Parents’ Consumer Preferences for Early Childhood Behavioral Intervention in Primary Care

Abstract

Objective

Early childhood parenting interventions are increasingly delivered in primary care, but parental engagement with those interventions is often suboptimal. We sought to better understand parents’ preferences for the content and delivery method of behavioral health guidance in pediatric primary care, and to determine the relationship of those preferences with demographic characteristics, child behavior problems, and parenting style.

Method

Participants were 396 parents of young children recruited from primary care offices. We collected measures of parental preferences (including behavioral topics, intervention strategies, and methods of delivery) for behavioral intervention in primary care, child behavior symptoms, parenting style, and demographic characteristics. Descriptive statistics were used to identify parents’ most preferred behavioral topics and intervention delivery methods. We used a hierarchical regression approach to determine whether parenting style predicted parents’ preferences beyond demographic and child-level factors.

Results

Nearly all parents (96%) endorsed a behavioral topic (e.g., aggression) as important. Most preferred to receive intervention during routine medical appointments. Child behavior problems correlated with parents’ overall interest in behavioral guidance, but clinically significant symptoms did not differentiate interest in any single topic. Socioeconomic factors and negative parenting practices predicted some parental preferences. Notably, lax parenting generally predicted higher interest in behavioral intervention, whereas hostile and physically controlling parenting predicted lower interest.

Conclusion

Most parents are interested in behavioral guidance as part of primary care, but their preferences for the content and delivery of that guidance varies by known socioeconomic, child, and parenting risk factors. Tailoring intervention to parents’ preferences may increase engagement with available interventions.

Promoting safe and effective child discipline practices is an important strategy for preventing child maltreatment and improving population health outcomes, and pediatric primary care had been identified as a key setting for intervening on parenting.¹ Strategies for promoting effective parenting in primary care including the integration of behavioral health specialists, technology-based interventions, and enhanced training for primary care providers (PCPs) show promise,² but questions persist regarding how to optimize parents’ participation in interventions. A recent systematic review of primary care parenting interventions found wide variability in rates of study enrollment (30–100%) and retention (38–100%) and noted “far from perfect” engagement with interventions. The authors concluded, “There are likely unmeasured characteristics that influence both parental participation in a behavioral health program and the environment in which a child’s behavior patterns develop.”³ Variable participation in behavioral interventions may reflect heterogeneity in parents’ attitudes towards those interventions, which in turn may be related to variables of clinical interest, such as parenting practices.

Parents’ preferences for behavioral health intervention in primary care may be impacted by several factors including both the informational content and delivery methods of those interventions. Research consistently indicates that many parents want more guidance than they receive on child discipline and other behavioral topics as part of their child’s primary care,^4–7 but the exact nature of this need from a parental perspective remains unclear. For example, colloquial use of the term “discipline” may refer to any number of parenting practices (e.g., reprimanding, praising, reasoning) to address a variety of child behaviors (e.g., aggression, non-compliance, emotion regulation), but previous research has treated “discipline” as a unitary concept. Identifying the specific child behaviors for which parents require assistance, rather than general categories (e.g., “externalizing behavior”) is important because the effectiveness of discipline strategies may vary based on the type of behavior being addressed.⁸ In addition to the informational content of interventions, it’s important to appreciate parents’ preferences for how that content is delivered. For example, some research indicates parents generally prefer self-administered interventions (e.g., video-based programs) to traditional modes of counseling;⁹ however, little is known regarding how these preferences relate to other variables of clinical interest (e.g., baseline parenting practices) in the primary care context.

A refined knowledge of parents’ priorities and preferences for the delivery of behavioral health care in primary care is needed to guide the development and dissemination of a spectrum of family-centered interventions.¹⁰ In order to better understand parental perspectives on the delivery of behavioral health information in pediatric primary care, we surveyed parents of young children. We sought to characterize parents’ attitudes towards delivery of behavioral health information in primary care, and assess the relationship between parent attitudes and demographic, parenting, and child characteristics of interest. We hypothesized that clinically significant behavior symptoms and associated risk factors, including lower socioeconomic status and more negative parenting styles, would be associated with greater reported importance of behavioral topics and interest in behavioral services.

METHODS

Design

This study used a cross-sectional survey approach. The methods were developed in collaboration with a stakeholder advisory panel consisting of two community pediatricians, one community psychologist working in pediatric primary care, one parent of young children from the community, and one professional parent advocate. This study was approved by the human subjects institutional review board at Oregon Health & Science University.

Participants

English and Spanish speaking parents of children between the ages of 18 months and 5 years 11 months were recruited through five pediatric practices, two that were university-affiliated (one within an urban academic medical center, one suburban) and three that were private practices (one urban, one suburban, one rural). Each of these clinics offered some co-located behavioral health services consistent with Level 3 of the Standard Framework for Levels of Integrated Healthcare.¹¹ All participants were recruited between March 2018 and May 2018. Parents were made aware of the study by clinic staff and screened for interest and eligibility via tablet computer. If interested in participation, parents provided their contact information and indicated a preference for either electronic or paper survey administration. Prospective participants were then contacted by email or mail and were given further information about the study and access to the survey, which was available in both English and Spanish. Electronic surveys were completed via REDCap,¹² and paper surveys were completed and returned via pre-paid envelopes. All participants received a $25 gift card for survey completion.

A total of 454 parents were screened for interest and eligibility. Because the screening was introduced by clinical staff rather than researchers, it is unclear whether a significant number of parents declined screening. Twenty parents were deemed ineligible because their child’s age was outside the target range. Amongst the 434 eligible participants who were invited to participate, 396 (91%) completed at least one measure and were included in the analyses. Ten surveys (3%) were completed in Spanish. Demographic characteristics of the sample are reported in Table 1.

Table 1.

Participant characteristics (N = 396)

Characteristic	Value
Parent
Age, years, M (SD)	33.2 (6.6)
Female sex, %	85
Ethnicity, %
Hispanic/Latino	15
Non-Hispanic/Latino	77
Unknown	8
Race, %
White	62
Asian	11
Black	2
American Indian/Alaska Native	2
Other	4
Multiracial	12
Unknown	7
Marital Status, %
Married	76
Widowed	<1
Divorced or separated	7
Remarried	1
Never married	16
Number of children, M (SD)	2.1 (1.1)
Parenting Situation, %
Single Parenting	11
Couple Parenting, same household	85
Co-parenting, separate households	5
Education, %
High school diploma or less	13
Vocational school/some college	20
College degree	41
Graduate/professional degree	26
Annual Household Income, %
$25,000 or less	11
$25,001-$49,999	24
$50,000-$79,999	28
$80,000-$119,999	15
$120,000-$149,999	9
$150,000 or more	13
Child*
Age, years, M (SD)	3.5 (1.3)
Female sex, %	46
First born child, %	62

Note:

^*Parents were asked to report on their oldest child in the target age range.

Measures

The primary outcome measure was the Behavioral Information Preferences Scale (BIPS), which we created for this study to measure parental preferences for behavioral intervention in primary care. We developed an initial item pool through review of common behavior symptom inventories,^13–15 research examining parents’ behavior concerns in primary care,^4–6 and literature reviews of behavioral interventions in primary care.^2,16,17 Items were then reviewed by the stakeholder advisory panel and a group of pediatric psychologists (N=8), and piloted with parents of young children (N=9) in order to ensure readability and acceptability of content. The resulting measure consisted of 39 five-point Likert-type items (1=Not at all, 5 = Very) divided into three sections: Behavior Topics, Delivery Method, and Intervention Approach. The English version was translated into Spanish using a professional medical translation service.

Following data collection, we used a principal components analysis item-reduction approach to derive composite scores from each section. The Behavior Topics section asked parents to rate the importance of individual child behavior topics and resulted in two subscales: Emotions and Conduct (12 items; α = .96 in this sample), which assesses topics related to internalizing and externalizing concerns (e.g., “Teaching children not to be aggressive towards others, like hitting, kicking, and biting”), and Healthy Habits (6 items; α = .84), which assesses interest in topics related to daily habits or developmental skills (e.g., “Helping children fall asleep and sleep through the night on their own”). The Delivery Methods section assessed parents’ interest in different modalities of behavioral information delivery and yielded three subscales: Auxiliary Care (5 items; α = .79), which measures interest in interventions that involve in vivo interaction with another person (e.g., “Talking to a behavioral expert over the phone talking to a behavioral specialist”); Multimedia Resources (9 items; α = .90), which measures interest in text (e.g., “Paper handouts from my child’s doctor”) or technology-based (e.g., “Mobile apps for smartphones or tablets”) interventions; and Usual Care, a single item that assesses interest in discussing behavioral topics with the child’s PCP during routine visits. The Intervention Approach section (6 items), which prompted parents to rate the helpfulness of receiving information about different behavior intervention strategies (e.g., “Setting up appropriate expectations for child behaviors”), possessed borderline psychometric qualities, so a composite score was not calculated.

Child Behavior Symptoms

Child behavior symptoms were assessed via the Eyberg Child Behavior Inventory (ECBI),¹³ a common and well validated measure of child disruptive behavior. The ECBI yields two subscales. The Intensity Scale (ECBI-I; 36 items; α = .92) measures the frequency of child disruptive behaviors. The Problem Scale (ECBI-P; 36 items; α = .92) measures the degree to which a parent considers behaviors problematic. Raw scores can be converted to T-scores and established clinical cut-offs.

Parenting Style

We used the Multidimensional Assessment of Parenting Scale (MAPS) to assess parenting style.¹⁸ The MAPS consists of 34 items and produces broadband scores for Positive Parenting (16 items; α = .89) and Negative Parenting (18 items; α = .88), as well as subscale scores for Positive Reinforcement (4 items; α = .71), Proactive Parenting (4 items; α = .79), Warmth (3 items; α = .75), Supportiveness (3 items; α = .79), Lax Control (7 items; α = .84), Hostility (4 items; α = .90), and Physical Control (4 items; α = .90). The MAPS possesses a theoretically coherent factor structure and demonstrates strong psychometric properties.¹⁸

Demographics

We collected demographic information including parent and child age, sex, racial/ethnic identity, family composition, and household income.

Analyses

We used descriptive statistics to characterize the sample and calculated correlation coefficients to assess bivariate relationships between variables. We examined responses to the BIPS in multiple ways. In order to determine which specific child behaviors, delivery methods, and intervention approaches were most important to parents, we calculated the proportion of parents who provided a rating of 4 or 5 (e.g., “Important” or “Very Important”) for each BIPS item. We then conducted Mann-Whitney U tests to determine whether parents’ responses to individual items varied significantly across groups based on child age (above or below 3 years) or the ECBI-I clinical cutoff score for each individual item. We were interested in examining these relationships, because behavioral guidance in primary care may be based on child age (e.g., recommended anticipatory guidance topics¹⁹) or triggered by positive behavioral symptom screening.²⁰ To correct for multiple comparisons, we used Holm’s sequential Bonferroni procedure within each subscale.²¹

We conducted a series of hierarchical linear regressions to evaluate child characteristics, including the presence of significant problem behavior (as defined by ECBI-P clinical cutoff scores), and parenting style as potential predictors of parents’ preferences for behavioral information in primary care above and beyond parent and family demographic characteristics. We conducted a separate regression using each subscale of the BIPS as an outcome measure. We carried out statistical diagnostics and visual residual plot inspection to ensure the data met assumptions for linear regression.

Of the 396 participants, 374 (94%) completed all measures and 28 (7%) had at least one missing value, resulting in 4% missing data overall. Little’s Missing Completely at Random test did not indicate any systematic bias of missing data, X² (452) = 490.88, p = .100. To replace missing values, we imputed five datasets and generated pooled estimates of those datasets. The reported results reflect the pooled estimates. All analyses were performed using SPSS version 25.0.²²

RESULTS

General Preferences and Priorities for Behavioral Guidance

Table 2 displays the percentage of parents who positively endorsed each item of the BIPS. In the Behavior Topics section, 96% of parents rated at least one behavior as important or very important, and aggression was most commonly endorsed as “Very Important” (51%). Aggression was also the only individual behavior that varied significantly by child age, with parents of younger children rating the topic as more important. Most other behaviors trended in this direction but differences were not significant. There were no significant differences based on the ECBI-I clinical cutoff score.

Table 2.

Parent endorsement of child behavior topics, delivery methods, and intervention approaches for primary care behavioral guidance.

	Percentage of Parents Endorsing
		Child age (years)			ECBI-I cutoff
BIPS Items	%	< 3	≥ 3	p	Below	Above	p
Behavior Topics Section1
Emotions and Conduct subscale
Tantrums	77	82	74	.10	78	77	.68
Coping with difficult emotions	74	78	72	.09	74	77	.87
Aggression	74	81	69	.004*	74	75	.71
Self-calming	72	78	68	.05	72	72	.42
Destruction	70	77	65	.11	71	64	.81
Listening to instructions	67	70	65	.14	67	71	.70
Misbehavior in public	67	67	67	.45	66	72	.41
Sharing	66	70	64	.11	67	64	.89
Following rules	61	65	59	.13	61	61	.86
Being honest	60	64	58	.31	60	62	.88
Inappropriate speech	52	56	50	.21	51	58	.35
Transition between activities calmly	50	53	48	.22	49	55	.07
Healthy Habits subscale
Picky eating	73	77	70	.01	72	74	.88
Sleep	68	73	65	.03	67	72	.35
Toileting	64	68	62	.17	64	67	.48
Limiting screens and toys	63	64	62	.20	62	67	.79
Attentiveness	55	58	54	.70	54	59	.21
Bad habits	51	49	52	.86	51	49	.57
Delivery Methods Section2
Usual Care
PCP – routine visit	81	85	78	.05*	80	84	.60
Auxiliary Care subscale
Behavior specialist – routine visit	61	62	60	.33	58	75	.003*
PCP – separate visit	52	61	46	.01*	50	61	.03
Behavior specialist – separate visit	48	51	46	.42	46	58	.02
Behavior specialist – phone	46	53	42	.05	43	62	.006*
Group class/seminar	35	42	31	.02	34	39	.13
Media Resources subscale
Websites	49	53	47	.14	47	61	.03
Mobile apps	48	52	46	.17	46	59	.008*
Books	47	53	43	.09	47	48	.37
Patient portal	47	48	47	.45	45	59	.04
Paper handouts	45	51	42	.32	44	51	.11
Online programs	45	45	44	.50	42	58	.002*
Videos	41	44	40	.48	39	51	.002*
Podcasts	32	36	29	.21	30	41	.04
Social media	22	21	22	.81	20	32	<.001*
Intervention Approach Section3
Typical/atypical behavior	89	91	88	.29	90	86	.41
Positive parent-child relationship	83	84	83	.83	85	77	.15
Stress management	80	81	79	.88	80	78	.64
Set expectations for child behavior	75	77	74	.33	75	77	.33
Consequences for misbehavior	69	68	69	.82	69	68	.65
Reward/approval of good behavior	68	70	66	.43	67	71	.56

Note:

^*Statistically significant Mann-U Whitney value using Holm’s Sequential Bonferroni Procedure. BIPS = Behavioral Information Preferences Scale; ECBI-I = Eyberg Child Behavior Inventory Intensity Scale. PCP=Primary care physician.

¹Endorsed as “Important” or “Very Important”

²Endorsed as “Interested” or “Very Interested”

³Endorsed as “Useful” or “Very Useful”

Within the Delivery Methods section, 81% of parents endorsed interest in talking to their child’s doctor about child behavior during routine visits, and parents of younger children were significantly more likely endorse interest in discussing behavioral topics with the PCP during both routine and separate visits. Parents of children with clinically significant behavior problems reported significantly higher interest in alternative delivery methods, including meeting with a behavioral specialist during routine visits and over the telephone, mobile apps, online programs, videos, and social media.

Parents most commonly endorsed education about which behaviors are typical, building a positive parent-child relationship, and stress management as useful intervention approaches. There were no differences between child age groups or ECBI cutoff groups on the Intervention Approach items.

Bivariate Associations

Table 3 displays bivariate associations between the BIPS subscales, demographic variables, child behavior symptoms, and parenting practices. Parent-rated importance of Emotions & Conduct behavioral topics was negatively correlated with child age and household income, and positively correlated with behavioral problems and total negative parenting. The Healthy Habits subscale showed a modest positive correlation with behavior problems. Interest in Auxiliary Care and Media Resources methods of delivery were both negatively correlated with income and positively associated with both child behavior symptoms and negative parenting practices. Negative parenting was the only significant correlate of interest in Usual Care.

Table 3.

Bivariate relationships between demographic characteristics, child behavior symptoms, parenting style, and behavioral information preferences.

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19
Demographics
1. Child age	-	.24**	−.03	.10	.17**	.00	.09	−.01	−.12*	.04	.14**	.27**	−.03	.09	−.10*	−.09	−.07	−.11**	−.09
2. # children in family		-	−.03	.08	.07	.05	.08	.05	.02	.02	−.06	−.04	−.10	−.00	.00	.02	−.06	−.08	−.04
3. Income			-	−.04	−.08	.01	.03	−.06	.05	−.03	−.10	.05	−.10	−.17**	−.12*	−.08	−.14**	−.11*	−.06
ECBI
4. Intensity scale				-	.63**	−.19**	−.10	−.13**	−.20**	−.20**	.41**	.47**	.29**	.20**	.04	.03	.21**	.17**	.01
5. Problem scale					-	−.19**	−.15**	−.07	−.21**	−.19**	.35**	.39**	.19**	.22**	.15**	.10*	.13**	.11*	.00
MAPS
6. Total positive						-	.78**	.85**	.76**	.87**	−.45**	−.30**	−.37**	−.35**	−.09	−.07	−.16**	−.10	.00
7. Proactive parenting							-	.57**	.41**	.56**	−.37**	−.20**	−.45**	−.20**	−.12*	−.05	−.20**	−.11	.02
8. Positive reinforcement								-	.55**	.67**	−.36**	−.24**	−.28**	−.29**	−.01	−.01	−.07	−.05	.00
9. Warmth									-	.57**	−.32**	−.25**	−.20**	−.28**	−.10	−.09	−.11*	−.11*	.00
10. Supportiveness										-	−.41**	−.29**	−.27**	−.37**	−.07	−.07	−.14*	−.05	.03
11. Total negative											-	.81**	.69**	.79**	.16**	.08	.18**	.16**	.11*
12. Hostility												-	.36**	.53**	.02	−.03	.11*	.08	−.03
13. Lax control													-	.24**	.22**	.12*	.34**	.25**	.11*
14. Physical control														-	.13*	.09	−.02	.04	.05
BIPS
15. Emotions & Conduct															-	.79**	.39**	.36**	.38**
16. Healthy Habits																-	.38**	.37**	.46**
17. Auxiliary Care																	-	.49**	.23**
18. Media Resources																		-	.20**
19. Usual Care																			-

Note:

^*p ≤ .05

^**p ≤ .01

^***p ≤ .001.

BIPS = Behavioral Information Preferences Scale; ECBI = Eyberg Child Behavior Inventory; MAPS = Multidimensional Assessment of Parenting Scale. Correlation coefficients are Pearson’s R, except for BIPS-Usual Care and Income, which are Spearman’s Rho.

Predictors of Parents’ Reported Preferences

Table 4 presents results of multiple regression models for the BIPS Behavior Topics and Delivery Methods subscales. Parent-rated importance of Emotions and Conduct topics was associated with race/ethnicity, such that minority status predicted higher interest (β = −.23, p ≤ .001). Clinically significant behavior problems (β = .12, p = .005) and lax parenting (β = .17, p = .005) positively predicted interest in Emotions and Conduct topics. For the Healthy Habits subscale, parent-rated importance was positively predicted by racial/ethnic minority status (β = −.18, p ≤ .001), clinically significant behavior problems (β = .12, p = .006), and negatively predicted by parenting hostility (β = −.14, p = .04). In both the Emotions and Conduct and Healthy Habits subscales, younger child age was predictive in the second step of the model, but this relationship was attenuated when parenting style was added in the third step.

Table 4.

Results of multiple linear regressions of parent and child characteristics on parent-rated importance of behavioral topics in primary care

	Behavior Topics subscales						Delivery Methods subscales
	Emotions and Conduct			Healthy Habits			Media Resources			Auxiliary Care
Predictor Variables	Δ R2	β at entry	β at final	Δ R2	β at entry	β at final	Δ R2	β at entry	β at final	Δ R2	β at entry	β at final
Step 1 (demographics)	.08***			.05***			.04**			.03*
Sexa		−.03	−.04		−.04	−.05		.00	−.01		−.01	−.04
Race/ethnicityb		−.27***	−.23***		−.21***	−.18***		−.14**	−.08		−.05	.02
Number of children		.01	.03		.03	.04		−.08	−.04		−.06	−.02
Household income		−.09	−.05		−.06	−.03		−.10*	−.08		−.14**	−.12**
Step 2 (child characteristics)	.05***			.03**			.03**			.01
Age (years)		−.13*	−.10		−.11*	−.09		−.10*	−.12*		−.07	−.06
Sex		.09	.08		.03	.02		−.03	−.03		−.02	−.01
Problem behaviorc		.18***	.15**		.14**	.12*		.13**	.10		.10	.03
Step 3 (parenting style)	.04*			.02			.05**			.12***
Proactive parenting			−.03			.04			.01			−.05
Positive reinforcement			.12			.09			.03			.09
Warmth			−.08			−.08			−.11			−.07
Supportiveness			.01			−.05			.06			−.09
Hostility			−.12			−.14*			.03			.07
Lax control			.17**			.10			.21***			.30***
Physical control			.10			.09			−.06			−.18**
R2 for total model	.17			.10			.11			.16
F for total model	5.46***			3.01***			3.37***			5.17***

Note:

^acoded as 1=female, 2=male

^bcoded as 1=not a racial/ethnic minority, 2=racial/ethnic minority

^cEyberg Child Behavior Intentory-Problem Scale, coded as 0=below cutoff, 1=at or above cutoff

^*p ≤ .05

^**p ≤ .01

^***p ≤ .001.

When controlling for other factors, scores on the Auxiliary Care subscale were predicted by household income (β = −.12, p = .01), lax parenting (β = .30, p ≤ .001), and physically controlling parenting (β = −.18, p = .003). Notably, the two negative parenting practices were predictive in opposite directions, such that lax parenting was associated with increased interest in auxiliary delivery methods, whereas higher physical control (i.e., use of corporal punishment) was predictive of lower interest. Scores on the Media Resources subscale were predicted by child age (β = −.12, p = .04) and lax parenting (β = .21, p ≤ .001) in the final model. Beta weights for race/ethnicity, household income, and problem behavior were significant at entry into the model, but no longer contributed significantly when parenting style was considered. The overall model for Usual Care was non-significant (F[14,382]= 1.20, p = .28).

DISCUSSION

This investigation increases knowledge of parents’ priorities and preferences for early childhood behavioral health care in primary care. The results indicate that parents almost universally consider behavioral topics an important part of primary care services in early childhood, but preferences for the content and delivery of behavioral care vary based on a variety of individual factors. Topics related to disruptive behaviors and emotion regulation were rated as important most often, which seemingly corresponds with the previously identified need for discipline counseling in primary care.^4–6 Notably, while total child behavioral symptoms were correlated with overall importance of the behavioral categories captured by the Emotions & Conduct and Healthy Habits subscales of the BIPS, the presence of clinically significant behavioral symptoms did not discriminate interest in any single behavioral topic, and behavioral symptoms were less predictive overall than was hypothesized. This finding is consistent with research in older children showing that distressing subclinical behavioral problems are quite common²³ and underscores the normalcy of parents having some behavioral counseling needs. Delineating singular behavioral concerns from clinical disorders may be an important distinction for intervention stratification, as multi-component treatment packages often target constellations of symptoms (e.g., parent management training for disruptive behavior), whereas more discrete interventions can be used to target individual behaviors (e.g., time-out for aggression). Devising discrete interventions that can be customized to target parents’ priority child behaviors is an important goal for providing family-centered care, and our findings identify a number of potential behavioral targets. The child behavior intervention literature provides several potential strategies for intervention (see Riley and Freeman²⁴ for examples), but with little exception these have not been studied in primary care.

Developing and testing discrete interventions for discrete behaviors may be especially important given parents’ apparent preference for behavioral care to be delivered in the context of routine visits. Talking with a child’s PCP or a behavioral specialist during routine care were the two most commonly preferred modes of delivery across all subgroups, consistent with lower engagement rates observed for interventions requiring multiple visits in clinical research.³ This finding underscores a mismatch between parents’ consumer preferences and the preponderance of integrated behavioral health research, which has often focused on multi-session treatments. For example, group-based intervention is a relatively common modality in the literature,³ but this was the least commonly endorsed of the Auxiliary Care options. Interventions delivered the context of well-child care are unlikely to meet the needs of all families, but may provide an important “step” of care within a stepped-care approach.²⁵ Tiered models of integration that include provision of services during well-child care in addition to co-located services, such as the Primary Care Behavioral Health model²⁶ or the Healthy Steps program,²⁷ may provide the best opportunities to engage and benefit a wide swath of parents. Comparative effectiveness research to determine who is likely to benefit from various levels of intervention is needed.

Beyond engaging parents, interventions must meet the needs of PCPs and diverse clinical settings if they are to be widely adopted and implemented. Time constraints, insufficient training, and inadequate reimbursement are oft cited barriers to the provision of behavioral guidance in primary care,^28–31 and any intervention strategy that fails to address these barriers is unlikely to gain traction. Direct provision of extensive behavioral guidance by PCPs may be infeasible, but nascent evidence indicates that embedded behavioral consultants may alleviate the PCP time-burden associated with behavioral concerns and lead to increased revenues.^32,33 Reimbursement mechanisms for behavioral consultations during well-child care have emerged,³⁴ and use of alternative payment methods is associated with higher levels of integration,³⁵ so full integration is increasingly viable (though access to reimbursement mechanisms varies). Currently, the supply of mental health professionals who meet competencies for practicing in pediatric primary care³⁶ is insufficient to meet demand, but there is evidence that non-mental health staff, including nurses can function as behavioral health consultants with augmentative training.³⁷ Low-cost multimedia interventions, such as Play Nicely for prevention of aggression,³⁸ may also provide benefits without exacerbating demands on PCPs. Research focused on the dissemination and implementation of primary care parenting interventions is needed to identify impediments and facilitators of high-fidelity delivery and maintenance.

Socioeconomic factors appear to play a role in parents’ preferences for behavioral guidance in primary care. Identification as racial/ethnic minority predicted greater importance of behavioral topics as part of primary care even when controlling for other factors, and lower income predicted more interest in auxiliary care methods. This may reflect disadvantaged groups’ higher levels of unmet needs in primary care⁶ and a preference for integrated services given racial/ethnic and socioeconomic disparities in access to traditional mental health care.³⁹ Thus, it may be particularly important to assess the behavioral health needs, care preferences, and perceived barriers to care of low income and racial/ethnic minority families in the primary care setting. Research is needed to examine what factors drive these preferences (e.g., convenience of primary care setting) and to determine whether tailoring behavioral intervention to the preferences of underserved groups can reduce health disparities.

This is the first study to indicate that parenting style is predictive of attitudes towards behavioral health services in primary care, an important finding as most evidence-based parenting programs are designed to promote an authoritative parenting style. Lax parenting in particular was predictive of greater importance of behavioral topics related to conduct and emotions, and interest in all methods of intervention delivery. This suggests parents with lax style may recognize a need for information and skills to improve behavior management. Conversely other negative parenting practices were not predictive of greater interest in behavioral services and in some cases predicted lower interest. It is particularly concerning that higher physical control was predictive of lower interest in auxiliary behavioral services, as parents who use corporal punishment may be at the highest risk for child maltreatment and other poor outcomes. More research is needed to determine the underpinnings of this association; it may be critical to engage these parents during well child care or to provide access to media-based interventions. Alternatively, motivational techniques or other strategies to enhance engagement may be especially needed to engage parents who use more physical control. Future research should consider parenting style when evaluating the reach and impact of parenting interventions in primary care.

The results should be interpreted with consideration of the limitations of this study sample and methods. While we sampled from multiple communities, all were part of the Pacific Northwest of the United States and the outcomes may reflect regional attitudes and experiences. The racial/ethnic characteristics of participants were representative of the region, but the majority White sample precluded a more detailed assessment of differences between racial/ethnic minority groups, and Black parents were particularly underrepresented. Future studies should use more targeted recruitment methods to ensure that socially and economically disadvantaged groups are appropriately represented. The recruitment rate was high, but it is unclear how agreement to participate in this study might correlate with interest in primary care behavioral services or whether the 9% that declined participation represented a categorically different population. The study was cross-sectional and it is unclear how parents’ reported preferences are related to actual engagement with behavioral resources, which is likely determined by multiple interrelated factors.

CONCLUSION

Overall, our findings indicate that parents of young children are highly interested in behavioral guidance in primary care, particularly in the context of routine care, and that parents’ preferences for behavioral care vary based on demographic, child, and parenting-level variables. Developing interventions that can be matched to parents’ individual priorities and preferences, both with regards to content and delivery modality, may increase engagement and ultimately improve child health.