What’s NewThis is the first such study to rigorously evaluate a model for enhancing pediatric primary care by addressing prevalent risk factors for child maltreatment in a low-risk population. The findings offer encouraging support that the SEEK model is a practical and promising approach.
Child abuse and neglect are common problems in the United States, with approximately 794 000 or 10.6 per 1000 children identified as maltreated annually.
1U.S. Department of Health and Human Services
Administration on Children Youth and Families. Child Maltreatment 2009.
Maltreated children are at risk of many short- and long-term consequences, including further injuries, developmental problems, and psychological, learning and conduct disorders.
2- Hussey J.
- Chang J.
- Kotch J.
Child maltreatment in the United States: prevalence, risk factors, and adolescent health consequences.
, 3- Stirling J.J.
- Amaya-Jackson L.A.
Understanding the behavioral and emotional consequences of child abuse.
, 4- Dubowitz H.
- Papas M.A.
- Black M.M.
- et al.
Child neglect: outcomes in high-risk urban preschoolers.
, 5- Lansford J.E.
- Dodge K.A.
- Pettit G.S.
- et al.
A 12-year prospective study of the long-term effects of early child physical maltreatment on psychological, behavioral, and academic problems in adolescence.
, 6- Dube S.R.
- Miller J.W.
- Brown D.W.
- et al.
Adverse childhood experiences and the association with ever using alcohol and initiating alcohol use during adolescence.
, 7- Hillis S.D.
- Anda R.F.
- Dube S.R.
- et al.
The association between adverse childhood experiences and adolescent pregnancy, long-term psychosocial consequences, and fetal death.
The effects of child maltreatment may persist into adulthood, manifesting in an array of physical, social, and mental health problems.
8- Dube S.R.
- Anda R.F.
- Felitti V.J.
- et al.
Childhood abuse, household dysfunction, and the risk of attempted suicide throughout the life span: findings from the Adverse Childhood Experiences Study.
, 9- Malinosky-Rummell R.
- Hansen D.J.
Long-term consequences of childhood physical abuse.
, 10Child abuse, neglect, and adult behavior: research design and findings on criminality, violence, and child abuse.
, 11- Dube S.R.
- Felitti V.J.
- Dong M.
- et al.
The impact of adverse childhood experiences on health problems: Evidence from four birth cohorts dating back to 1900.
, 12- Felitti V.J.
- Anda R.F.
- Nordenberg D.
- et al.
Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study.
Individual and societal costs of child maltreatment are enormous; a recent estimate was $95 billion per year related to medical and mental health treatment, lost productivity, and crime.
13Total Estimated Cost of Child Abuse and Neglect in the United States.
Results
As expected, nearly 95% of
SEEK mothers had brought their child for at least one well child visit before recruitment and were therefore exposed to the model before the initial assessment; 70% had more visits. Thus, responses to the initial survey probably reflect early effects of
SEEK rather than a baseline. The mean CTSPC score for Psychological Aggression and Minor Physical Assault was greater in the control group than the
SEEK group at each time point (
Table 3). In multivariable analyses in which we controlled for potential confounders (mother’s education, age, marital status, family income, and child’s ethnicity) and accounted for the clustering of observations within participants and practices,
SEEK mothers reported less frequent Psychological Aggression (effect size = −0.16,
P = .006) and fewer Minor Physical Assaults (effect size = −0.16,
P = .019) initially and 12 months later (Psychological Aggression, effect size = −0.12,
P = .047; Minor Physical Assault, effect size = −0.14,
P = .043;
Table 3). Findings at 6 months were in the same direction, albeit not statistically significant. The frequencies of reported Severe and Very Severe Physical Assault were extremely low (<1% of the sample) and were excluded from the analyses. The intraclass correlations within practice estimated from these models were very low (0 for Psychological Aggression, 0.01 for Minor Physical Assault), suggesting that after adjustment for socioeconomic differences, there was no association between practice and these outcomes.
Table 3Parent−Child Conflict Tactics Scales (CTSPC) Scores Comparing SEEK∗SEEK = Safe Environment for Every Kid; CI = confidence interval.
and Control Maternal Parenting Behaviors Initially and at 6 and 12 Months To assess whether differences between the study groups in CTSPC initial outcomes might have been the result of uncontrolled differences between the groups (rather than an early effect of SEEK), we compared the CTSPC scores on the initial assessment for the 547 families exposed to SEEK before the initial assessment to those of 28 families in the SEEK group who were unexposed to SEEK at the time of the initial assessment. “Unexposed” refers to those who had not had a checkup in a SEEK practice or had not completed a PSQ before the initial assessment. Exposed and nonexposed SEEK families differed only on child age; exposed children averaged 15 months younger. We thus controlled for child’s age in the analyses. Within the SEEK group, nonexposed mothers reported more Psychological Aggression (Mean score =14.0, SD =11.6) initially than did exposed mothers (M = 5.8, SD = 9.8; P = .03). Nonexposed mothers reported more incidents of Minor Physical Assault (M = 5.2, SD = 7.5) than did exposed mothers (M = 2.5, SD = 6.7), although this difference was not statistically significant (P = .201). Comparing SEEK nonexposed mothers and controls (n = 28 and 48, respectively), we found no differences in terms of reported Minor Physical Assault (B = −.03, P = .915) and Psychological Aggression (B = 1.69, P = .566). These findings supported using the initial data as an early outcome.
Before SEEK, 46 (8%) intervention families and 25 (5%) controls had one or more problems related to possible abuse or neglect documented in the medical records. During the project, the proportions were 85 (14%) and 45 (9%), respectively. This difference was not statistically significant (OR = 1.14, P = .76) after adjusting for the random effect of practice and the number of pre-study problems.
There were relatively few child protective service reports (
Table 4). Most reports were for neglect (50%) or physical abuse (32%). After taking into consideration pre-
SEEK differences, we found no statistically significant difference between groups during
SEEK (P = .69).
Table 4Number and Proportion of SEEK∗SEEK = Safe Environment for Every Kid; CPS = Child Protective Services.
and Control Families in the Four CPS Reporting Categories P = .69 for difference between groups with respect to decline in CPS reports during SEEK, on the basis of the Fisher exact test.
The time HPs in SEEK and control practices spent discussing psychosocial concerns during the study was nearly identical (median, 37.0 vs 37.5 seconds; interquartile range, 59.5 vs 60.0, respectively). There was also no significant difference in the average total time spent on visits (SEEK 17.5 minutes vs controls 16.3 minutes, P = .18).
Discussion
These findings provide further evidence that the
SEEK model of enhanced pediatric primary care may help prevent maltreatment. It is especially important that this was in a relatively low-risk population.
SEEK mothers reported less Psychological Aggression and fewer Minor Physical Assaults at the initial and 12-month assessments, with moderate effect sizes.
23Statistical Power Analysis for the Behavioral Sciences.
For example, at 12 months, the assault rate was 0.2 SD lower in the
SEEK group compared with controls. Although many of these instances may not meet legal definitions of maltreatment, ample evidence indicates that experiences such as hitting children (ie, corporal punishment) jeopardize their development.
24Relationships between parents’ use of corporal punishment and their children's endorsement of spanking and hitting other children.
, 25On hitting children: a review of corporal punishment in the United States.
, 26- Taylor C.
- Manganello J.
- Lee S.
- et al.
Mothers’ spanking of 3-year-old children and subsequent risk of children’s aggressive behavior.
, 27- Zolotor A.
- Theodore A.
- Chang J.
- et al.
Speak softly—and forget the stick. Corporal punishment and child physical abuse.
, 28- Straus M.
- Sugarman D.
- Giles-Sims J.
Spanking by parents and subsequent antisocial behavior of children.
, 29- Miller-Perrin C.L.
- Perrin R.D.
- Kocur J.L.
Parental physical and psychological aggression: psychological symptoms in young adults.
Psychological maltreatment is defined by the American Academy of Pediatrics as a repeated pattern of damaging interactions between caregiver and child
30- Hart S.N.
- Brassard M.R.
- Davidson H.A.
- et al.
Psychological maltreatment.
; it may be the most damaging of all forms of maltreatment, even though it seldom leads to child protective service involvement.
31- Kairys S.W.
- Johnson C.F.
- Committee on Child Abuse and Neglect
The psychological maltreatment of children—technical report.
It is clear that protective services reports reflect only a small fraction of the maltreatment children experience; they are guided by state laws that generally focus on relatively egregious circumstances. We suggest that the definition of maltreatment be based on scientific evidence of what harms children. Psychological Aggression and Minor Physical Assault were prevalent and potentially damaging; there is a need to reduce these experiences that at a minimum constitute harsh punishment, and may indicate maltreatment.
As expected, child maltreatment was infrequent—when measured by child protective service reports or documentation in medical records. This poses a challenge for evaluating efforts to prevent maltreatment With a relatively low base rate in all but the greatest-risk populations it is very difficult to show decreased rates of reported or documented maltreatment. Direct observation is naturally very difficult, making researchers mostly reliant on self-report measures. These too have their limitations, especially when ascertaining socially undesirable information.
The previous
SEEK study was conducted in a very high-risk urban, mostly African-American population.
SEEK reduced maltreatment—measured by self-report, medical records and child protective service reports.
15- Dubowitz H.
- Feigelman S.
- Lane W.
- et al.
Pediatric primary care to help prevent child maltreatment: the Safe Environment for Every Kid (SEEK) Model.
Findings in the current study involving mostly middle-income white families were statistically significant, but not as strong.
15- Dubowitz H.
- Feigelman S.
- Lane W.
- et al.
Pediatric primary care to help prevent child maltreatment: the Safe Environment for Every Kid (SEEK) Model.
This raises the question of whether the model should be used only in high-risk populations. However, even modest reductions in potentially damaging experiences can have valuable, far-reaching benefits at a population level; the present sample likely represents many American families. It is noteworthy that even in this relatively low-risk population, whereas some risk factors were reported infrequently (eg, intimate partner violence), others were quite prevalent (eg, alcohol abuse: 8%). Furthermore, even if a significant reduction in child protective service reports is difficult to demonstrate, helping address prevalent psychosocial problems such as maternal depression or alcohol abuse should strengthen families, support parents, and improve children’s health, development and safety.
32- Olson A.
- Kemper K.
- Kelleher K.
- et al.
Primary care pediatricians’ roles and perceived responsibilities in the identification and management of maternal depression.
, 33- Weissman M.M.
- Pilowsky D.J.
- Wickramaratne P.J.
- et al.
Remissions in maternal depression and child psychopathology: a STAR*D-child report.
SEEK involves a modest yet substantive change in current practice. For example, screening for parental depression seldom occurred prior to the study or in control practices (data not shown).
34Practice improvement: child healthcare quality and Bright Futures.
The study required a commitment to attend training sessions and complete periodic questionnaires. It is very encouraging that 75% of practices agreed to participate, as did all the HPs in those practices. This reflects substantial interest among pediatricians and nurse practitioners to respond to the psychosocial problems facing many families.
31- Kairys S.W.
- Johnson C.F.
- Committee on Child Abuse and Neglect
The psychological maltreatment of children—technical report.
, 35Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents.
, 36- Hagan J.F.
- Shaw J.
- Duncan P.
Bright Futures Guidelines for Health Supervision of Infants, Children, and Adolescents.
, 37- Dubowitz H.
- Lane W.G.
- Semiatin J.N.
- et al.
The Safe Environment for Every Kid (SEEK) model: impact on pediatric primary care professionals.
With such interest, changes to pediatric primary care practice are clearly possible, particularly since implementing
SEEK was mostly straightforward. There are naturally challenges. Finding time for training is not easy, nor is changing practice behavior. Importantly,
SEEK HPs showed improvement in their comfort level and perceived competence addressing the targeted problems, sustained 36 months after the initial training.
34Practice improvement: child healthcare quality and Bright Futures.
In developing
SEEK, we were very practical, recognizing cost and time constraints in a busy practice. Having assistance from a social worker seemed important, complementing HPs efforts to address identified problems. To limit costs, the social worker divided her time among the 7
SEEK practices, while being available to
SEEK HPs and parents during regular hours. Surprisingly, despite excellent working relationships, she was underused, and, much of her work was by phone. It may be possible to lower program costs by having a social worker cover more practices and provide assistance only by phone. Data support the effectiveness of such psychosocial phone interventions.
38Information in practice: Telephone consultations.
, 39- Simon G.
- VonKorff M.
- Rutter C.
- et al.
Randomized trial of monitoring, feedback, and management of care by telephone to improve treatment of depression in primary care.
Alternatively, it is possible that the HP or someone else in the practice could perform the key function of facilitating appropriate referrals.
Despite AAP recommendations for screening,
40- Committee on Child Abuse and Neglect
The role of the pediatrician in recognizing and intervening on behalf of abused women.
, 41- Thackeray J.
- Hibbard R.
- Dowd M.
Intimate partner violence: the role of the pediatrician.
pediatricians often raise concerns about the time required to address identified problems. Using waiting room time takes good advantage of this opportunity, that is, having parents complete the PSQ while they are waiting should save time during the visit and enable HPs to efficiently focus on identified problems. Indeed, we found that
SEEK did not add time to visits. A comparable study in which primary care pediatricians were trained to address children’s behavior problems did not find significant increases for any of the four levels of office visits or for health maintenance visits.
42- Gadomski A.
- Wissow L.S.
- Slade E.
- et al.
Training clinicians in mental health communication skills: impact on primary care utilization.
How may
SEEK have influenced the outcomes? We previously reported that
SEEK HPs had significant and sustained improvements in attitudes and behavior regarding addressing the targeted psychosocial problems compared to controls.
34Practice improvement: child healthcare quality and Bright Futures.
For example, screening for depression occurred far more frequently in
SEEK practices, and when problems were identified, some action was almost always documented. At 12 months,
SEEK mothers reported significantly greater satisfaction with their parenting (
P = .02), with a trend (
P = .06) in same direction at 6 months. There were also encouraging findings regarding intimate partner violence. At 6 months,
SEEK mothers reported fewer physical assaults by them toward their partners (odds ratio [OR] 0.44,
P = .049) and at 12 months by their partners toward them (OR 0.47,
P = .045) compared with controls. These findings may partially explain the apparent effectiveness of
SEEK. There remains a need to better elucidate how
SEEK may effect change.
There are several strengths to this study. We used the conservative intention to treat approach; some families had relatively little exposure to SEEK. We had excellent retention of practices, HPs and participants during a 30-month period. We were unsure whether HPs might resent the time required or if parents might find the PSQ intrusive. However, there were very few complaints and SEEK was well accepted. We used rigorous statistical approaches to minimize potential limitations, such as carefully controlling for group differences and potential confounders.
Study Limitations
The study also has several limitations. We could not collect baseline data; the 18 months needed to recruit the sample precluded waiting to implement SEEK. Thus, most participants were exposed to the model before the initial survey. However, nonexposed SEEK mothers reported more Psychological Aggression than those exposed and nonexposed SEEK mothers and controls did not differ in terms of this outcome or Minor Physical Assault. These findings support the early influence of the model and the use of the initial assessment as early outcome data. Given more power, the trend for Minor Physical Assault would probably also have been significant. It is also possible, however, that the initial findings reflect baseline differences between groups. The power was also limited for the outcomes that occurred rarely, such as reports to child protective services. Thus, the lack of a significant finding does not rule out possible impact of SEEK.
Randomization was not entirely successful. We needed to add 2 control practices to have a similar number of HPs across groups. Also, there were some socioeconomic differences between the groups. We adjusted for differences between the groups with respect to measured variables (income, education, marital status, ethnicity, age) by using a regression model. The very low intraclass coefficients indicate no association between practice and the outcomes. In addition, we adjusted for unmeasured differences between the practices by including a random effect for practice in the model. The greater adversity in the SEEK group, however, makes the findings more remarkable; differences between the groups favored the null hypothesis.
Future Directions
There is great interest to find promising strategies to help prevent child maltreatment. After 2 rigorous studies, the
SEEK model appears to be one, and, by addressing prevalent family problems it may also enhance children’s health, development, and safety. This fits well with the mission of pediatrics and Bright Futures.
35Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents.
, 36- Hagan J.F.
- Shaw J.
- Duncan P.
Bright Futures Guidelines for Health Supervision of Infants, Children, and Adolescents.
, 37- Dubowitz H.
- Lane W.G.
- Semiatin J.N.
- et al.
The Safe Environment for Every Kid (SEEK) model: impact on pediatric primary care professionals.
There has been considerable interest in the United States in replicating this model. Some may argue the evidence is not enough to justify going to scale. Others may think it is more than adequate and may also point to many areas of practice based on scant evidence. Importantly, the model does not appear to have negative outcomes and should substantially enhance pediatric primary care, especially as it did not involve more professional time, and there is evidence that the program may in fact be cost saving.
43Lane WG, Frick K, Dubowitz H, et al. Cost-Effectiveness analysis of the SEEK (A Safe Environment for Every Kid) child maltreatment prevention program. American Public Health Association 139th Annual Meeting and Exposition. Washington, DC, November 1, 2011.
There are many practical issues to consider. How do we encourage those providing pediatric primary care to adopt this model? Possible approaches include developing online training and ongoing support and technical assistance. Another issue concerns the social worker in the model, a challenge for many given the finances of pediatric primary care. As suggested previously, the facilitation of referrals could probably be accomplished by HPs and/or office staff. Such questions raise the issue of fidelity to the model tested. In developing SEEK we recognized the heterogeneity among HPs and practices and deliberately sought to make the model somewhat flexible. Nevertheless, some core components do appear important: preparing HPs to help address the targeted problems, the PSQ (or similar tool) to screen systematically, ability to link families needing help to community resources, and the availability of necessary resources. In sum, it seems reasonable to cautiously disseminate and replicate the SEEK model, without overpromising, while continuing to assess its effectiveness.