Abstract
A first-line screening instrument, the Preschool Inventory of Trauma Symptoms (PITS), was developed to assess trauma symptoms with a diverse sample of 150 toddlers and preschoolers (M = 2.49 years; SD = 1.12). Items reflected the current trauma literature, assessment measures, and diagnostic criteria for very young children. A principal component analysis produced a 34-item, four-factor model: Arousal and Hyper-Reactivity, Fearful Attachment, Intrusion and Re-Experiencing, and Avoidance and Negative Cognition and Mood. One validity scale, Response Style, was also developed. All scales significantly correlated (r = .45 to .81; p < .01) with preestablished trauma measures and demonstrated adequate internal consistency (α = .68 − .87). A receiver operating characteristics curve analysis identified a cut-score with good discrimination ability (.88), sensitivity (.81), and specificity (.81). In a preliminary pilot study, PITS also was found to be sensitive to trauma symptom change following participation in an evidence-based trauma informed treatment program. A copy of the PITS is included in the Appendix for free use by qualified professionals.
Since the publication of the seminal article on the relationship between adverse child experiences and health risk behavior and disease in adulthood (Felitti et al., 1998), the level of interest in trauma during childhood has escalated dramatically over the past 20 years for professionals and laypersons alike (The National Child Traumatic Stress Network, nctsn.org). Trauma refers to an event or circumstance that poses a serious threat to self or others and is coupled with extreme disturbances in behavior and/or mood (American Psychiatric Association [APA], 2013). A recent national report found 9.1 victims of maltreatment per 1,000 children in the population; unfortunately, the youngest children (3 years of age and younger) are the most vulnerable (U.S. Department of Health & Human Services, 2018). Exposure to trauma during early childhood places individuals at elevated risks for several dysfunctional pathways (Bonanno, 2004; De Young, Kenardy, & Cobham, 2011). Maltreatment that occurs during this critical period of development can disrupt emerging capacities including self-regulatory processes that can have both short- and long-term consequences (Choi & Graham-Bermann, 2018). Toddlers who have been traumatized may exhibit impairments in the areas of attachment (Zeanah et al., 2004), externalizing and internalizing behaviors (Pears, Kim, & Fisher, 2008), developmental brain differences (De Bellis et al., 2002), and manifestations of traumatic stress responses that differ from adults (Modrowski, Miller, Howell, & Graham-Bermann, 2013). Moreover, exposure to trauma that starts as early as infancy can alter a child’s long-term ability to manage stress both affectively and behaviorally (Schore, 2001). Complicating the picture further, very young children from families exposed to poverty face a disproportionate risk of exposure to trauma due to factors associated with poverty such as living in unsafe neighborhoods, increased family stress, reduced families resources, lower parent education levels, racial discrimination, experiencing significant barriers to receiving basic health-care or appropriate mental health services, and the ongoing challenges associated with daily living (Collins et al., 2010).
The importance of researching and providing mental health services to very young children who have experienced trauma has recently emerged as a focal topic in the literature, largely dispelling the prior belief that very young children are robust to the effects of early exposure to potentially traumatic events (Miller-Graff, Galano, & Graham-Bermann, 2016; Scheeringa, Zeanah, Myers, & Putnam, 2005). Despite this growing awareness, as Miller-Graff, Galano, and Graham-Bermann (2016) pointed out, all areas of preschool trauma, including assessment, diagnosis, and treatment, remain highly understudied. As a first step to meeting the mental health needs of this younger population is a comprehensive intake evaluation that includes an initial screening for possible trauma symptoms. Presently, there are very few valid, cost-effective, and efficient instruments for trauma symptom screening, particularly in very young children from diverse backgrounds living in poverty. In their review, Strand, Sarmiento, and Pasquale (2005) reported that only 4 of 35 screening tools for assessing trauma in children and adolescents were designed for preschool children. In a more recent review of studies on the trauma symptoms in children ages 0–6, Choi and Graham-Bermann (2018) concluded that many of the studies failed to use a developmental framework and instead applied adult-oriented trauma symptom frameworks in measurement and treatment. This is an unfortunate omission, given that trauma symptoms may present differently at different ages even within the early childhood period. Presently, mental health providers often will use well-established measures such as the Child Behavior Checklist (Achenbach, 1991) in their intake assessments. While such broad-spectrum measures provide valuable clinical information, they are at risk of missing important trauma symptoms. One of the first instruments developed to specifically assess trauma in children was the Trauma Symptom Checklist for Young Children (TSCYC), aged 3–12 years old (Briere, 2005). This measure while psychometrically sound with significant research support does not extend below the age of 3 and is relatively long. More recently published measures have tended to focus on older children from families often with unreported socioeconomic levels (Grasso, Felton, & Reid-Quinones, 2015; Lang & Connell, 2017). Also, most available measures do not consider the newer diagnostic criteria for Post-Traumatic Stress Disorder in Children 6 years of age and younger in the Diagnostic and Statistical Manual of Mental Disorders (DSM-V; APA, 2013). While this diagnosis does not cover the full range of trauma symptoms presented by younger children and may even exclude children in need of trauma-informed care (Love & Fox, 2019), it does include important criteria when considering appropriate diagnoses for reimbursement purposes and to further research on the range of symptoms that may need to be considered in future editions of the DSM.
The primary aim for this study was to develop a screening instrument to measure the broad range of symptoms associated with trauma in very young children in a sample of diverse families living in poverty. A secondary aim was to pilot if this new screening instrument would be sensitive to changes in trauma symptoms following participation in an evidence-based trauma-informed treatment program. Best practices in assessment include a multistage screening process to efficiently assess children for developmental problems and mental health concerns (e.g., Carter, Briggs-Gowan, & Davis, 2004). First-line screeners allow quick identification of children in need of further evaluation and possible treatment services. These first-line screeners play a vital role in early detection and should be short, inexpensive, easy to administer, and score to help promote use among a variety of mental health and related medical professionals who serve very young children that have experienced potentially traumatic events.
Method
Participants
A Midwestern University’s Institutional Review Board (IRB) approved this study, and the child’s legal guardian signed an IRB-approved informed consent form prior to participation. Participants had been referred to a clinic in a large urban area in the Midwest that specialized in treating a diverse population of very young children with significant behavior and emotional problems including trauma, the majority who were living in poverty (Fox, Keller, Grede, & Bartosz, 2007). Children with autism spectrum disorders, severe to profound intellectual disabilities, or serious physical illnesses were not included in this study. Participants included children aged 1–5 years old who were consecutively referred to the clinic and whose legal guardians agreed to participate. The resulting diverse sample of 150 children had an average age of 2.49 years (SD = 1.12); 65.3% were males with 52.0% of the sample African American, 14.0% Latino/a, 10.7% Caucasian, and 33.3% identified as multiracial. The sample included twenty eight 1-year-olds, 51 at 2 years, 40 at 3 years, 25 at 4 years, and 5 at 5 years, which is typical of the distribution referred to the clinic. Most of families (89.9%) had annual incomes below the federal poverty level. The average education of the legal guardian, the majority of whom were mothers, was below the 12th-grade level. Participants received no compensation to be included in this study.
Measures
The demographic/clinical interview form included questions to begin to establish rapport with the primary caregiver and obtain background information. The Traumatic Events Screening Inventory–Parent Report Revised (TESI-PRR; Ghosh-Ippen et al., 2002), a 24-item scale with a yes–no response format, was used to assess the children’s previous exposure to potentially traumatic events potentially traumatic events (PTEs). Items demonstrated test–retest reliability ranging from .50 to .79 interrater reliability for the original TESI (Ford & Rogers, 1997) was reported to range from 0.73 to 1.0 for the different types of traumatic events (Gray & Slagle, 2006). The TSCYC (Briere, 2005) is a 90-item checklist to assess post-traumatic stress symptoms and comorbid difficulties for children ages 3–12. Gilbert (2004) found that the TSCYC has good concurrent validity with other parent report measures. The coefficient αs for the clinical scales for this measure ranged from .81 to .93, the validity scales coefficient αs ranged from .36 to .73, and the test–retest correlation coefficients for TSCYC scales ranged from .37 to .96 (Briere, 2005; Briere et al., 2001). The eight clinical scales are (1) Post-traumatic Stress—Intrusion, (2) Post-traumatic Stress—Avoidance, (3) Post-traumatic Stress—Arousal, (4) Sexual Concerns, (5) Dissociation, (6) Anxiety, (7) Depression, and (8) Anger and Aggression. There is also a composite post-traumatic stress score consisting of the first three scales labeled Post-traumatic Stress–Total (PTS-TOT). Subscales one through five (i.e., 45 items) were used as being most consistent with the literature on trauma in very young children. For the final analysis, only the three post-traumatic scales were used as the factors that emerged in the new screening instrument were best aligned with these constructs. The Response Level (RL) Scale was also administered to examine whether concurrent validity could be established with the current measure’s proposed validity scale. The Pediatric Emotional Distress Scale (PEDS; Saylor, Swenson, Reynolds, & Taylor, 1999) is a 21-item measure that was developed to quickly assess behaviors identified as significantly elevated after trauma exposure. The PEDS consists of three factors including anxious/withdrawn, fearful, and acting out. Although the entire measure was administered for the study, only the Fearful subscale was used in final analysis as the factors that emerged in the new screener were best aligned with this construct. Saylor, Swenson, Reynolds, and Taylor (1999) reported the following reliabilities for the PEDS’ total score: α = .85, test–retest reliability =.56, and interrater reliability = .77.
Development of an Item Pool for New Screening Measure
Initial items were sampled from a review of the trauma literature in very young children, available assessment instruments for older children, and diagnostic criteria for Post-traumatic Stress Disorder for Children 6 Years and Younger (APA, 2013). In addition, items to assess overly favorable and unfavorable informant responding were included in the measure as part of a Response Style Scale. All 65 initial items were written in simple, noncolloquial language. The response format chosen was a 4-Point Frequency Likert-type Scale: 4 = always/almost always (occurs daily), 3 = often (occurs weekly), 2 = sometimes (occurs biweekly), and 1 = almost never/never (occurs less than monthly or never).
Content validity
The initial item pool of 65 items was reviewed by a focus group of 32 parents that were drawn from families with children who attended a large birth-to-three agency; 95% of families using these services met the criteria for poverty. The parent sample was primarily females (96%) that included 45% African Americans, 20% Caucasian, 20% Latino/a, and 15% multiracial. Parents individually rated each item for clarity. The rating for clarity used the following scale: 1 = did not understand item, 2 = need more information, 3 = somewhat clear, and 4 = clear meaning. Items that scored below a 2.5 on clarity were removed or modified. A second focus group included seven professionals with an average of 6.57 years (SD = 4.64) experience as child therapists and 5.35 (SD = 3.04) years of experience working with young children with trauma. The remaining 63 items were rated again on clarity and additionally rated on relevance of assessing trauma symptoms in young children. The rating for relevance used the following markers: 1 = not at all relevant, 2 = little relevance, 3 = some relevance, 4 = good relevance, and 5 = excellent relevance. Items scores below a 3 on relevance and 2.5 on clarity were considered for removal or modification. The final initial pool included 56 items. Finally, following the parent and expert reviews, a professor with a reading specialization reviewed the revised items to reduce the reading level further and increase the clarity. The reading level for the resulting measure was at a Flesch–Kincaid 3.7 grade level. This new screening measure was entitled, the Preschool Inventory of Trauma Symptoms (PITS).
Procedures
As part of the clinic’s general intake, caregivers of children who met the study’s eligibility criteria were given the opportunity to participate in this study. Only caregivers with primary responsibility for the child were included (e.g., biological parents, foster parents, grandparents). All interviews and the administration of the study’s instruments were conducted in the participants’ homes. Except for the demographic form and the TESI-PRR, the remaining instruments (i.e., TSCYC, PEDS, PITS) were administered in random order to avoid possible order effects. The measures were administered by professional counselors or clinical social workers. All measures were read item by item to the caregivers to avoid the potential impact of reading difficulties. Clinic staff provided treatment to the children of caregivers who exhibited trauma symptoms and had previous exposure to potentially traumatic events. Staff had been trained and regularly supervised in an in-home, evidence-based, trauma-informed program specifically designed for very young children in poverty (Love & Fox, 2019).
Results
Of the present sample, 81.4% of caregivers reported at least one potentially traumatic event; the average number of PTEs was 2.50 (SD = 2.11). The most frequently sited PTEs were having witnessed or heard threats of domestic violence (50.7%), separation from primary caregiver (46.3%), witnessed or experienced physical, sexual, or emotional abuse and/or neglect (40%), witnessed or experienced severe illnesses/injuries and/or medical procedures (30.7%), had seen family members arrested or in jail (24.5%), experienced the death of someone close (14.3%), and had seen or had been in serious accidents (10.0%), among other less frequent PTEs (e.g., animal attack, homelessness).
To determine the factor structure of the final 56 items, a principal component factor analysis with promax rotation was used. The 7 items used in the Response Style Scale were not included in the analysis as they were not theoretically related to the construct of trauma. The critical eigenvalues were set at one. The Kaiser–Meyer–Olkin measure of sampling adequacy was .77, which is above the recommended value of .6 (Tabachnick & Fidell, 2013). Bartlett’s test of sphericity was significant, χ2(1,540) = 3,916.67, p < .001, indicating that the correlation matrix was not an identity matrix and was appropriate for a factor model; thus, correlations were large enough to warrant a factor analysis. Additionally, communalities for all items were above .50, which provided support for the adequacy of the sample size of 150 (Worthington & Whittaker, 2006). A parallel analysis was conducted, which randomly generated eigenvalues over 1,000 iterations. In addition to the parallel analysis, the scree plot was also examined to determine how many factors to retain. Results of the parallel analysis and Scree Plot supported a four-factor model, with eigenvalues for the real data being larger than the simulated data for the first four factors. Items that failed to load on any factor (<.4) were removed. To achieve the goal of developing a relatively short screening measure with approximately the same number of items retained for each factor, additional items were eliminated based on their factor loadings, so that only the strongest loading items related to each of the four factors were retained. After each removal, the analysis was rerun. The final solution accounted for 49.17% of the variance in the sample and resulted in retention of the empirically supported four-factor model. The first factor of the PITS consisted of 8 items and was labeled Arousal and Hyper-Reactivity (ARH). The second factor consisted of 7 items and was labeled Fearful Attachment (FA). The third factor consisted of 7 items and was labeled Intrusion and Re-Experiencing (I). The fourth factor consisted of 6 items and was labeled Avoidance and Negative Cognition and Mood (AVN). Table 1 lists the standardized loadings for each of the items and their respective factors. One item (“Is shy”) loaded similarly strong on two factors and was retained as it captured an important temperament issue that could be useful for conceptualizing and treating a child’s trauma.
Preschool Inventory of Trauma Symptoms’ Item Loadings on the Four Factors.
Clinically significant symptoms on the trauma subscales were determined using a 1.5 standard deviation elevation above the mean score. While the 1.5 standard deviation elevation was used to determine cut points for the clinical scales, clinicians may wish to consider lower scores for treatment consideration to avoid missing children who may be experiencing significant symptoms. Cut points for overly positive and overly negative response styles were also similarly calculated, with 1.5 standard deviation above the mean for overly positive and 1.5 standard deviation below the mean for overly negative responses. Higher scores indicated the responder was endorsing more negative items about their child (e.g., lying, whining, being hard to be around), whereas lower scores indicated overly positive responding. The PITS subscale and total composite (TC) score statistics, including cut scores, are reported in Table 2.
Preschool Inventory of Trauma Symptoms Subscale and Total Scale Descriptive Statistics and Cut Scores Suggesting Possible Clinical Significance.
The correlation between the subscales theoretically related to trauma was examined to determine whether an overall trauma composite score for the measure would be appropriate. Correlations should be in the slight to moderate range, meaning that correlations should fall between .2 and .7 (Hamill, Brown, & Bryant, 1992). These correlations are large enough to indicate a relationship but small enough as to imply that constructs are still empirically related, but separate. Results of correlation analyses between the subscales of the measures used in this study indicated significantly small (.21) to moderate correlations (.45; all ps < .01) between all subscales on the measure, which provided support for the creation of a trauma composite total score. There were no differences in TC scores on the PITS when examining age, F(5, 142) = .38, p = .86, gender, F (1, 146) = .36, p = .55, or between families above or below the federal poverty line, F (1, 145) = 1.26, p = .264. As a result, the creation of separate norms was not warranted at this time. A cutoff score for clinical significance for the PITS-TC score was determined through a Receiver Operator Characteristics (ROC) Curve Analysis. For this analysis, the TSCYC PTS-TOT was used as the output variable to establish whether a child had clinically significant symptoms of trauma. The area under the ROC curve was .88 (95% CI [.82 to .94]; p < .001). Consequently, there is an 88% likelihood that if a clinically significant cut score on the TSCYC TC score was obtained, then a randomly selected child would have a higher TC score than would a randomly selected child who did not meet the clinical threshold on the TSCYC. ROC curve areas of .80 to .90 are considered good discriminators (Swets, 1996). Because the PITS is meant to be a first-line screening tool, there is an emphasis on minimizing false-negative results; therefore, false-negative results were weighted higher than the false positives when deriving a cut score. The optimal criterion score considered the cost of different decision categories (e.g., false positive) using the generalized Youden index (Schisterman, Perkins, Liu, & Bondell, 2005), with 1 as the value for the cost of a false positive and to 1.5 as the value for false negative. The final cut score for clinical significance on the PITS-TC was 51. Sensitivity for this cut score was .81 and specificity was also .81, which met Mouthaan, Sijbrandij, Reitsma, Gersons, and Olff’s (2014) recommendations.
Reliability
The coefficient αs were .85 for the AHR subscale, .81 for the FA subscale, .75 for the I subscale, .68 for the AVN subscale, and .72 for the Response Style subscale. The internal consistency for the TC Trauma Scale (all subscales except the Response Style subscale) was .87. The average inter-item correlation was .41, .37, .30, .25, .26, and .20, respectively, for each of the domains.
Concurrent Validity
To further assess validity of the PITS, the correlations between the TSCYC subscale scores were examined between similar constructs on the PITS. The TSCYC RL correlation also was examined for the PITS Response Style Scale. Additionally, because the construct of FA emerged and did not theoretically correlate with any subscale on the TSCYC, the Pediatric Emotional Distress Scale’s Fearful subscale (Saylor et al., 1999) was used to establish concurrent validity. To be accepted as evidence of concurrent validity, the correlation coefficient between the two instruments needed to reach or exceed the minimum of r = .35 (Hamill et al., 1992). As shown in Table 3, all PITS subscales, the overall Trauma Composite (PITS-TC) Scale, and the Response Style subscale met these criteria for demonstrating concurrent validity.
Concurrent Validity of the Preschool Inventory of Trauma Symptoms and Other Measures of Trauma for Young Children.
Note. TSCYC PTS-I = Trauma Symptom Checklist for Young Children–Post-traumatic Stress Intrusion Scale; TSCYC PTS-AV = Trauma Symptom Checklist for Young Children–Post-traumatic Stress Avoidance Scale; TSCYC PTS-AR = Trauma Symptom Checklist for Young Children–Post-traumatic Stress Arousal Scale; PEDS-F = Pediatric Emotional Distress Scale’s Fearful subscale; TSCYC RL = Trauma Symptom Checklist for Young Children–Response Level; TSCYC PTS-TOT = Trauma Symptom Checklist for Young Children–Post-traumatic Stress Total; PITS-I = Preschool Inventory of Trauma Symptoms–Intrusion and Re-Experiencing; PITS-AVN = Preschool Inventory of Trauma Symptoms–Avoidance and Negative Cognition and Mood; PITS-ARH = Preschool Inventory of Trauma Symptoms–Arousal and Hyper-Reactivity; PITS-FA = Preschool Inventory of Trauma Symptoms–Fearful Attachment; PITS-RS = Preschool Inventory of Trauma Symptoms–Response Style; PITS-TC = Preschool Inventory of Trauma Symptoms–Trauma Composite score.
**p < .01.
Treatment Sensitivity
As an initial effort to determine whether the PITS-TC scores and subscale scores would be sensitive to change following participation in an in-home, trauma-informed treatment specifically designed for very young children in poverty (Fox, 2018), we conducted a separate pilot study with 93 children referred to our clinic. This treatment program has been analyzed for efficacy through four separate randomized controlled studies (Carassco & Fox, 2012; Fung & Fox, 2014; Harris, Fox, & Love, 2015; Love & Fox, 2019) with similar samples of children. We did not include a control group in this preliminary study. These children were referred to our clinic because it specialized in treating very young children (Fox et al., 2007) and were currently experiencing one or more trauma symptoms. The sample for this phase of the study was 93 diverse children (African American = 32.3%, Latino/a = 43.0%, European American/White = 9.7%, multiracial = 14.0%, and 1% identified as other with a mean age of 2.6 years (SD = 1.10); 70% were males. The average number of treatment sessions was 9 (SD = 2.87). Paired t tests were used to compare scores between intake and posttreatment scores for the four PITS subscales and TC score. The resulting p values were all significant and ranged between .03 and .001; the total effect sizes based on Cohen’s d were as follows: PITS-ARH subscale = .79, PITS-FA subscale = .45, PITS-I subscale = .39, PITS-AVN subscale = .25, and PITS TC Score = .71.
Discussion
Recognizing both the immediate impact and the long-ranging implications of potentially traumatic event(s) exposure in young children, a growing need has arisen to properly screen children who may need intervention services for trauma symptoms related to PTEs early in life (Miller-Graff et al., 2016), particularly for those living in poverty (Collins et al., 2010). This study developed a 34-item, caregiver report instrument to screen for trauma symptoms in very young children living in poverty. Factor analysis for this instrument, referred to in this article as the PITS, identified a four-factor model: (1) AHR, (2) FA, (3) I, and (4) AVN. Three of the PIT’s factors aligned with domains of the preschool subtype of PTSD in the DSM-V, and the fourth factor aligned with the literature regarding impairments in attachment following PTE exposure (e.g., Pickreign Stronach et al., 2011; Pynoos et al., 2009; Zeanah et al., 2004). Increased arousal, including irritability, aggressive behavior, and fussiness, has consistently been noted in young children exposed to trauma (Gigengack, van Meijel, Alisic, & Lindauer, 2015; Modrowski et al., 2013; Pynoos et al., 2009). In fact, symptoms of hyperarousal are among the most frequent symptoms reported in children with maladaptive response following PTE exposure (Gigengack et al., 2015). Consistent with the literature, the AHR factor yielded the strongest loading factor on the PITS. The PITS also assessed impairments in attachment, a domain rarely assessed by current instruments. Young children with PTE exposure display less secure and more disorganized styles of attachment (Pickreign Stronach et al., 2011; Pynoos et al., 2009; Zeanah et al., 2004). The quality of the parent–child relationship is also inversely related to PTE exposure and development of psychopathology (Fergusson, Boden, & Horwood, 2008; Milot, St-Laurent, Ethier, & Provost, 2010). Very young children relive and reexperience trauma differently than adults, frequently describing their trauma via story narrative or reenactment through play (Miller-Graff et al., 2016; Modrowski et al., 2013; Pynoos et al., 2009). Consistent with this literature, frequently endorsed symptoms of reexperiencing (e.g., via play) strongly loaded on the PITS-I factor. Avoidance symptoms less frequently endorsed than any other domain (e.g., Intrusion) among the children with PTE exposure (Scheeringa, Peebles, Cook, & Zeanah, 2001), likely do to the internal nature of the experience that is hard to capture from a behavioral standpoint (Pynoos et al., 2009). Numerous works cited the DSM-IV’s requirement of three avoidance symptoms as one of the largest hurdles for the accurate diagnosis of PTSD in young children (Gigengack et al., 2015; Scheeringa, Meyers, Putnam, & Zeanah, 2012; Scheeringa et al., 2001). The new criteria for “PTSD for Children 6 Years and Younger” in the DSM-5 also included “increased frequency of negative emotional states under Criterion C (Avoidance), replacing the DSM-IV’s symptoms of “emotional constriction and estrangement from others” (APA, 2013; Pynoos et al., 2009). In alignment with DSM-5 criteria, items associated with AVN, loaded on one factor, further bolstering the clinical and diagnostic utility of the PITS. The composite score provided a quick means to flag children at risk for maladaptive responses following PTE exposure. All subscales and the composite measure of the PITS correlated strongly with preestablished measures of trauma such as the TSCYC. The final PITS item pool had a Flesh–Kincaid reading grade level of 3.7, making it simple enough for most parents to complete independently, further reducing time and expense. The preliminary psychometric properties including reliability and validity measures of the PITS were found to be within acceptable ranges. The TC cutoff score of 51 was found to have strong sensitivity and specificity. Positively, all four clinical subscales of the PITS and the TC score were sensitive to changes expected from an evidence-based, trauma-informed treatment program with small to large effect sizes.
Limitations
In the process of developing any new assessment instrument, several limitations were anticipated, and others became apparent following the analyses of the PITS data. Although the PITS fulfills an important need for effectively screening very young children from diverse backgrounds, conducting future research in populations of different socioeconomic backgrounds may improve the utility of the scale. Due to the early age of the children in the sample, child caretakers reported these data. Because caregiver-report data can be influenced by bias, the measure included a Response Style Scale; however, the inclusion of other clinical measures (e.g., trauma-focused clinical interview, direct observation) along with this data would strengthen the existing measure. Additionally, because there are limited psychometrically sound measures that assess trauma symptoms for 1- to 2-year-old children, the ROC analysis output variable (TSCYC-PTS-TOT) used the norms for the 3- to 4-year-olds to determine clinically significant trauma symptoms. The 5-year-old sample also was small. Although no significant differences emerged among the different age-groups for the PITS-TC, caution should be used when interpreting scores for these age-groups. Additional measures of reliability, such as interrater reliability and test–retest reliability, would strengthen existing psychometric information. The current factor structure should be confirmed with another sample. The authors encourage such research and have provided the instrument in the Appendix for use by qualified professionals, without cost. Although no significant differences emerged among the different age-groups for the PITS-TC, additional caution should be used when interpreting cut scores for this age-group as they are based on relatively stringent but arbitrary statistics (1.5 standard deviations above mean), particularly for the youngest children (1–2 years of age). The treatment sensitivity of the PITS reported in this study should be considered as only a preliminary finding as no control group was included. However, four previous randomized controlled studies of this evidence-based, trauma-informed treatment program did show that it significantly reduced trauma symptoms compared with a control group using more established trauma measures (Carassco & Fox, 2012; Fung & Fox 2014; Harris et al., 2015; Love & Fox, 2019). Given the paucity of sound screening measures combined with the growing recognition of the critical need to identify, assess, and treat trauma symptoms in very young children (nctsn.org), having a relatively short screening measure with at least the potential to accurately assess trauma symptoms following trauma-informed treatment programs will be a welcome addition to the various welfare and child protective service agencies who routinely serve the population sampled in this study. Finally, although a wide variety of PTEs were included in this sample, it would be important for future research to examine how trauma symptoms vary among different PTEs.
Clinical Implications
From a clinical standpoint, the PITS provides mental health clinicians and other medical professionals with an efficient means to assess whether concerns for trauma were present and to determine the areas of greatest impact (e.g., avoidance, arousal). Additionally, the presence of the attachment subscale highlighted important developmental factors and potential treatment goals, given the correlation between the quality of the parent–child relationship and resilience following PTE exposure (Fergusson et al., 2008). We recommend if trauma symptoms are found on the PITS, clinicians should determine contributing factors of these symptoms through a combination of methods such as a more trauma-focused clinical interview to identify potential triggers, the use of available scales that measure potentially traumatic events such as the TESI-PRR (Ghosh-Ippen et al., 2002), or other assessment measures to assist in accurately conceptualizing the trauma symptoms and providing appropriate treatment. Included in the Appendix to this article is a user-friendly copy of the PITS including scoring and interpretation information.
Footnotes
Appendix
Child’s Name:__________________ Parent’s Name:________________ Date:______________ Therapist:________________
Authors’ Note
Sara H. Bollens is now affiliated with Regional Center of Orange County, Santa Ana, CA, USA
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported in part by grants from Charles D. Jacobus Family Foundation, Helen Bader Foundation, Hearst Foundation, Roger and Cindy Schaus Family, and the Substance Abuse and Mental Health Administration (Grant Number: 1U79SM063056-1).
