Development and Initial Validation of a Measure of Parents’ Preferences for Behavioral Counseling in Primary Care

Abstract

Introduction:

There is a significant need to understand the factors that contribute to parents’ consumer preferences for behavioral health services in pediatric primary care; however, no validated measure of such preferences exists. We developed the BIPS (Behavioral Information Preferences Scale), a measure of parents’ preferences for delivery of behavioral guidance in pediatric primary care, and assessed its psychometric properties.

Method:

An initial item pool consisted of three sections: Behavior Topics, Intervention Approach, and Delivery Methods. In addition to the BIPS, parents of young children (N=396) completed measures of child behavior problems and parenting self-efficacy. We conducted principle component analyses and examined correlations of the resulting factors.

Results:

The Behavior Topics section resulted in a two-factor solution (Conduct/Emotions and Healthy Habits), as did the Intervention Approach section (Behavior Change and Psychoeducation), whereas the Delivery Methods yielded three factors (Usual Care, Auxiliary Care, and Media Resources). Patterns of association with parent reported child behavior problems and parenting self-efficacy were indicative of construct validity for the Behavior Topics and Media Resources sections.

Discussion:

The BIPS holds potential for informing the design and dissemination of primary care parenting interventions.

Early childhood is a critical period for preventing later psychological and behavioral problems (O’Connell, Boat, & Warner, 2009). Nearly all young children are seen repeatedly for health supervision in a primary medical home, and primary care-based parenting interventions hold strong potential to prevent poor child outcomes and promote long-term population health (Perrin, Leslie, & Boat, 2016). Parents are likely to seek behavioral guidance from their child’s primary care provider (PCP) over other options (Taylor, Moeller, Hamvas, & Rice, 2013), and behavioral concerns are the primary reason for 15–20% of all pediatric visits (Williams, Klinepeter, Palmes, Pulley, & Foy, 2004). Unfortunately, the opportunity to provide behavioral guidance is primary care is often unfulfilled. Parents of young children consistently report their need for information on behavioral topics such as discipline, toilet training, and sleep is unmet in primary care settings, and parents with lower socio-economic status are especially likely to report unmet needs (Combs-Orme, Holden Nixon, & Herrod, 2011; Olson et al., 2004; Regalado, Larson, Wissow, & Halfon, 2010; Schuster, Duan, Regalado, & Klein, 2000).

Researchers have developed several promising methods of delivering parenting interventions to address the unmet need for behavioral guidance in pediatric primary care, including augmentative training for PCPs, integration of behavioral health specialists, and multi-media based interventions (Glascoe & Trimm, 2014; Shah, Kennedy, Clark, Bauer, & Schwartz, 2016). Each of these strategies has shown promise in clinical trials, but their real-world impact appears to be limited by suboptimal permeation into the intended patient populations. A recent systematic review of parenting interventions in primary care found wide variability in parents’ engagement with available interventions and identified a critical need to better understand factors that influence intervention uptake by families (Brown, Bignall, & Ammerman, 2018). This finding is especially pertinent given that increased access and engagement with interventions is one of the foremost goals of integrating behavioral services into the primary care environment (Hodgkinson, Godoy, Beers, & Lewin, 2017; Leslie et al., 2016).

Study of patient and caregiver preferences holds potential for improving the uptake and efficiency of health interventions (Ostermann, Brown, de Bekker-Grob, Mühlbacher, & Reed, 2017), and Sanders and Kirby (2012) have described how understanding and incorporating the perspectives of parents as consumers is integral to developing and disseminating evidence-based parenting programs. Parents’ preferences and priorities for consuming behavioral guidance via primary care may be considered in at least three domains: the intervention target (i.e., the child behaviors of interest), the approach (i.e., the mechanism for affecting change), and the mode of delivery (i.e., the format in which the intervention is provided). To date, assessment of parental attitudes in these domains have generally been limited to either post-intervention satisfaction ratings of a single intervention without comparator or cross-sectional surveys about individual modes of delivery (e.g., Combs-Orme et al., 2011; McGoron & Ondersma, 2015; Metzler, Sanders, Rusby, & Crowley, 2012; Riley, Freeman, & Marshall, 2016; Scholer, Hudnut-Beumler, & Dietrich, 2010). While useful, such research does not evaluate how parents’ consumer preferences impact initial intervention engagement or behavioral outcomes, nor does it assess how those preferences are related to other variables of interest (e.g., socioeconomic status or symptom severity).

Currently, there is no validated measure of parents’ consumer preferences for primary care-based behavioral intervention for use across studies, populations, or interventions. Such a measure would aid behavioral intervention research at the development, evaluation, and implementation stages by providing a valid method for prospectively assessing parents’ attitudes towards potential interventions. For example, researchers could capture the intervention preferences of a given sub-population of parents in order to engineer interventions that meet those preferences. Quantification of parents’ consumer preferences would also facilitate investigations into the factors that impact treatment uptake by parents and the comparative effectiveness of preference-tailored versus non-tailored approaches. Beyond research, a measure of parental preferences for behavioral guidance may hold clinical utility. Given the growing availability of various parenting interventions in various formats (Brown, Raglin Bignall, & Ammerman, 2018; Peacock-Chambers, Ivy, & Bair-Merritt, 2017), health care organizations that are considering adopting an intervention could sample their patient population to determine which approach is most well-matched their particular needs. At the patient level, provision of patient- and family-centered care involves incorporating family preferences into the treatment plan (Committee on Hospital Care & Institute for Patient- and Family-Centered Care, 2012), so a measure of those preferences would hold potential for facilitating shared-decision making and enhancing the personalization of care.

To advance the study of parenting interventions in pediatric primary care, we created the BIPS (Behavioral Information Preferences Scale), a measure parents’ priorities and preferences for behavioral health information in primary care. The goals of this study were to develop the BIPS and evaluate its initial psychometric properties.

Methods

Participants and procedures

We recruited parents/caregivers of children aged 18 months to 5 years 11 months through five pediatric primary care clinics in the Pacific Northwest United States. We focused on this age range, because of the growing emphasis on early childhood parenting interventions in primary care as a method to improve developmental outcomes (Garner et al., 2012; Peacock-Chambers et al., 2017). Parents were eligible to participate if they attended a clinic visit with a child in the specified age range and spoke English or Spanish. Two clinics (one urban, one suburban) were affiliated with an academic medical center and three (one urban, one suburban, one rural) were independent practices. Parents were recruited through study flyers distributed by clinic staff and electronic kiosks in clinic waiting rooms. Participating parents completed study measures in their preferred format (electronic or paper). Participants received a $25 gift card upon completion of measures. All methods were approved by the Human Subject Institutional Review Board at Oregon Health & Science University (# 16468).

Instrument development

We developed the BIPS as part of a larger project examining parents’ attitudes towards behavioral services in primary care (Riley et al., 2019). We were interested in creating a measure of parents’ interest in specific behavioral topics, the perceived helpfulness of behavioral supports, and potential modes of intervention delivery. A community stakeholder advisory panel consisting of two pediatricians, one primary care psychologist, one parent of young children, and one parent advocate (i.e., a provider of professional support and advocacy for parents of child with special health care needs in medical and educational systems) advised on the development of the BIPS in order to verify the content was meaningful and acceptable for administration. To create an initial item pool, we reviewed several measures of child behavior symptoms (Achenbach & Edelbrock, 1983; Eyberg & Pincus, 1999; Jellinek et al., 1988), literature pertaining to parents’ behavioral concerns in primary care (Combs-Orme et al., 2011; Olson et al., 2004; Schuster, Duan, Regalado, & Klein, 2000), and reviews of relevant interventions (Glascoe & Trimm, 2014; Leslie et al., 2016; Shah et al., 2016). We attempted to keep the item pool relatively small, because the advisory panel stressed the importance of limiting participant burden. The initial item pool was reviewed by the advisory panel and a group of content-expert pediatric psychologists (N=8) in order to generate additional items for consideration and to establish consensus around item groupings. Each of the experts independently reviewed the generated items in order to evaluate whether they were essential to their assigned sections. Ratings were summarized as content validity ratios (CVR), a statistical measure of consensus that ranges from −1.00 to +1.00 (Lawshe, 1975). With an expert panel of eight members, a minimum CVR of +.75 is required to meet a 0.05 significance level, meaning agreement by and least six of our content experts; we only retained items that met this threshold.

The resulting draft instrument consisted of three sections: Behavior Topics (18 items), Intervention Approach (6 items), and Delivery Methods (15 items). Each item prompted parents to rate either the importance of a behavioral topic, helpfulness of an intervention approach, or interest in a delivery modality as part of their child’s care on a five-point rating scale (1 = Not at all, 5 = Very). We recruited nine consecutive parents at one of the clinics to complete a cognitive interviewing task (Peterson, Peterson, & Powell, 2017). This consisted of prompting each parent to read the BIPS and express their perception of each item’s intent. Cognitive interviewing resulted in minor wording changes to items and section prompts to ensure clarity. English items were translated in to Spanish using a professional translation service. The entire measure can be viewed as a supplement.

Validation measures

As there are no existing measures of parents’ preferences for primary care-based behavioral intervention, selecting comparative measures for investigating construct related validity is challenging. Andersen’s Model of Health Care Utilization posits that individual “need factors” are important determinants of health services utilization (Andersen, 1995). We reasoned that parent-reported child behavior problems and parenting self-efficacy in behavioral domains are likely need factors for behavioral guidance, so we included measures of each of these constructs.

The Eyberg Child Behavior Inventory (ECBI; Eyberg & Pincus, 1999) is a common measure of child behavioral symptoms that possesses strong psychometric properties. The 36 items yield scores on both an Intensity Scale (ECBI-IS; α = .94 in this sample), which measures the frequency of problem behaviors, and a Problem Scale (ECBI-PS; α = .94), which measures how problematic parents find child behaviors.

The Self-Efficacy for Parenting Tasks Index-Toddler Scale (SEPTI; Van Rijen, Gasanova, Boonstra, & Huijding, 2014) is a reliable 26-item instrument that measures parents’ self-efficacy in four domains: nurturance (α = .88), discipline (α = .74) , play (α = .80) , and routine. (α = .79). To accommodate a slightly wider age range, we modified the SEPTI by changing the word “toddler” to “child.”

Previous research indicates that parents who identify as racial/ethnic minorities and those who are socioeconomically disadvantaged report significantly higher unmet needs for behavioral counseling in primary care (Combs-Orme et al., 2011). As such, we presumed that higher importance of behavioral topics and greater interest in a variety of delivery methods by racial-ethnic/minorities and persons with lower income would be indicative of construct validity for a measure of parental preferences for primary care behavioral health. We therefore included race/ethnicity status and household income as additional validation variables.

Analyses

We conducted factor analyses with each section of the BIPS using principal components analysis with varimax rotation (because there was no a priori reason to expect an orthogonal solution) to evaluate the preliminary factor structures. We utilized principal components analysis rather than exploratory factor analysis, because 1) we imposed some structure on the data a priori by dividing it into sections, 2) preliminary analyses indicated a high level of correlation amongst items in each section, and 3) our goal was to summarize observed responses, not identify latent constructs. We followed standard procedures of retaining factors (DeVellis, 1991). Items with a factor loading score of at least 0.40 were retained, as is conventional in health research (Lawlor, Ebrahim, May, & Smith, 2004). The factor extraction process identified how many summary scores were needed to adequately reflect the variability measured by the items within each section of the BIPS and which items to assign to the different summary scores.

To determine the degree to which all items assigned to a factor were measuring the same underlying construct, we calculated Cronbach’s alpha and item-total correlations.

We calculated correlation coefficients to test the hypotheses that lower parenting self-efficacy (particularly in the domain of discipline), higher child behavior problems, and lower household income would be associated with greater interest in behavioral topics, as well stronger interest in alternative modes of behavioral intervention delivery. To test the hypotheses that parents who identified as racial/ethnic minorities would report higher importance of behavioral topics, greater perceived helpfulness of interventions strategies, and higher interest in alternative delivery modalities, we carried our t-tests and Mann-Whitney U tests for continuous and ordinal variables, respectively.

Results

Study information was provided to 547 parents, of whom 454 were screened for eligibility. Amongst those who were eligible, 396 participants (91%) completed the BIPS, and 369 (85%) also completed the ECBI and SEPTI. Table 1 summarizes the demographics of the participants. We conducted all analyses using data from English-language measures only, as well as English and Spanish measures combined. There were no significant differences in outcomes based on the inclusion of participants who completed the measures in Spanish (n=10), so herein we report the pooled results.

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
Parenting Situation, %
Single Parenting	11
Couple Parenting, same household	85
Co-parenting, separate households	5
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
Number of children, M (SD)	2.1 (1.1)
Child*
Age, years, M (SD)	3.5 (1.3)
Female sex, %	46
First born child, %	62

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

Factor structure

Each section of the BIPS was separately evaluated by factor analysis as a way to explore the measure’s construct validity (an estimate about the appropriateness of inferences drawn from test scores regarding individual standings on a construct). An iterative approach was utilized in that a single factor solution was first explored, followed by two, three and four factor solutions. Exploratory solutions beyond two factors for the Behavior Topics section, two factors for the Intervention Approach section, and three factors for the Delivery Methods section produced factors that were conceptually confusing with limited additional variance captured; therefore, these solutions were deemed irrelevant.

The best fit for the BIPS Behavior Topics section was a strong two-factor solution (i.e., two factors had an eigenvalue >1) that accounted for 70% of the total explained variance of scores. The factors that emerged corresponded conceptually to constructs of 1) Conduct & Emotions and 2) Healthy Habits. The best solution for the BIPS Intervention Approach section was a strong rotated two-factor solution that accounted for 74% of total explained variance. Conceptually, those factors corresponded to constructs of 1) Change Strategies and 2) Psychoeducation. The BIPS Delivery Methods section showed a strong three-factor solution. The rotated three-factor solution accounted for 60% of the total explained variance of scores. The factors that emerged corresponded conceptually to constructs of 1) Multimedia Resources; 2) Auxiliary Care; and 3) Usual Care. Of note, the Usual Care factor consisted of only one item, “Talking to my child’s doctor during normal visits.” Traditionally, item reduction strategies are justified when a factor emerges with a single item. In this case, we elected to retain the item as its own factor, because it holds conceptual relevance and captures what is presumably the most common form of behavioral guidance.

Table 2 presents the loading scores of items within each section of the BIPS. Table 3 indicates the factor to which each item was designated.

Table 2.

Principal Component Analysis Total Variance Summary: Behavior Topics and Delivery Methods Sections of the BIPS

	Rotation Sums of Squared Loadings
Component	Total	% of Variance	Cumulative %
Behavior Topics Section
(1) Conduct & Emotions	7.978	44.322	44.322
(2) Healthy Habits	4.634	25.747	70.068
Intervention Approach Section
(1) Behavior Change	2.776	46.262	46.262
(2) Psychoeducation	1.664	27.728	73.991
Delivery Methods Section
(1) Multimedia Resources	4.981	33.203	33.203
(2) Auxiliary Care	2.811	18.743	51.947
(3) Routine Care	1.258	8.388	60.335

Table 3.

Principal Component Correlation Matrices: Behavior Topics, Intervention Approach, and Delivery Methods sections of the BIPS

Items	Component
	1	2	3
Behavior Topics Section
Teaching children to follow household rules	.833		-
Teaching children to listen to and follow directions	.848		-
Teaching children not to be aggressive towards others, like hitting, kicking, and biting	.832		-
Teaching children not to be destructive with things, like throwing or breaking objects	.859		-
Teaching children to play nicely with others, like sharing and taking turns	.832		-
How to handle temper tantrums	.719		-
Managing misbehavior in public places, like grocery stores or restaurants	.822		-
Teaching children to stay on-task and focus	.577		-
Teaching children to be honest and truthful	.701		-
Helping children with difficult emotions, like feeling sad, frustrated, angry, or scared	.663		-
Helping children transition from one activity to another without problems	.603		-
What to do if children say inappropriate things, like insults or bad words	.667		-
Helping children learn how to calm themselves when upset	.689		-
Teaching children to use the toilet properly		.559	-
Helping children have appropriate limits with toys, TV, computers, or other activities		.626	-
Helping children fall asleep and sleep through the night on their own		.705	-
Managing picky eating and helping children eat healthfully		.778	-
How to break bad habits like thumb sucking or nail biting		.776	-
Intervention Approach Section
Setting up appropriate expectations for child behaviors	.739		-
Rewarding and showing approval for good behaviors	.887		-
Creating consequences for misbehavior	.833		-
Building a positive parent-child relationship	.672		-
Managing the stress of parenting		.614	-
Knowing what behaviors are typical or unusual for your child's age		.918	-
Delivery Methods Section
Paper handouts from my child’s doctor	.734
Videos that could be watched on a computer, TV, phone, or other device	.764
Websites with general information	.858
Online programs you can sign up for	.816
Books	.669
Mobile apps for smartphones or tablets	.740
Messages through my child’s patient portal or medical record	.575
Messages through social media, like Facebook or Twitter	.602
Podcasts (audio files that can be downloaded to a mobile device or computer)	.642
Talking to my child’s doctor during a separate visit focused on behaviors		.799
Talking to a behavioral expert as part of my child’s normal medical visits		.735
Talking to a behavioral expert as part of one or more separate visits		.875
Talking to a behavioral expert over the phone		.515
Attending group classes or seminars with other parents		.558
Talking to my child’s doctor during normal visits			.787

Internal consistency

Within the Behavior Topics section, Cronbach’s alpha was .96 for Conduct & Emotions and .84 for Healthy Habits. Within the Intervention Approach section, Cronbach’s alpha was .87 for the Change Strategies and .63 for Psychoeducation. The Multimedia Resources and Auxiliary Care factors of the Delivery Methods section produced alphas of .90 and .79, respectively. The single-item Usual Care factor was not assessed for internal validity.

Construct related validity

Subscale scores were generated by totaling the sum of items in each derived factor. Table 4 displays a correlation matrix of scores from the BIPS and construct validity measures.

Table 4.

Correlation matrix for BIPS, ECBI, and SEPTI scores (N=369).

Measure	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. BIPS-CE	-	.79**	.67**	.47**	.39**	.36**	.42**	.02	.13*	−.04	−.15**	.01	−.16**	−.14**
2. BIPS-HH		-	.66**	.51**	.38**	.38**	.51**	.01	.09	.01	−.09	.00	−.08	−.10
3. BIPS-CS			-	.67**	.39**	.41**	.41**	.02	.09	.02	−.13*	.01	−.05	−.09
4. BIPS-PE				-	.33**	.34**	.39**	−.01	.05	.06	−.04	−.02	.01	.05
5. BIPS-AC					-	.49**	.28**	.20**	.12*	−.14**	−.37**	−.13*	−.15**	−.14**
6. BIPS-MR						-	.23**	.16**	.09	−.09	−.16**	−.04	−.18**	−.12*
7. BIPS-UC							-	.01	−.01	.09	−.02	−.02	.01	−.06
8. ECBI-intensity								-	.63**	−.26**	−.41**	−.32**	−.35**	−.06
9. ECBI-problem									-	−.26**	−.37**	−.33**	−.34**	−.07
10. SEPTI-nurturance										-	.30**	.46**	.47**	−.06
11. SEPTI-discipline											-	.37**	.49**	.10
12. SEPTI-play												-	.42**	.02
13. SEPTI-routine													-	.09
14. Income														-

Note: CE = Conduct and Emotions, HH = Healthy Habits, BC = Behavior Change, PE = Psychoeducation, AC = Auxiliary Care, MR = Multimedia Resources, UC = Usual Care. Coefficients for BIPS-Usual Care and Income are Spearman’s rho; all others are Pearson’s r.

^*p < .05.

^**p ≤ .01.

For the BIPS Behavior Topics section, the Conduct and Emotions subscale was significantly correlated in the hypothesized direction with the ECBI-PS, the SEPTI Discipline and Routine subscales, and household income. Conduct and Emotions scores were significantly higher for minority parents than non-Hispanic White parents, t(390) = 5.55, p ≤ .001. Healthy Habits scores were not significantly correlated with the other measures, but minority parents again rated topics more important on average t(390) = 4.06, p ≤ .001.

Within the BIPS Intervention Approach section, the Change Strategies subscale was negatively correlated with SEPTI Discipline scores, but no other significant relationships were observed. Minority parents scored significantly higher on the Change Strategies subscale; t(390) = 2.01, p = .05; but there was no difference on the Psychoeducation subscale, t(390) = .74, p = .46.

For the BIPS Delivery Methods section, the Auxiliary Care subscale was significantly correlated in the expected direction with all ECBI and SEPTI scores, as well as household income. Minority parents reported significantly higher interest in Auxiliary Care t(390) = 1.96, p = .05. The Multimedia Resources subscale was significantly correlated with the ECBI-IS, Discipline and Routine scores of the SEPTI, and household income. Minority parents scored significantly higher on the Multimedia Resources scale, t(390) = 3.26, p =.001. The single-item Usual Care subscale was not correlated with any other measures, nor did parents differ significantly by minority status, U = 15878.00, p =.24.

Discussion

The purpose of this investigation was to develop and evaluate the psychometric properties of a measure of parents’ consumer preferences for behavioral guidance in primary care, the BIPS. Principal component analyses yielded a strong two-factor solution that explained 70% of the variance of the 18-item Behavior Topics section, a two-factor solution that explained 74% of the variance of the 6-item Intervention Approach section, and a three-factor solution that accounted for 60% of variance in the 15-item Delivery Methods section. The derived factors showed excellent internal consistency with the exception of the Psychoeducation subscale, which was marginal. The emerged factors are conceptually coherent: The Conduct and Emotions factor of the Behavior Topics section appears to measure parents’ interest in topics related to externalizing and internalizing behavior problems, whereas the Healthy Habits factor assesses interest in more general developmental skills and routines. Similarly, the Change Strategies items of the Intervention Approach section focus on methods of altering child behavior, whereas the Psychoeducation items focus on understanding or coping with child behavior, rather than modification. The Delivery Methods subscales were also delineated in a logical fashion: The Usual Care item captures interest in behavior information during routine clinical care with a child’s PCP, whereas the Auxiliary Care items pertain to augmentative care experiences that involve direct interaction with other people, and the Multimedia Resources section items focus on text or digital-based information sources. Internal consistency and logical coherence are strengths of the BIPS.

The pattern of associations of the Behavior Topics subscales with established measures of child behavior problems and parenting self-efficacy provides some evidence of construct validity. The Conduct and Emotions subscale was significantly correlated with the ECBI-PS scale, but not the ECBI-IS. This is consistent with a measure of parent preference, because the ECBI-IS measures the frequency of child behaviors, whereas the ECBI-PS evaluates parental perception of whether those behaviors are problematic. The Conduct and Emotions subscale was also significantly correlated with parenting self-efficacy in the Discipline and Routine domains. Given that establishing consistent routines and effective discipline practices are primary strategies for preventing and addressing behavioral problems, this finding adds to the evidence for convergent validity of the Conduct and Emotions subscale. Parents with lower income and racial/ethnic minority status scored more highly on the Conduct and Emotions scale, consistent with previous research documenting higher unmet needs in these groups. Overall, there was a lack of evidence for convergent validity for the Healthy Habits subscale. It was surprising that the Healthy Habits subscale did not correlate with any SEPTI subscales, particularly the Routine subscale. In retrospect, a lack of correlation with the ECBI subscales is less surprising, given their focus on conduct problems. It may be that there is little variance in the population regarding Healthy Habits topics, in that they are generally of high importance to most parents.

The Delivery Methods section subscales also demonstrated associations with the ECBI and SEPTI that indicated construct validity. The Auxiliary Care and Multimedia Resources domains were generally positively correlated with child behavior problems, negatively correlated with aspects of parenting self-efficacy, and negatively correlated with income. Racial/ethnic minority parents reported higher interest on both subscales. Parents’ interest in discussing behavioral topics with their child’s doctor during routine care was not related to any other assessed variable. The existing literature indicates that a wide swath of parents appreciate discussing behavioral topics with their child’s PCP, so it is unsurprising that the Usual Care subscale/item did not vary commensurate with other measures.

The Intervention Approach section showed the weakest evidence of construct validity. While the Change Strategies subscale was correlated with parent self-efficacy in the domain of discipline in the expected direction, the Psychoeducation subscale did not demonstrate any significant relationships, so there is little evidence of validity.

Overall, the BIPS yielded pattern of associations suggesting the subscales may represent different levels of care, such that preferences for higher tiers of care are associated with indicators of need, but lower tiers are not. The Healthy Habits and Usual Care subscales represent informational content and care delivery that is consistent with routine well-child care, which likely appeals to a wide swath of parents. It is thus logical these subscales demonstrated fewer significant relationships. The Emotions and Conduct, Auxiliary Care, and Multimedia Resources subscales represent information and delivery methods above and beyond typical care. These subscales showed more relationship to indicators of need, including child behavior symptoms, lower parenting self-efficacy, minority status, and lower household income.

Across BIPS subscales, statistically significant associations with validation measures were generally small in magnitude. This was expected, as even parents with relatively well behaved children may be highly interested in behavioral topics. For example, aggression is nearly universal in early childhood, so parents may have significant interest in this topic in the absence of pathology. Similarly, parents of children with clinically significant behavior problems may not necessarily be interested in services to ameliorate those problems as evidenced by low engagement and retention rates of parenting programs (Brown et al., 2018).

Implications and Potential Uses

Notably, the BIPS Behavior Topics and Delivery Methods sections map relatively well onto the existing spectrum of behavioral interventions that have been studied in primary care. For example, reviews of developmental-behavioral interventions in primary care identify PCP-delivered verbal advice (Usual Care), use of written and technology-based information systems (Multimedia Resources), and the integration of behavioral health providers or specially trained medical providers (Auxiliary Care) as commonly employed methods (Brown et al., 2018; Glascoe & Trimm, 2014; Shah et al., 2016). Each of these approaches possesses strengths and weaknesses, and determining which strategy is most appropriate for any given child or family is a nuanced process. Inclusion of the BIPS in clinical research could shed light on the role of parental preferences in that process in several ways. For example, the BIPS could be used to characterize the consumer preferences of a given population of parents, helping researchers to design stakeholder-informed interventions. The BIPS can also be used to investigate factors that influence parents’ preferences for behavioral care. Brown et al. have speculated that parents’ engagement in available parenting interventions may be associated with environmental risk factors for the development of behavior problems. If parent preferences as captured by the BIPS are associated with known risk factors (e.g., parent mental health, SES, corporal punishment), it would hold implications for maximizing the reach of primary care interventions into their intended populations by tailoring to the preferences of those most at-risk. The BIPS could also be used in clinical trials to assess the degree to which parental preferences at baseline influence treatment engagement and clinical outcomes.

In addition to research applications, the BIPS may hold clinical utility. Primary care practices that are engaged in behavioral care quality improvement efforts could assess parents’ preferences within their patient population in order to identify areas of need and inform resource planning. We do not recommend the BIPS for use as a scale with individual patients in its initial form, because we have not yet established population norms or clinical cut-off scores. However, the BIPS could be used as an inventory to guide clinical decision-making in a manner similar to pre-visit questionnaires included in the Bright Futures Tool and Resource Kit (Hagan, Shaw, & Duncan, 2008). For example, the items of the Behavioral Topics section items could be distributed with pre-visit paperwork in order to identify the topics parents are most interested in discussing. Prioritizing the focus of guidance in this manner could be especially important for behavioral topics, as they are disproportionately time-consuming and often left unaddressed (Combs-Orme et al., 2011; Cooper, Valleley, Polaha, Begeny, & Evans, 2006).

Limitations and Future Directions

This study possessed several limitations, most significantly a smaller than ideal initial item pool. This was the result of both the secondary nature of measure development in the context of a larger project, as well as concerns about participant burden conveyed by our stakeholder advisory panel. The 6-item Intervention Approach section was particularly limited, and this may explain why that section demonstrated the weakest psychometric properties. Another limitation was the inclusion of relatively few comparison validity measures. Inclusion of measures of general child development or parenting stress may have been useful. Ideally, future investigations of the BIPS will allow for expanded item pools, additional validation measures, and more robust psychometric evaluation. Some additional limitations were associated with the participant sample. While we sampled from primary clinics in several communities, all were within the Pacific Northwest region of the United States and it is unclear how the BIPS would perform in broader samples. Further, while inclusion of a Spanish version is a positive, too few were recruited to adequately evaluate the Spanish-only version of the BIPS. In future investigations, larger and more heterogeneous samples would be useful for determining the consistency of the BIPS factor structure across subgroups and establishing age-based norms. Ultimately, predictive validity may be the best test of the BIPS and similar measures, and future studies should determine whether the BIPS can prospectively predict parents’ level of engagement with available behavioral interventions. Finally, while the BIPS was created specifically for primary care, integration of behavioral services into inpatient and outpatient subspecialty settings is also needed (Samsel, Ribeiro, Ibeziako, & DeMaso, 2017). The BIPS could potentially be adapted for particular clinical populations and settings by including condition-specific potential topics (e.g., non-medical strategies to manage pain) and delivery methods (e.g., post discharge phones consultation with a behavioral specialist).

Conclusion

This study provides initial evidence for the validity of some components of the BIPS, the first measure of parents’ priorities and preferences for behavioral guidance in primary care. The Behavior Topics and Delivery Methods sections of the BIPS in particular displayed robust factor structures, strong internal consistency, and logically coherent associations with child behavior symptoms, parenting self-efficacy, and sociodemographic characteristics. These scales may aid research focused on the design or dissemination of behavioral parenting interventions in primary care and inform clinical initiatives. Due to marginal psychometric properties and a lack of evidence for construct validity, the Intervention Approach items are not recommended for use as subscales at this time; however, individual items of these scales may provide useful descriptive information in some instances and could be included as a supplement to the other subscales.