A Parent Coach Model for Well-Child Care Among Low-Income Children: A Randomized Controlled Trial
OBJECTIVE: The goal of this study was to examine the effects of a new model for well-child care (WCC), the Parent-focused Redesign for Encounters, Newborns to Toddlers (PARENT), on WCC quality and health care utilization among low-income families.
METHODS: PARENT includes 4 elements designed by using a stakeholder-engaged process: (1) a parent coach (ie, health educator) to provide anticipatory guidance, psychosocial screening and referral, and developmental/behavioral guidance and screening at each well-visit; (2) a Web-based tool for previsit screening; (3) an automated text message service to provide periodic, age-specific health messages to families; and (4) a brief, problem-focused encounter with the pediatric clinician. The Promoting Healthy Development Survey–PLUS was used to assess receipt of recommended WCC services at 12 months’ postenrollment. Intervention effects were examined by using bivariate analyses.
RESULTS: A total of 251 parents with a child aged ≤12 months were randomized to receive either the control (usual WCC) or the intervention (PARENT); 90% completed the 12-month assessment. Mean child age at enrollment was 4.5 months; 64% had an annual household income less than $20 000. Baseline characteristics for the intervention and control groups were similar. Intervention parents scored higher on all preventive care measures (anticipatory guidance, health information, psychosocial assessment, developmental screening, and parental developmental/behavioral concerns addressed) and experiences of care measures (family-centeredness, helpfulness, and overall rating of care). Fifty-two percent fewer intervention children had ≥2 emergency department visits over the 12-month period. There were no significant differences in WCC or sick visits/urgent care utilization.
CONCLUSIONS: A parent coach–led model for WCC may improve the receipt of comprehensive WCC for low-income families, and it may potentially lead to cost savings by reducing emergency department utilization.
- CAB —
- community advisory board
- CAHMI —
- Child and Adolescent Health Measurement Initiative
- ED —
- emergency department
- EHR —
- electronic health record
- ES —
- effect size
- PARENT —
- Parent-focused Redesign for Encounters, Newborns to Toddlers
- PC —
- parent coach
- WCC —
- well-child care
- WVP —
- well-visit planner
What’s Known on This Subject:
The current structure of well-child care (WCC) in the United States cannot support the vast array of preventive needs of families. There are few evidence-based comprehensive models for WCC that can serve as sustainable alternatives to our current delivery of WCC.
What This Study Adds:
This parent coach–led WCC intervention demonstrated robust improvements in the receipt of WCC services (eg, psychosocial screening, health education and guidance, developmental guidance and screening), experiences of care, and reduced emergency department visits.
Well-child care (WCC) visits from age 0 to 3 years are a key element of pediatric primary care; for many families, these visits provide the only opportunity to identify and address important social, developmental, behavioral, and health issues before a child begins school.1,2
In Donabedian’s Quality Framework, health care outcomes are determined by the structure and processes of care.3,4 The structure→process→outcome model, further developed by Starfield,5 serves as the conceptual model for this study by demonstrating how the structure and processes of a system of care affect the eventual outcomes.3,4 Our current WCC structure has difficulty supporting the vast array of preventive care needs among low-income families. Studies have shown that these visits do not provide adequate preventive and developmental services and that most parents leave the visit with unaddressed psychosocial, developmental, and behavioral concerns.6–11 Key structural problems in WCC include: (1) reliance on high-level clinicians (pediatricians, family physicians, and nurse practitioners) for basic, routine WCC services; (2) limitation to a clinician-directed well-visit for the wide array of education and guidance services in WCC; and (3) lack of a systematic, patient-driven method for visit customization to meet families’ needs.1,2,12–15
Various tools and strategies to redesign the structure of WCC have been proposed and studied; however, there are few evidence-based comprehensive models of care to serve as feasible and sustainable alternatives to the current WCC structure in most small community practices.16,17 Our comprehensive model (ie, the Parent-focused Redesign for Encounters, Newborns to Toddlers [PARENT]) was developed to address these key structural deficiencies in our current WCC system by decreasing reliance on the physician for routine WCC and shifting many WCC services to a health educator.18 Our objective was to test the effectiveness of the PARENT intervention in providing comprehensive, nationally recommended WCC services, while optimizing WCC, urgent care, and emergency department (ED) utilization among low-income families.
Intervention Components and Process
The intervention was designed by using a 12-month structured process in partnership with 2 independent pediatric practices (practices A and B comprising 6 and 2 clinicians, respectively) that serve predominantly Medicaid-insured populations within Los Angeles County. Key stakeholders from the practice sites, including parent representatives, pediatricians, and medical assistants, formed a community advisory board (CAB). Over a series of meetings, the CAB designed the new WCC system by using qualitative data from stakeholder interviews,19–22 a systematic literature review,16 and an expert panel process.18 The CAB was involved in all aspects of intervention development (including parent coach [PC] curriculum, training, and ongoing education), implementation, and testing, and the CAB continued regular meetings throughout the trial to monitor progress of the intervention.
PARENT includes the following elements: (1) a PC (ie, health educator) to serve as the primary provider of anticipatory guidance, psychosocial screening and referral, and developmental/behavioral guidance and screening at each well-visit; (2) a Web-based tool to customize the visit to parents’ needs and facilitate previsit screening; (3) an automated text message service for periodic, age-specific health messages to families; and (4) a brief, problem-focused encounter with the pediatric clinician.
The PC was a Master’s level Spanish/English-bilingual health educator with no experience in child health. She received 80 hours of training. This training consisted of self-directed learning based on Bright Futures Guidelines, Third Edition1; relationship building with community organizations near each clinical site; mock visits; and pediatrician-observed visits with feedback at the practices. During the study, the PC split her time between the 2 clinical sites, seeing intervention patients 2 days per week at each site. The fifth day of the week was spent as “Parent Coach Office Hours,” during which she conducted parent follow-up calls and visit reminders and was available to answer parents’ preventive care–related questions.
The well-visit planner (WVP) was developed by the Child and Adolescent Health Measurement Initiative (CAHMI) and is a Web-based parent tool that allows parents to: (1) select priorities for their child’s well-visit; (2) complete screening questions; and (3) receive anticipatory guidance and a tailored visit guide (a publicly available version of the WVP is available at www.cahmi.org/projects/wvp).23,24 We worked with CAHMI to translate the tool into Spanish; adapt psychosocial questions to meet the needs of our urban, low-income population; and enable automatic upload and display of WVP results in the well-visit encounter by using the practices’ electronic health record (EHR).
The WVP asks parents a set of standardized, previsit questions anchored to the Bright Futures Guidelines, Third Edition,1 and tailored for each visit (eg, 4-month visit). All content was specified and tested by CAHMI in collaboration with experts and parents. Where available, validated question sets assessing child and family health and risks were used (eg, Patient Health Questionnaire–2 for depression25). At the time of this randomized controlled trial, the WVP had been tested among pediatric practices in Portland, Oregon, and was available for well-visits for children ages 4 through 36 months.,23
Text Message Service
The library of text messages were adapted from Healthy-TXT, a proprietary text messaging service (Healthy-TXT, LLC, Chicago, IL), to the needs of the practice sites. Parents were offered the text messaging service at study enrollment; parents texted their child’s birth date and clinic name to a designated number and received a welcome message from their clinic, with subsequent biweekly messages throughout the study. Messages focused on age-appropriate anticipatory guidance, health education, and reminders for well-visits. Most messages included a link to an educational Web site (eg, healthychildren.org) with a video or written information on that specific topic or included the clinic’s telephone number for visit scheduling. At any time, parents could text STOP to end the service. Sample messages are available from Healthy-TXT upon request (healthy-txt.com).
Upon arrival, the child was registered, weighed, and measured by the medical assistant, and then roomed (ie, usual care process). For intervention patients, the PC would come into the room and use the data uploaded from the WVP into the EHR to: (1) discuss parent-selected priorities for anticipatory guidance topics (generally up to 3 topics); (2) review any red flags from the WVP results; and (3) conduct developmental screening (at the 9- and 15-month well-visits) or autism screening (at the 18- and 24-month well-visits). If the WVP results indicated a need for community referrals for psychosocial needs, the PC would provide those referrals at this time. Sessions with the PC lasted ∼20 minutes. The PC documented her visit with the family in the EHR, highlighting at the top of the encounter page the results of developmental and behavioral screening and any issues that the clinician needed to review.
After the PC completed her time with the family, the clinician would enter the examination room, conduct a physical examination, and review the PC notes within the EHR. The clinician would address any PC findings that needed further clinical investigation (eg, concern for speech delay on developmental milestones history) and any final parent concerns.
Control patients received usual care (ie, clinician well-visits scheduled in 15- or 20-minute slots, for practice A and B, respectively). Most WCC services (anticipatory guidance, developmental guidance) were performed during the clinician visit as usual care at the practice sites. During the study period, practice A implemented quality improvement efforts that consisted of routine developmental screening administered by medical assistants at the 9- and 15-month visits.
At both practices, all clinicians (6 clinicians at practice A and 2 clinicians in practice B) and support staff (medical assistants) participated in visits for intervention and control participants. The study was approved by the University of California Los Angeles Institutional Review Board.
Parents or legal guardians arriving at 1 of the clinical sites (practices A and B, described earlier) for an acute or well-visit with infants ≤12 months were approached by a UCLA research assistant or clinic staff to discuss participation in a WCC redesign study. Interested parents were screened for eligibility and consented either in person or over the telephone through a research assistant. Participants (henceforth “parents”) were ineligible if they: (1) did not speak English or Spanish; (2) were <18 years of age; (3) planned to move outside of Los Angeles County or change primary care providers within the next 12 months; (4) had limited availability for Monday through Friday visit scheduling; or (5) were currently employed by the participating pediatric practice.
Participants were randomized to a study group at enrollment by using a computer-generated random allocation sequence. The enrollment research assistant referred to the allocation sequence to determine group assignment only after the eligibility and consent process was complete. There was no blinding of participants or research staff.
Upon enrollment, parents completed a baseline survey conducted by a telephone or in-person interview. Demographic data were collected on the infant, parent, and household.
At 12 months’ postenrollment, parents participated in a 25-minute interview that used the Promoting Healthy Development Survey–PLUS,7,26,27 a parent survey developed and tested by CAHMI to evaluate the quality and content of recommended WCC services. This survey is the telephone and in-person interviewer-administered version of the self-administered Promoting Healthy Development Survey.7 It is endorsed by the National Quality Forum and has been used by state Medicaid agencies, health plans, and pediatric practices and nationally through the National Survey of Early Childhood Health.28 It is available in English and Spanish at an eighth-grade reading level and has strong internal consistency (Cronbach’s α = 0.80).29 We also included 1 item from the Consumer Assessment of Healthcare Providers and Systems Clinician & Group Survey for overall parent rating of care.30
A brief survey was conducted at 6 months’ postenrollment to boost participant retention, as well as qualitative interviews of a 25% random sample of intervention parents at the end of the study. We present here the 12-month parent survey findings on the primary study outcomes.
Our primary outcome measures assessed the following: (1) receipt of preventive care services; (2) parent experiences of care; and (3) health care utilization (Table 1). Composite outcomes used multiple survey items and have been previously tested for internal consistency. Details of psychometric analyses and concurrent validity for the Promoting Healthy Development Survey can be found elsewhere.29
Before enrollment, we calculated that a total of 240 enrolled participants would provide 80% power to detect a small to medium effect size (ES)31 for the 12-month composite scores for receipt of preventive care services measures. This calculation assumed that 80% of enrolled participants (n = 192) would complete the study through 12 months.
All analyses were performed by using an intention-to-treat analysis to avoid bias from selective disenrollment from the intervention. Nondichotomous outcome measures were linearly transformed to a scale of 0 to 100 for ease of interpretation and averaged in groups to create composite scores. Differences were examined between the intervention and control groups on baseline characteristics, and on each outcome measure, by using t tests for continuous outcomes and Pearson’s χ2 tests for dichotomous outcomes; ES values were also reported when differences reached statistical significance. Because continuous outcome measures were not normally distributed, sensitivity analyses were conducted by using a nonparametric test (Wilcoxon rank-sum test) to examine differences between the intervention and control groups. An additional sensitivity analysis was conducted that assessed intervention effects, with adjustment for any characteristics that were found to be significantly different between the control and intervention groups at baseline.
A total of 251 parents were enrolled, randomized, and completed a baseline survey (June–December 2013); 226 (90%) parents completed the 12-month assessment (Fig 1). The only statistically significant difference between the intervention and the control groups on baseline characteristics was for parent-reported child health status. Fewer intervention parents than control parents selected “excellent” versus “good” to describe their child’s health (Table 2).
Mean child age at enrollment was 4.5 months. Seventy-seven percent of index children were Latino, and 18% were African American; 45% lived in households in which Spanish was the primary language; 53% had the highest household education level as high school or less; and 64% had an annual household income less than $20 000.
Intervention families had a mean ± SD of 3.3 ± 1.5 visits with the PC (range, 0–7). Nine intervention families had no encounters with the PC. Sixty-two percent of intervention parents used the text message service, and 84% used the WVP (almost exclusively at the practice site) during at least 1 PC-led well-visit.
Receipt of Preventive Care Services
Compared with the control group, intervention parents scored significantly higher on all receipt of preventive care services measures (Table 3): anticipatory guidance (mean score, 89.3 vs 77.4; ES, 0.49), health information (mean score, 96.3 vs 89.6; ES, 0.30), and psychosocial assessments for family risks (mean score, 93.3 vs 70.9; ES, 0.62) and tobacco, drug, and alcohol use (mean score, 98.3 vs 91.9; ES, 0.29). Intervention parents were more likely to report receiving a structured developmental screening (92.2% vs 81.1%; ES, 0.12) and having their developmental and behavioral concerns addressed (90.2% vs 73.8%; ES, 0.28).
Experiences of Care
Intervention parents had higher mean scores for helpfulness of care (91.3 vs 82.1; ES, 0.47) and family-centeredness of care (96.3 vs 92.4; ES, 0.30), and slightly higher scores for overall rating of care (94.5 vs 91.7; ES, 0.24).
Health Care Utilization
Intervention children were less likely to have ≥2 ED visits (10.4% vs 21.6%; ES, 0.47). Ten parents total (6 from the control group and 4 from the intervention group) reported a hospitalization for the index child (data not shown). There were no significant differences between the intervention and the control groups for the proportion of children up-to-date on well-visits or with ≥2 sick/urgent care visits.
Sensitivity analyses found no differences in the statistical significance of results at the 0.05 threshold when assessed by using Wilcoxon rank-sum tests. Adjustments for parent-reported child health status in linear and logistic regressions had negligible effects on outcomes (Supplemental Table 4).
A PC-led intervention designed to improve WCC for publicly insured infants and toddlers resulted in robust improvements in parent-reported quality of WCC and a substantial reduction in ED utilization. Among the preventive care services examined, the largest improvements were found in psychosocial screening for family risks, with a 23-point difference between the intervention and the control. Using Cohen’s classification,31 this ES could be described as medium. Unpublished data from the development of the WVP indicate that its use among clinicians in other settings also improves identification of psychosocial concerns.23
We do not have data to determine if intervention parents were more likely to use community resources for various psychosocial needs; however, the PC provided parents with community referrals based on identified needs from the WVP. Other interventions have been designed to help clinicians identify psychosocial needs32–35; some have demonstrated increased referral rates and parent enrollment into community resources.32
The PARENT intervention includes 4 main elements; the PC serves as the core of the intervention, with support from the WVP to guide the well-visit encounter and text messages to provide additional information between visits. The inclusion of nonphysicians as part of a team approach to care is widely recognized as a key component in preventive care quality improvement and WCC delivery system redesign.13,36–38
In a recent review of tools and strategies to improve WCC, we found evidence suggesting that inclusion of nonphysician providers into a health care team could improve receipt of anticipatory guidance and developmental and behavioral services.16,17,39–42 In Healthy Steps, a physician and child developmental specialist (typically a nurse, social worker, or early childhood educator) provide WCC in partnership. In a large, national, 3-year trial, intervention parents reported receiving more anticipatory guidance and more family-centered care and were more likely to be up-to-date on WCC.39 There were no statistically significant differences in hospitalizations or ED use in general, but intervention children did have slightly lower rates of ED utilization for an injury-related cause (9% vs 11%).
Similarly, the PARENT intervention uses a nonphysician provider as part of a team-based approach to WCC. The role of the PC, however, is not to enhance WCC as currently provided by the pediatrician43 but to instead serve as the primary provider of anticipatory guidance, psychosocial screening, and developmental/behavioral guidance and screening. Reliance on a physician as the primary provider of routine WCC services is thus reduced.
We found a significant effect of the intervention on ED utilization, with a 52% reduction in the proportion of children with ≥2 ED visits, which would represent a substantial savings in costs over a large population of children. The rate of ED utilization in our control group is similar to national data, in which 15.3% of Medicaid-insured children aged <6 years made ≥2 ED visits in the previous 12 months.44 We do not have data to explain why intervention parents made fewer ED visits; however, based on what is known regarding parents’ reasons for ED use, we can hypothesize.45–47 Increased focus on parental concerns and anticipatory guidance during well-visits may have reduced the need for potentially unnecessary ED visits or for injury-related ED use. In addition, the relationship with the parents and PC may have reduced parent need for ED use to gain reassurance on nonurgent concerns.
We found no difference between the 2 groups on well-visit or sick/urgent care utilization. This outcome suggests that the decreased ED utilization did not seem to be replaced by increased urgent care or sick visit utilization.
The primary cost of this intervention was the PC salary at approximately $20 per hour plus benefits, at 50% effort per practice (or 16.7% effort per physician). Using a fringe benefit rate of 45% and a per-physician patient panel size of 2000, this calculation would translate into an annual cost of $10 282 per physician or $5.14 per patient. Second, ongoing costs of PARENT include maintenance of the WVP site and EHR linkage (which in large part could potentially be supported by the practice’s information technology and EHR vendors) and the automated text message service (estimated annual cost of $1200 per practice). Even with the potential for larger panel sizes per physician (because less physician effort is required for WCC), small independent practices serving a predominantly Medicaid managed care population may need incentives to make this intervention financially sustainable. Incentives for reducing overall costs of care (including ED use) or for providing higher quality of care and experiences of care to families may help make PARENT a more sustainable model for care.48
Our study has some limitations. We are unable to report which elements of PARENT are driving the intervention effects. To enhance generalizability of our findings, we will need to test the intervention across multiple clinical sites with multiple PCs. The WCC quality of the usual care provided at the practice sites likely exceeded usual care nationally in various domains (eg, developmental screening),10,29 which may also limit generalizability. We also recognize the potential for contamination. It is possible that clinicians spent less time on education and guidance during well-visits on control days as well as their intervention days when the PC was involved. However, the difference between parent ratings of overall care for the intervention and control were small, and parent experiences of care in the control group were generally very good. Finally, our sample was limited to a predominantly low-income, minority population. Our aim was to develop and test a community-specific model of WCC to improve care for publicly insured children in the low-income communities served by the partner clinical sites; thus, our findings may not be generalizable to other populations, such as privately insured children or children in families with higher socioeconomic status.
Our findings suggest that this PC-led model for WCC can improve the receipt of comprehensive WCC for low-income families and potentially lead to cost savings by reducing ED utilization. Replication of these findings across a larger number of clinics and practices will inform the broader use of this intervention in primary care.
The authors thank Healthy-TXT for providing the text message service as part of this intervention, Office Ally for its work in incorporating the WVP results into the health record, and CAHMI for support and maintenance of the WVP for the practice sites.
- Accepted November 19, 2015.
- Address correspondence to Tumaini R. Coker, MD, MBA, Mattel Children’s Hospital, UCLA, 10833 Le Conte Ave, Room 12-436, Los Angeles, CA 90095. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (K23-HD06267), the Health Resources and Services Administration (R40-MC21516), and the UCLA Kaiser Permanente Center for Health Equity. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Hagan JF,
- Shaw JS,
- Duncan PM
- ↵Simon GR, Baker C, Barden GA III, et al; Committee on Practice and Ambulatory Medicine; Bright Futures Periodicity Schedule Workgroup. 2014 Recommendations for pediatric preventive health care. Pediatrics. 2014;133(3):568–570pmid:24567012
- Donabedian A.
- Bethell C,
- Reuland CH,
- Halfon N,
- Schor EL
- Halfon N,
- Regalado M,
- Sareen H, et al
- Bethell C,
- Reuland C,
- Schor E,
- Abrahms M,
- Halfon N
- Halfon N,
- Stevens GD,
- Larson K,
- Olson LM
- Coker TR,
- Thomas T,
- Chung PJ
- Kuo AA,
- Inkelas M,
- Lotstein DS,
- Samson KM,
- Schor EL,
- Halfon N
- Coker TR,
- Windon A,
- Moreno C,
- Schuster MA,
- Chung PJ
- Coker T,
- Casalino LP,
- Alexander GC,
- Lantos J
- Coker TR,
- Chung PJ,
- Cowgill BO,
- Chen L,
- Rodriguez MA
- Mooney K,
- Moreno C,
- Chung PJ,
- Elijah J,
- Coker TR
- Bethell C
- The Child and Adolescent Health Measurement Initiative
- Arroll B,
- Goodyear-Smith F,
- Crengle S, et al
- Bethell C,
- Reuland C,
- Latzke B.
- Child and Adolescent Health Measurement Initiative
- Bethell C,
- Peck C,
- Schor E
- Agency for Healthcare Research and Quality
- Cohen J
- Olson AL,
- Dietrich AJ,
- Prazar G,
- Hurley J
- Green LV,
- Savin S,
- Lu Y
- Taylor EF,
- Machta RM,
- Meyers DS,
- Genevro J,
- Peikes DN
- Farber ML
- Zuckerman B,
- Parker S,
- Kaplan-Sanoff M,
- Augustyn M,
- Barth MC
- National Center for Health Statistics
- Gindi RM,
- Jones LI
- Margolius D,
- Bodenheimer T
- Copyright © 2016 by the American Academy of Pediatrics