majority of treatment studies involving young children with behaviour problems used a group-based, parent education class format in clinics or community-based sites. Families living in poverty often experience significant barriers that limit their access to such programs including transportation difficulties, caring for multiple children, problems keeping schedule appointments, and reduced motivation (Snell-Johns, Mendez & Smith, 2004). As such, this traditional model of delivering parent management programs may not meet the needs of these families.
The purpose of the present study was to examine the effectiveness of a parent treatment program for very young children with behaviour problems, most of whom also had significant developmental delays and were living in poverty. The treatment program was adapted from an existing parenting program with proven efficacy (Fox & Nicholson, 2003). Because of the multiple barriers these families living in poverty face in accessing mental health services, the program was designed to be individually delivered to families in their homes.
- Top of page
- Key Practitioner Message:
At intake, families who completed the treatment program did not differ from families who were non-completers on any of the study’s outcome measures. Completers participated in an average of 12.76 weekly, in-home sessions (SD = 5.30) over a mean of 4.91 months (SD = 2.57) with an average attendance rate of 79.2% (SD = 14.99). Non-completers participated in an average of 3.92 sessions (SD = 4.48) over a mean of 3.64 months (SD = 3.17) with an average attendance rate of 51.6% (SD = 36.37). Completers and non-completers significantly differed on all three participation measures (p < .05).
Although treatment completers participated in the intake and post-treatment evaluations, missing data for the study’s measures occurred for a variety of reasons. The child’s significant oppositional behaviour often precluded their participation in the direct observational measures. Additionally, some parents refused to play with their children to allow us to observe and rate the quality of the parent-child interactions. There were situations where clinicians chose to initiate treatment at the expense of data collection because the child’s behaviour was sufficiently severe to cause harm. At times, the numbers of different people present in the home jeopardised the confidentiality of the caregiver or created a noisy and chaotic atmosphere that was not suited for data collection. For some families, the caregivers did not understand the self-report instruments, thus invalidating their responses.
For the treatment completers, repeated measures, multivariate analyses of variance (MANOVA) were used to assess changes from intake to treatment completion. Significant MANOVAs were followed up with univariate F tests to determine the source of the significance (Table 1). The first MANOVA was computed for the observational measures of the parent-child interactions and revealed a significant time effect [F(2,65) = 22.37, p < .01] with a moderate effect size (.41), which was due to significant improvement in both the child and parent behaviour scores. A significant time effect also was found for the children’s compliance, parent use of the child’s name and the parent complimenting the child during compliance testing [F(3,46) = 17.19, p < .01]; the effect size was moderate (.53). Based on the ECBI, children’s behaviour problems improved significantly over time [F(2,89) = 39.35, p < .001] with a moderate effect size (.47) that was due to significant decreases in both the intensity and number of children’s problem behaviours. The Parent Behavior Checklist showed significant change over time [F(3,81) = 7.39, p < .001] with a small effect size (.22) that was due to a significant increase in parent expectations.
Table 1. Treatment outcomes based on parent and child measures at intake and treatment completion
|Child behaviour scores||18.30||2.02||20.96a||2.91||39.65||1,66||.38|
|Parent behaviour scores||20.72||3.15||23.03a||2.65||24.95||1,66||.27|
|% Compliance to parental requests||36.81||25.73||74.71a||42.87||37.37||1,48||.44|
|Parent uses child’s name||1.53||0.68||1.78b||0.47||4.02||1,48||.08|
|Parent praises child’s compliance||0.89||0.77||1.61a||0.61||33.33||1,48||.41|
|Eyberg Child Behavior Inventory|
|Parent Behaviour Checklist|
|Parent-Child Relationship Scale||58.34||12.41||72.54a||13.39||114.52||1,81||.59|
Eyberg and Pincus (1999) recommended a t-score of 60 as a cutoff score to determine if the child’s scores on the ECBI’s intensity and problems scales were clinically significant. The proportion of children who met the cutoff score at intake but not at treatment completion changed significantly for the intensity (χ2 (1) = 17.41, p < .001) and problem scores (χ2 (1) = 18.08, p < .001). For the intensity measure, 64.5% met the cutoff criteria at intake compared to 47.3% at treatment completion; for the problem measure, 63.7% met the cutoff criteria at intake compared to 34.1% at treatment completion.
Of the children who received a formal DSM Axis I psychiatric diagnosis at intake, 79.3% were oppositional defiant disorder, 8.5% separation anxiety disorder, 2.4% attention deficit hyperactivity disorder, and the remaining 9.8% were other disorders (autism, conduct disorder, reactive attachment disorder). At intake, 82.7% of the sample received a psychiatric diagnosis; at the end of treatment, 21.4% of the children met the criteria for a psychiatric diagnosis.
In order to assess caregiver satisfaction with the parent management program, total scores were computed by summing the parent ratings for the seven items comprising this scale with a possible range of scores from 7 (low satisfaction) to 49 (high satisfaction). The average score on this measure was 44.40 (SD = 4.00).
- Top of page
- Key Practitioner Message:
The results of this study demonstrated that the parent management program was an effective intervention program for toddlers with behaviour problems living in poverty. Parent-child interactions improved from pre- to post-treatment based on direct observations of the parents and children in their homes. Not only did the quality of their interactions and reciprocity improve, but children’s compliance to parental requests also increased significantly. The latter finding may be due to the parents’ improving their skills at obtaining their child’s attention before giving a request and following their child’s compliance with social rewards. These direct observational data of improvement in the children’s behaviours are particularly compelling as very young children normally will not ‘fake’ behaviours to present themselves in a socially desirable manner. Consequently, observing these young children in their natural settings, while inherently challenging for optimal data collection, provided a reliable and valid method for assessing change.
The study’s findings also indicated that the parent treatment program was associated with significant reductions in the frequency and severity of the children’s behaviour problems. The effect sizes for these pre to post treatment changes were moderate which is consistent with other research (Behan & Carr, 2000; McIntyre, 2008). Parents also increased their expectations for their children over the course of treatment. The increase may be due to parents adjusting their expectations to the child’s developmental growth over the course of treatment (average of five months). In addition, unreasonable parental expectations were consistently challenged throughout treatment. Parental discipline scores did not change at post treatment which is not consistent with previous research (Nicholson et al., 1999, 2002). This finding may be due in part to the parents’ reluctance to honestly share their negative parenting practices at intake with an unfamiliar clinician, perhaps in part out of a fear of being reported to the authorities for child abuse. Throughout treatment, clinicians did report increased parental use of effective limit setting strategies (time-out). The majority of children with a psychiatric diagnosis at intake no longer met the criteria for a psychiatric diagnosis at treatment completion. This finding supports providing early intervention for children before their behaviour patterns become more resistant to change. Parents reported high levels of satisfaction with the treatment services. These results support the findings from the literature regarding the effectiveness of behavioural family interventions for young children with behaviour problems and developmental disabilities (Gavidia-Payne & Hudson, 2002).
One limitation of the present study was the absence of a control group. The efficacy of the treatment strategies that were employed in this study has been well established in the literature. However to our knowledge, no effort has been made to apply these strategies to very young children with delays and behaviour problems from families in poverty in their home settings. While we had initially intended to include a wait-list control group, we quickly learned that this would not be reasonable given the significant difficulty we experienced in initially engaging our families and in maintaining them throughout the treatment program. Our treatment attrition rate of 57%, which is higher than the 33% reported in other treatment studies for families of children with developmental disabilities (Roberts et al., 2006) and the 50% rate for children from low-income families (Nicholson et al., 1999), exemplifies the inherent challenges of providing mental health services to this population. Our analyses of family intake data also suggested that it would be difficult to identify those families who were likely to benefit from the parenting program based on this information alone. We also did not include a follow-up condition to determine if the treatment effects were maintained over time. We are presently engaged in a follow-up study that will essentially repeat the study’s outcome measures one to three years after treatment completion. However, we are already finding this study has inherent limitations. One characteristic of our families is their transient nature. In addition to moving frequently, families routinely have their telephone services disconnected. Consequently, locating these families for follow-up has been challenging.
The sample for this study was somewhat heterogeneous including young boys and girls with and without developmental delays, families living in and not in poverty, caregivers with and without marital support systems, and different ethnic groups. The treatment program selected for this study has been previously empirically validated for all of these child and family variables with the exception of the presence of a developmental delay. In a recent study comparing this treatment program between children with and without delays and behaviour problems (Holtz et al., in press), results showed that the parent management program was equally effective for both groups. Consequently, practitioners should feel confident in using this treatment program with young children coming from a variety of different family backgrounds.
This treatment program included several components including non-directive play, teaching parents to monitor their thoughts and feelings when interacting with their child, instructing caregivers on how to maintain appropriate developmental expectations for their children, procedures to strengthen children’s prosocial behaviours, limit setting strategies, and parent incentives. In addition, this program was tailored to meet the unique circumstances of each family and delivered in their homes. The study’s design did not permit us to ferret out the unique contributions of the separate treatment components to the program’s effectiveness. However, the use of multiple strategies to address the myriad of child, family, and environmental factors that contribute to behaviour problems in young children is consistent with other programs reported in the literature (Eyberg et al., 1995; Sanders et al., 2003; Webster-Stratton, 2001), and most importantly, shares their adherence to a foundation in social learning theory and cognitive behavioural treatment approaches.
Clearly, engaging these at risk families in early mental health intervention efforts is important. In the absence of quality parental involvement, many young children who need these mental health services will not receive them in the early childhood period, when they are likely to have their greatest impact. New strategies will be needed to identify these children with significant behaviour problems as early as possible and to attempt to overcome family barriers that will interfere with their participation in treatment. New research is emerging to help screen for these early behaviour problems in very young children from low-income families (Holtz, Fox, & Meurer, 2008) and to begin to more systematically address barriers to treatment attendance and adherence (Nock & Ferriter, 2005). The present study showed that families who drop out of treatment did so around the fourth session. We currently have modified our parent incentive system to provide a larger incentive after the third session and an even larger one at treatment completion rather than smaller incentives at each treatment session to increase caregivers’ motivation to complete the treatment program. We know how to effectively change young children’s behaviour problems. Now we have to get better at engaging families who resist our treatment efforts.