Parenting Interventions in Pediatric Primary Care: A Systematic Review

In this article, we review the evidence for implementation of parenting programs in pediatric primary care, which are effective at preventing and treating children’s behavioral health problems.

Abstract

CONTEXT:

More than 4 decades of research indicate that parenting interventions are effective at preventing and treating mental, emotional, and behavioral disorders in children and adolescents. Pediatric primary care is a viable setting for delivery of these interventions.

OBJECTIVE:

Previous meta-analyses have shown that behavioral interventions in primary care can improve clinical outcomes, but few reviews have been focused specifically on the implementation of parenting interventions in primary care. We aimed to fill this gap.

DATA SOURCES:

We reviewed 6532 unique peer-reviewed articles published in PubMed, the Cumulative Index to Nursing and Allied Health Literature, and PsycInfo.

STUDY SELECTION:

Articles were included if at least part of the intervention was delivered in or through primary care; parenting was targeted; and child-specific mental, emotional, and behavioral health outcomes were reported.

DATA EXTRACTION:

Articles were reviewed in Covidence by 2 trained coders, with a third coder arbitrating discrepancies.

RESULTS:

In our review of 40 studies, most studies were coded as a primary. Few researchers collected implementation outcomes, particularly those at the service delivery system level.

LIMITATIONS:

Including only published articles could have resulted in underrepresentation of implementation-related data.

CONCLUSIONS:

Parenting interventions delivered and implemented with fidelity in pediatric primary care could result in positive and equitable impacts on mental, emotional, and behavioral health outcomes for both parents and their children. Future research on the implementation strategies that can support adoption and sustained delivery of parenting interventions in primary care is needed if the field is to achieve population-level impact.

Parenting interventions are considered best practice for preventing the most common mental and behavioral health disorders of childhood and adolescence, such as depression, anxiety, and externalizing disorders,^1–8 while also promoting positive development and adaptation.^9,10 However, parenting programs are not implemented widely in primary care, despite intuitive appeal and fit within this service context.⁹ In this review, we delve into the state of the literature on implementing parenting interventions in primary care to inform directions for taking these programs to scale.

The 2017 to 2018 National Survey of Children’s Health revealed that ∼1 in 5 children in the United States aged 3 to 17 years (21.9%) had a diagnosed mental, behavioral, or developmental disorder.¹¹ Among youth with concerns severe enough to warrant intervention, <50% received treatment in the previous year, and significant disparities existed for ethnic minority youth, low-income families, and the uninsured.¹² Four decades of research provide evidence that these interventions improve key parenting domains, such as positive behavior support, monitoring and limit setting, and family relationship building,^13,14 which in turn prevent children’s emotional (eg, anxiety, depression)⁵ and behavioral disorders (eg, substance use, conduct problems),¹⁵ can mitigate the effects of socioeconomic adversity,^16,17 and improve school readiness, positive peer relationships, and well-being.¹⁸ Parenting intervention effects are robust and often long- lasting across multiple health and developmental outcomes^19,20 and have been shown to be acceptable and effective with racially and ethnically diverse families.^21,22 Despite the strong evidence base, parenting interventions have not been taken to scale.^14,23

Methods

Achieving Greater Reach of Parenting Programs

Pediatric primary care provides an opportune setting for parenting interventions aimed at improving children’s emotional and behavioral health.²⁴ Beginning in infancy, most children in the United States receive primary health care. In the 2018 National Health Interview Survey, it was estimated that more than three-quarters of all children <18 years have had contact with a primary health care professional in the past 6 months.²⁵ Thereby, primary health care is a compelling service context as it creates a tight net for reaching families in need of intervention. With the realization that children are nearly always accompanied by their parents to well-child visits, recommendations regarding child behavior are often given to parents in this context, engendering low stigma.^9,26 In addition to child health, the health of caregivers is a growing focus within pediatric primary care. For example, postpartum depression screening, a known risk factor for adverse child outcomes across the life course, is now covered under children’s health insurance when delivered at well-child visits.²⁷ This shift to supporting caregiver well-being provides an opening for the delivery of parenting programs in this context.

Parenting interventions can be delivered by using a variety of models appropriate for pediatric settings. Kolko and Perrin²³ describe 5 evidence-based models of effective delivery of behavioral health services in pediatric primary care: (1) coordination with an external behavioral health clinician, (2) consultation between behavioral health and the primary care clinician, (3) training the primary care clinician in behavioral health interventions, (4) on-site intervention by behavioral health (often called co-located care), and (5) on-site collaborative and integrated care. With its success with adults, particularly for depression, collaborative care and medical home models for mental, emotional, and behavioral health in children have garnered recent attention. However, results are equivocal, and implementation is thus far limited.^28–30 It is unclear which of these care delivery models for parenting interventions are more common in primary care and which result in better implementation outcomes, which are defined as the effects of deliberate and purposive actions to implement new programs, practices, and services³¹ and include such common metrics as acceptability, adoption, appropriateness, cost, feasibility, fidelity, penetration and reach, and sustainability.^31,32

There has been limited implementation of empirically supported interventions that prevent mental and behavioral health disorders in pediatric primary care.¹⁵ This lack of uptake suggests a substantial infrastructure deficit as well as an area with tremendous growth potential.^9,15,24,33 Primary care–based interventions have the potential to address health inequity by reaching families of socioeconomic and racial and ethnic diversity and marginalized populations who often have difficulty accessing specialized mental health care in other contexts or who tend to avoid specialty mental health care because of stigma.³⁴ In the 2017 National Survey of Drug Use and Health, among adolescents aged 12 to 17 years who received behavioral health services, only 3.3% of youth reported receipt in primary care compared with 13.3% and 13.6% of services received in education or outpatient mental health settings, respectively.³⁵

Challenges to Implementation at Scale in Primary Care

Before parenting interventions can be implemented at scale in primary care, however, 4 major challenges need to be addressed. Such interventions must be feasible to deliver; integrated across and within existing primary care and child health care systems; culturally responsive²¹ to socioeconomically disadvantaged, historically marginalized, and underserved populations; and adaptable to different organizational structures and cultures.³⁶ These challenges underscore the tension between health care infrastructure and the characteristics of interventions that, although being considered evidence based, may not be feasible for implementation in systems that are often resource scarce and can have incongruent priorities and funding sources.³⁷ Adaptation of programs for delivery in the primary care context can increase feasibility but might not overcome all barriers to adoption and sustainment.³⁸ Many examples exist of adapting evidence-based parenting interventions for racial and ethnic minority families to improve cultural responsivity,³⁹ but few examples exist of adapting these programs for the primary care service context.³⁸

Effectiveness of Behavioral Health Interventions in Pediatric Primary Care

In recent meta-analytic studies, researchers have found that the effects of prevention programs targeting children and adolescents are significant when delivered in coordination with primary care. Asarnow et al⁴⁰ reviewed controlled studies of treatment and prevention programs for child and adolescent mental health delivered as integrated behavioral health care compared with primary care services as usual. In the results from 31 included studies (n = 12 129 participants), a significant overall effect was indicated (d = 0.32; 95% confidence interval [CI], 0.2–0.44; P < .001). Effects were smaller but still significant for prevention programs compared with treatment programs that targeted youth with elevated symptoms. However, only 9 prevention trials met the authors’⁴⁰ criteria, which included a rigorous randomized trial design. Relatedly, in a meta-analysis, Shah et al⁴¹ examined the effects of primary care–based interventions for the promotion of positive parenting for children enrolled at <3 years of age. Thirteen studies (n = 3807) were identified. Results were promising, with effects on parent-child interactions (d = 0.29; 95% CI, 0.06–0.52; P < .001) and participation in cognitively stimulating activities (d = 0.34; 95% CI, 0.03–0.54; P < .001). Authors of both these meta-analyses noted the need to better understand implementation of these effective practices, but neither group explicitly evaluated included studies for characteristics or outcomes related to intervention implementation.

A number of systematic reviews have also emerged in the literature recently, highlighting interest in this area. Brown et al⁴² found 44 studies of preventive interventions in primary care for children ages 0 to 5 years. Many had a parenting outcome but were not necessarily a formal parenting program. Rojas et al⁴³ identified 19 preventive interventions for child and adolescent mental and behavioral health problems, 10 of which were considered parenting programs. Studies were focused mostly on school-aged children (n = 6) and adolescents (n = 8) compared with earlier ages (infancy: n = 2; preschool age: n = 3). They also noted that there was a low degree of involvement from the primary care practices themselves, meaning that primary care was primarily used as a recruitment source for services provided in other settings rather than as a context for integrated delivery. In these reviews, it is indicated that parenting interventions can be effective when delivered in coordination with primary care, but little attention was given to sustainable implementation, possibly contributing to the low uptake in regular practice.

This Study

We report the findings of a systematic review of peer-reviewed research on the use of parenting programs to improve child and adolescent mental, emotional, and behavioral health outcomes that were either delivered in the context of pediatric primary care or coordinated with primary care through systematic referrals to services provided in other settings. The findings of this review are placed in the framework of the intervention research translation pipeline,⁴⁴ which describes the process of taking interventions to scale in a series of stages from preintervention, to testing interventions using efficacy and effectiveness designs, to dissemination and implementation research in real-world service delivery systems. With regard to parenting interventions, the field has provided sufficient evidence of efficacy and effectiveness already, so in our review, we follow the guidance of Glasgow et al⁴⁵ to focus on data reported to inform implementation. As such, the aims are to provide (1) a comprehensive view of the state-of-the-knowledge concerning implementation of parenting interventions in conjunction with the primary health care system and (2) evidence-based guidance to inform the development of effective strategies to enhance the adoption and sustainment of parenting programs in pediatric primary care settings.⁴⁶

Search

A comprehensive search of studies published since 1990 was undertaken in March 2018 in Google Scholar, PubMed, PsycInfo, and the Cumulative Index to Nursing and Allied Health Literature. Search terms included those related to child demographics, primary care (including family medicine, pediatric care, community health centers, federally qualified health centers), parental involvement in the intervention, and a range of mental, emotional, and behavioral health outcomes (see Supplemental Information for complete search strategy). A total of 6859 articles were returned, with 327 duplicates removed via Covidence⁴⁷ review management software and manual review (6532 unique articles). After title and initial full-text review by 2 trained coders (G.H.C., L.M.R.) and arbitration of discrepancies by a third (J.D.S.), 676 articles were deemed eligible for full-text review. A total of 636 were excluded for various reasons presented in the preferred reporting items for systematic reviews and meta-analyses flow diagram (Fig 1), the most common of which were the intervention not occurring in primary care (n = 381) and either no child outcome was reported (n = 41) or the child outcome reported was not a mental, emotional, or behavioral health outcome (n = 74). The final sample included 40 articles (references for included articles are in Supplemental Information). Data extraction was completed by using a shared Excel template. Ratings of trial design quality were completed by 2 authors (J.D.S., G.H.C.) using an established 8-point scale: (1) aims stated clearly, (2) “randomized controlled trial” or a trial with a “comparable comparison group,” (3) intervention described sufficiently to be replicable, (4) number of recruited subjects provided, (5) preintervention data provided, (6) level of attrition discussed, (7) results of all measured outcomes discussed, and (8) postintervention results provided for all intervention groups.⁴⁸ This study did not require institutional review board approval because it was not considered human subjects research.

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Inclusion and Exclusion

We included studies published in the English language of interventions that were targeting youth (birth to 18 years); explicitly targeting and reporting intervention effects on a child-level mental, emotional, and behavioral health outcome; and involving a distinct parental component to the intervention (defined as having at least one parent or caregiver being clearly identified as an intervention recipient, either individually or in a group setting; merely having a parent present while the child received an intervention or having the parent complete a questionnaire was not sufficient). Trials had to be quasi-experimental or fully randomized, have an appropriate comparison condition, and describe the location(s) at which the intervention was delivered. Multiple articles in which researchers reported on the same study were included if unique analyses were conducted (eg, mediation or moderation versus overall effect) or subgroups were assessed. Interventions that targeted parent behavior (eg, breastfeeding frequency) without a child mental, emotional, or behavioral target; targeted disease management behavior alone (eg, medication management); or did not include comparable pre- and postmeasures were excluded. Interventions in which teenage (<18 years old) mothers were targeted were excluded. Non–peer-reviewed publications, such as dissertations and conference proceedings, were excluded.

Results

Child Participants

The interventions predominantly targeted infants (∼2 weeks) to preadolescents (12.9 years). Among the 40 studies, using the Ages and Stages of Child Development taxonomy,⁴⁹ 10 targeted babies, infants, and toddlers (birth to 3 years); 5 targeted toddlers to preschoolers (3–6 years); 7 targeted preschoolers only (5–6 years); 1 targeted preschool to school age (6–12 years); 1 targeted preschool to adolescent (13–18 years); 6 targeted school age only; 4 targeted school age and adolescent; and 2 solely targeted adolescents (Table 1).

TABLE 1.

Study Characteristics

Source, Year	Sample Size	Developmental Stage(s)	Delivery Modality	Target Condition of Intervention	Clinician Type	Primary Care Physician Involvement	Intervention Setting	Type of Integration	Comparison Condition	Design Quality Rating
Bauer et al,50 2015	29 families	School age (6 y) to adolescent (18 y) (mean = 10.4 y)	Parent and child separate	Conduct or behavioral disorder	Physician	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Services as usual	8
Barlow et al,51 2007	131 women	Infant (recruited at birth)	Parent and child together	Other	Other	Referral	Home	External mental health clinician	Services as usual	8
Bayer et al,52 2010	733 mothers	Infant (6–7 mo) (mean = 7.0 mo)	Parent group	Conduct or behavioral disorder	Physician	Delivery, full intervention	Primary care	Colocation, colocated care, on-site intervention	Services as usual	8
Berkovits et al,53 2010	30 children	Preschool (3–6 y) (mean = 4.3 y)	Parent group	Conduct or behavioral disorder	Behavioral health specialist	Referral	Primary care	External mental health clinician	Active	8
Borowsky et al,54 2004	224 children	School age (7 y) to adolescent (15 y) (mean = 11.2 y)	Individual parent	Conduct or behavioral disorder	Behavioral health specialist, physician	Referral	Phone	Consultation with the primary care clinician	Services as usual	8
Chang et al,55 2015	501 mother–child pairs	Infant (6–8 wk) (mean = 1.67 mo)	Parent group	Other	Nurse	Delivery, partial intervention	Primary care	Integrated care or on-site collaborative care	Services as usual	7
Chavira et al,56 2014	48 parents	School age (8 y) to adolescent (13 y) (mean = 9.8 y)	Parent and child together	Anxiety	Physician	Referral	Primary care	Colocation, colocated care, on-site intervention	Active	8
Guilamo-Ramos et al,57 2011	264 mother–adolescent dyads	School age (11 y) to adolescent (14 y) (mean = 12.9 y)	Individual parent	Other	Physician, social worker	Referral	Primary care	Colocation, colocated care, on-site intervention	Services as usual	8
Hiscock et al,58 2008	328 mothers	School age (mean = 7.3 y)	Individual parent	Parent mental health and stress	Nurse	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	8
Hiscock et al,59 2007	328 mothers	Infant (mean = 7.3 mo)	Individual parent	Parent mental health and stress	Nurse	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	8
Hiscock et al,60 2008	733 mothers	Infant (6–7 mo) (mean = 7.0 mo)	Parent group	Conduct or behavioral disorder	Nurse, group facilitation expert	Delivery, full intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	7
Iloabachie et al,61 2011	83 adolescents	Adolescent (mean not reported)	Individual parent	Depression	Physician	Delivery, partial intervention	Primary care	Training the primary care clinician in mental health skills; referral	Active	8
Johnston et al,62 2006	439 mothers	Toddler (mean = 2.0 y)	Parent and child together	Conduct or behavioral disorder	Other (Head Start specialist)	Delivery, full intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	8
Kjøbli et al,63 2012	216 children	School age (mean = 7.4 y)	Individual parent	Parenting skills	Other (train primary care clinicians, staff in manualized intervention)	Delivery, partial intervention	Primary care	Training the primary care clinician in mental health skills	Active	8
Kolko et al,64 2011	163 children	School age (mean = 8.1y)	Parent and child together	Conduct or behavioral disorder	Nurse	Delivery, full intervention	Primary care	Integrated care, on-site collaborative care	Active (enhanced usual care)	8
Kolko et al,65 2014	321 caregivers	School age (mean = 8.0 y)	Parent and child together	Conduct or behavioral disorder	Behavioral health specialist, physician, physician assistant, nurse practitioner	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Active (enhanced usual care)	8
Lavigne et al,66 2008	117 parents	Preschool (3–6 y) (mean not reported)	Parent and child together	Conduct or behavioral disorder	Behavioral health specialist, nurse	Delivery, full intervention	Primary care	Colocation, colocated care, on-site intervention	Active	7
Lavigne et al,67 2008	117 parents	Preschool (mean = 4.6 y)	Individual parent	Conduct or behavioral disorder	Researcher, nurse	Delivery, partial intervention	Primary care	Colocation, colocated care, on-site intervention	Active	8
McFarlane et al,68 2005	233 women	Toddler (18 mo) to adolescence (18 y) (mean not reported)	Individual parent	Parent mental health or stress	Nurse	Delivery, full intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	7
Mendelsohn et al,69 2005	93 children	Infant (newborns)	Parent and child together	Conduct or behavioral disorder	Physician	Delivery, partial intervention	Primary care	Colocation, colocated care, on-site intervention	Services as usual	8
Mendelsohn et al,70 2007	99 children	Infant (newborns)	Parent and child together	Parenting skills	Researcher	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Services as usual	7
Minkovitz et al,71 2003	5565 children	Infant (enrolled up to 4 wk old)	Parent and child together	Conduct or behavioral disorder	Physician, nurse, health steps specialist	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Active	7
Minkovitz et al,72 2007	5565 children	Infant (enrolled up to 4 wk old)	Parent and child together	Conduct or behavioral disorder	Physician, nurse, health steps specialist	Delivery, partial intervention	Primary care	Integrated care, on-site collaborative care	Unspecified	7
Myers et al,73 2015	223 children	School age (5.5–12 y) (mean = 9.3 y)	Individual child	Conduct or behavioral disorder	Behavioral health specialist, physician	Referral	Web	External mental health clinician	Active	8
Patterson et al,74 2002	116 parents	Toddler (2 y) to school age (8 y) (mean not reported)	Parent group	Depression	Nurse, other health professional (health visitor)	Referral	Colocated	Colocation, colocated care, on-site intervention	No intervention	8
Perrin et al,75 2014	273 parents	Toddler (2 y) to preschool (4 y) (mean = 2.8 y)	Parent group	Conduct or behavioral disorder	Researcher	Delivery, partial intervention	Primary care	Colocation, colocated care, on-site intervention	Active (wait list)	8
Reedtz et al,76 2011	186 children	Toddler (2 y) to preschool (8 y) (mean = 3.9 y)	Parent group	Parenting skills	Nurse	Delivery, full intervention	Primary care	Colocation, colocated care, on-site intervention	No intervention	8
Reid et al,77 2013	178 parents	Toddler (2 y) to preschool (5 y) (mean = 3.2 y)	Individual parent	Conduct or behavioral disorder	Behavioral health specialist	Delivery, partial intervention	Phone	External mental health clinician	Services as usual	8
Sanders et al,78 2008	3004 parents	Preschool (4–7 y) (mean not reported)	Individual parent	Parenting skills	Physician, nurse	Delivery, partial intervention	Home	Training the primary care clinician in mental health skills	Services as usual	7
Saulsberry et al,79 2013	83 adolescents	Adolescent (mean = 17.4 y)	Individual parent	Depression	Physician, other health professional	Delivery, partial intervention	Primary care	Training the primary care clinician in mental health skills	Active	8
Schappin et al,80 2013	67 children	Preschool (mean = 3.7 y)	Parent group	Parenting skills	Other primary care health professional	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Active (wait list)	8
Schilling et al,81 2017	120 parents	Preschool (2–6 y) (mean not reported)	Parent group	Parenting skills	Behavioral health specialist	Delivery, full intervention	Primary care	Colocation, colocated care, on-site intervention	Services as usual	8
Shaffer et al,82 2017	321 parent-child dyads	Preschool (5 y) to school age (12 y) (mean = 8.0 y)	Individual parent	Conduct or behavioral disorder	Behavioral health specialist, physician	Referral	Primary care	External mental health clinician	Active (enhanced usual care)	7
Sourander et al,83 2016	464 parents	Preschool (4 y) (mean not reported)	Individual parent	Conduct or behavioral disorder	Behavioral health specialist, physician	Referral	Web	External mental health clinician	Active (education)	8
Spijkers et al,84 2013	81 families	School age (9–11 y) (mean = 10.6 y)	Parent and child together	Anxiety	Nurse	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Services as usual	8
Swift et al,85 2009	29 children	Toddler (2 y) to school age (12 y) (mean = 7.0 y)	Individual parent	Conduct or behavioral disorder	Behavioral health specialist	Delivery, full intervention	Phone	Training the primary care clinician in mental health skills	Active (wait list)	8
Turner et al,86 2006	30 families	Toddler (2 y) to preschool (<6 y) (mean = 3.1 y)	Individual parent	Conduct or behavioral disorder	Nurse	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Active (wait list)	8
Weisleder et al,87 2016	463 families	Toddler (14 mo) to preschool (36 mo) (mean not reported)	Individual parent	Parenting skills	Other health worker	Delivery, full intervention	Primary care	Colocation, colocated care, on-site intervention	Services as usual	8
Wissow et al,88 2008	403 youth	Preschool (5 y) to adolescence (16 y) (mean = 10.4 y)	Parent and child together	Anxiety	Physician	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Unspecified	8
Wissow et al,89 2011	403 youth	Preschool (5 y) to adolescence (16 y) (mean = 10.4 y)	Parent and child together	Parenting skills	Physician	Delivery, full intervention	Primary care	Training the primary care clinician in mental health skills	Services as usual	7

“Other” includes nonclinical staff in the primary care facility, such as administrative staff or study-trained interventionists whose roles were not otherwise specified, or external clinicians and staff who may have been responsible for partial intervention delivery, such as teachers. Design quality ratings are on an 8-point scale (described in Methods section).

Clinical Outcomes

Although we were not focused on intervention effects or synthesis of effect via meta-analysis in this review, we observed that the patterns of results on both child and parent outcomes were inconsistent, in that researchers of some studies reported immediate effects that trailed off and were not maintained, others did not have immediate effects but rather “sleeper effects”⁹⁰ that emerged after completion of the intervention, and a small proportion reported immediate effects that were also sustained. Many studies had small effects on child, parent, or both outcomes. No study reported null effects on all outcomes. Authors of recent meta-analyses of the effectiveness of parenting interventions in general have reported small to medium effects (standardized mean differences range from 0.12 to 0.53), with significant heterogeneity across studies and outcomes and inconsistency in immediate versus follow-up effects.^5,13,91,92 Effects reported for parenting interventions in primary care are similar in magnitude to those in general (standardized mean differences range from 0.29 to 0.34), but there are only 2 meta-analyses and far fewer total trials.^40,41

In our review, there was a high degree of heterogeneity among effects on targeted outcomes, which were most commonly disruptive behaviors (aggression, general externalizing behavior, conduct problems, or attention-deficit/hyperactivity disorder), followed by anxiety, depression, and developmental outcomes. Some studies had a primary focus on physical outcomes, such as sleep or well-child visit adherence, and studies in which researchers looked at infants or very young children examined temperament (but with a mental, emotional, and behavioral health outcome reported). Participant follow-up timeframes ranged from 2 weeks to 24 months, with the exception of the 1 population-level trial that had a 3-year follow-up.⁷⁸ Comparison conditions were services as usual (n = 20) and active intervention (n = 15; such as educational materials = 1; wait list = 4; enhanced usual care = 4; evidence-based program = 6), and 5 were unspecified.

Parent outcomes were also highly variable. The most common pertained to maternal mental health, parenting stress, and parenting practices, such as harsh discipline. Researchers of some studies looked at specific types of parenting skills, such as parent behavior in response to child misbehavior,^71,72 and understanding of child development and how to support their child.^62,79 Researchers of one study examined overall parental well-being.⁵⁹

Intervention Characteristics

Several studies were conducted by using manualized interventions, including Incredible Years, Triple P, Primary Care Triple P, Child-Adult Relationship Enhancement, and the Strongest Families Telephone Program, whereas other programs were unnamed. Of those providing sufficient information, interventions were delivered to parents alone (n = 13), with their children present (n = 12), or in group format (n = 9). Five interventions had a home-visiting or take-home component, and 5 had technology-based components or primary delivery strategy, such as phone- or Web-based delivery. Interventions typically occurred over 1 to 12 sessions, with 1 or more sessions per week up to 12 weeks, although 3 studies intervened periodically over 2- and 3-year periods.^51,55,87

Implementation-Related Characteristics

Researchers of approximately one-third of included studies described in detail how intervention deliverers and/or involved health care staff were trained. Among them, most provided training sessions beyond dissemination of training manuals (n = 10), and trainings lasted from 1 to 9 sessions (mode = 3) and took place over a few weeks or, in a few cases, several months. Most trainings involved some sort of manualized or discrete training protocol, followed by experiential training (eg, simulated patient interview).

Program Integration and Delivery Agents

Concerning program integration within primary care, involvement of the health care team was highly variable, including a pure integrative model (n = 8), co-located care (n = 9), and training the physician to deliver the program (n = 9). Researchers of 6 studies facilitated referrals to external mental health clinicians rather than delivering the intervention through primary care or in collaboration with primary care physicians. The variation in integration and delivery model adds to the difficulty of gauging program feasibility and sustainability in primary care. Model type was coded by using the Kolko and Perrin²³ models of effective delivery of behavioral health services in pediatric primary care and is summarized in Table 1. Future researchers should test the relative advantages and feasibility of implementation between 2 or more models as is being done in other areas of child health care.⁹³

Program delivery was also highly variable, with physicians only (n = 6); physicians or other health care provider (eg, medical assistant) (n = 3); specialty behavioral health clinician only (n = 6); physician or behavioral health clinician (n = 3); nurse only (n = 9); the researchers themselves (n = 3); or some other combination of health care and behavioral health staff, nonclinical staff in the practice, or graduate students outside of the practice (n = 7) (see Table 1). In sum, physicians were included in 14 of the 40 trials, and other health care staff (nurse practitioners, physician assistants, nurses) were responsible for delivering the full intervention in 11 studies. In the remainder, clinicians were responsible for part of the program, typically a motivational interview or introduction to the intervention followed by referral. Those that did not involve the physician or a clinician most commonly referred participants to an outside behavioral health specialist or trained research staff (often graduate trainees). Carefully considering who will be responsible for delivering preventive services in primary care has been noted as a significant barrier to implementation because physicians’ time is severely limited for such activities.⁹ Furthermore, specialty behavioral health clinicians are in short supply nationally and remain a luxury in pediatric primary care practices.⁹⁴ Greater attention to sustainable staffing models for parenting programs is sorely needed.⁹⁵

Implementation Outcomes

Fidelity was the most commonly measured and reported implementation outcome (n = 7), followed by parent’s or youth’s acceptability of the intervention (n = 5) (see Table 2). Among those reporting fidelity, researchers of 3 studies indicated that fidelity was assessed but did not report associated measures or outcomes^52,77,80; researchers of other studies stated that fidelity was “achieved” or there was a “high-level of fidelity” but did not specify what constituted fidelity, how it was measured, or any criteria. Most study groups used study-specific fidelity measures. Thus, it is challenging to judge the validity and reliability of the measures used. Only 2 trials reported on cost of the intervention. No trials measured or reported other common implementation outcomes³¹: acceptability of the intervention by the health care team, appropriateness, feasibility, reach and penetration rate, or program sustainability. Three studies reported on participation rate and dose of the intervention, which can be an indicator of implementation, but they did not interpret it in that context.

TABLE 2.

Implementation Outcome Measurement and Analysis

Source, Year	Implementation Outcome	Measure	Comparison	Result or Finding
Hiscock et al,59 2007	Satisfaction (with and usefulness of each intervention component) (parents), Acceptability	10-cm visual analog scales (helpful to unhelpful; did not receive or do this)	None	Satisfied (median 7.7; IQR, 5.5–9.2), useful (median 7.3; IQR, 3.0–9.1)
	Cost of intervention	Per hour cost based on research budget	Control	Intervention £96.93 (SD 249.37) versus control £116.79 (SD 330.31); mean difference £19.44 (95% CI, 83.70–44.81), P = .55
Hiscock et al,60 2008	Fidelity	Checklist at end of each session (nurse reported)	None	Full program was delivered for 95% of group sessions
Iloabachie et al,61 2011	Acceptability of program (adolescents)	Survey developed for this study (9 items relevant to acceptability), parent interviews (qualitative analysis)	None	Before and after (baseline and 6–8 wk postintervention): Internet site preferences, adolescents reported greater importance of chatting online as part of the intervention after participation than before (P = .01) intervention format (wanting more interactive format)
Kolko et al,64 2011	Fidelity	Mean percentage of correctly administered content or procedures in the 2 treatment protocols (observer rated)	None	Average rating 91.4 (SD = 9.1); fidelity was unrelated to either child dysfunction or child health
Mendelsohn et al,70 2007	Cost of intervention	No measure reported	None	∼$240 per child per year
Minkovitz et al,72 2007	Satisfaction (parents), Acceptability	Survey (no more details)	Control	Satisfaction was high
	Fidelity	Checklist	Control	Range: 48.2%–86.6%; higher in intervention than control
Myers et al,73 2015	Fidelity	Telepsychiatrists and therapists used checklists that outlined the essential treatment components to address during each session. All sessions were recorded, and sessions were randomly selected to rate the clinicians’ coverage of these components	None	High fidelity to protocols (other publications referenced but no results provided in this article)
Perrin et al,75 2014	Fidelity	Incredible Years program, Leader’s Process Checklist	None	Session activities were documented by group leaders after each session using the Incredible Years Leader’s Process Checklist and taped sessions that were randomly reviewed for fidelity. Over 90% of content and delivery components were delivered during each session. 100% of reviewed taped sessions had every parent speak at least once
Sanders et al,78 2008	Clinician training, adoption	375 clinician trained over 21 courses	None	73% of clinician became accredited
	Participation and dosage (intervention elements used)	N/A	None	29 television programs, 48 radio broadcasts, 58 newspaper or magazine articles, and 1750 Web site hits
	Participation and dosage (families)	N/A	Control group	7.5% of families participated in the program versus control (2.1%; χ2 = 48.93, P < .001)
Schilling et al,81 2017	Participation and dosage (parents)	N/A	None	43% of intervention parents attended 3 or more sessions (of 6)
	Fidelity	Checklist (clinician report)	None	Clinicians received weekly oversight and documented session activities after each intervention session according to a fidelity checklist
	Satisfaction (youth), Acceptability	Likert scale surveys and qualitative exit interviews	None	81% of youth completing the surveys reported being satisfied or very satisfied with the intervention. Qualitative analyses aligned with the survey results, expressing satisfaction with the intervention itself and with the delivery location (primary care)
Turner et al,86 2006	Satisfaction (parents), Acceptability	Client Satisfaction Questionnaire	None	Mothers reported high satisfaction (mean = 72.89; SD 11.48)
Saulsberry et al,79 2013	Fidelity	MI fidelity scale	Control (brief advice)	Clinicians in the treatment group had high fidelity (4.5 out of 5, SD = 0.83) to the MI model versus control fidelity to MI (1.02, SD = 0.07). Interviews were longer in the MI group (P = .002)
Wissow et al,88 2008	Adoption	N/A	None	84% of eligible clinicians accepted and received training once a lead clinician was identified at each site. No attrition

IQR, interquartile range; MI, Motivational interview; N/A, not applicable.

The review also reveals a lack of evidence for sustainment of parenting interventions. The studies were focused almost exclusively on intervention effects, which is meaningful but not sufficient for understanding the impact of delivering these programs in primary care over the long-term. Transporting interventions that were largely developed and tested in laboratory-based settings to existing systems, such as pediatric primary care, can be a tremendous challenge, yet doing so is imperative to achieving a public health impact on children’s behavioral health.¹⁵ Additionally, there was no formal reporting of procedures used to adapt programs to the primary care settings. More studies are needed that focus on the context of implementation, the adaptations required for a protocol to fit with primary care, and the predictors of sustained adoption. Researchers should consider the use of implementation trial designs^96,97 to place the focus on the system that is needed to support program adoption and delivery. Data obtained from such trials would inform translation of programs that show effects in primary care or have been found to be efficacious in other contexts.⁹⁸ This research is imperative to advance parenting interventions in primary care from efficacy and effectiveness to wide-scale implementation. Taken together, parenting interventions for youth mental, emotional, and behavioral disorders have not been sufficiently tested in a way that shows they can be adopted and sustained in primary care. At least 2 trials^99,100 that are known to the authors are underway in which researchers focus on both implementation and clinical effectiveness of a parenting program in primary care, known as a hybrid design,¹⁰¹ but more are needed.

Trial Quality

The included studies had the majority of the characteristics of a well-designed intervention study, as evidenced by ratings of 7 or above on all included studies (mean = 7.75). This indicated a high degree of scientific value.⁴⁸ Importantly, all studies had clearly stated aims, a comparable comparison condition, and postintervention results provided for all conditions.

Discussion

In this review, we add to the mounting evidence that parenting interventions can be effective at improving child mental, emotional, and behavioral health and parental outcomes when delivered in coordination with pediatric primary care. However, the way in which they are implemented matters, and this has not been the focus of research to date. Our findings also expose a pronounced lack of focus on implementation research methods and the context of implementation in the published literature. This is preventing the field from scaling up parenting intervention and achieving an equitable public health impact. There is sufficient evidence of effectiveness: it is time to research the implementation strategies that support adoption and sustained delivery of parenting interventions in primary care.

Limitations

The results of this review should be considered with the following caveats. Our approach limiting the search to published articles only introduces the potential for underrepresenting the literature as a result of not contacting any study authors while conducting the search and preparing the article and not examining unpublished “gray” literature. Further publication bias is possible with what we perceive to be a preference for effectiveness study results among many journals, with perhaps less of an interest in implementation processes and outcomes (particularly when researchers in the study primarily aim to establish the effectiveness base for parenting interventions in primary care and secondarily aim to examine implementation). This suggests that implementation data may have been collected but were not included in the published results of the included studies in this review. We also noticed that few studies in this review were conducted outside of the United States. This could be a function of the search terms, such as primary care, which might not be the preferred term internationally, or our inclusion of only English language publications.

Glossary

CI

confidence interval