During infancy, the American Academy of Pediatrics Bright Futures fourth edition health supervision guidelines recommend frequent well-child visits (WCVs) in which providers are expected to screen for and address maternal depression, intimate partner violence (IPV), and health-related social needs (HRSN). We spread an evidence-based approach that implements these recommendations (Developmental Understanding and Legal Collaboration for Everyone; DULCE) with 3 aims for 6-month-old infants and their families: 75% receive all WCVs on time, 95% are screened for 7 HRSNs, and 90% of families with concrete supports needs and 75% of families with maternal depression or IPV receive support.
Between January 2017 and July 2018, five DULCE teams (including a community health worker, early childhood system representative, legal partner, clinic administrator, pediatric and behavioral health clinicians) from 3 communities in 2 states participated in a learning collaborative. Teams adapted DULCE using Plan-Do-Study-Act cycles, reported data, and shared learning monthly. Run charts were used to study measures. The main outcome was the percent of infants that received all WCVs on time.
The percentage of families who completed all WCVs on time increased from 46% to 65%. More than 95% of families were screened for HRSNs, 70% had ≥1 positive screen, and 86% and 71% of those received resource information for concrete supports and maternal depression and IPV, respectively.
Quality improvement–supported DULCE expansion increased by 50% the proportion of infants receiving all WCVs on time and reliably identified and addressed families’ HRSNs, via integration of existing resources.
Early infancy is a time of joy and vulnerability. During this period of rapid brain development, infants are particularly sensitive to adverse and protective experiences.1–5 The physiologic, financial, and social impacts of caring for a newborn often cause parental stress. Children in low-income families, 44% of US children <3,6 are at-risk for worse health and developmental outcomes than their wealthier peers.1,7,8 High-quality, comprehensive pediatric interventions can help families overcome early adversity.1,4,5,9
Bright Futures Guidelines for Health Supervision of Infants, Children, and Adolescents (BF4) recommend 5 WCVs by 6 months of age, including a new visit at 1 month of age. Furthermore, the American Academy of Pediatrics’ Poverty and Child Health statement suggests practices that address social determinants of health: pediatricians are expected to screen for and address maternal depression, intimate partner violence (IPV), and health-related social needs (HRSN).10,11 Payers increasingly expect medical homes to address HRSN within value-based structures.12–14
Pediatric clinics are well-positioned to support families. Nearly all young children in the United States receive routine health care.15 The benefits of timely WCVs are well-established, including timely immunizations, early detection of developmental delays16 and feeding challenges,17 reduced emergency department use and hospitalizations,18 and pediatrician-delivered anticipatory guidance that parents value.19
However, care systems struggle to address all BF4 recommendations. Although in 2017 ∼90% of children <2 received a WCV in the past year,15 the proportion who receive all recommended visits is unknown. Most providers do not regularly screen for HRSN.20 In a recent survey, 33% of US physician practices reported not screening for any HRSN; only 15% reported screening for 5 HRSN.21 Finally, clinics often lack reliable processes for offering HRSN resources or helping families access support.22 Barriers include redesign costs (training, protocols, access to reliable information about community services)23 and lack of financial incentives.24
Developmental Understanding and Legal Collaboration for Everyone (DULCE) is an evidence-based pediatric primary care approach for families with infants from birth through 6 months of age that overcomes common barriers to implementing BF4 recommendations. DULCE embeds a community health worker (family specialist; FS) within a cross-sector team that includes an early childhood system representative, legal partner, clinic administrator, and pediatric and behavioral health clinicians. The team works together to link families to needed resources. A randomized controlled trial revealed that DULCE reduced emergency department visits, increased preventive care adherence, and accelerated access to concrete supports among participating families.25
In 2016, the Center for the Study of Social Policy (CSSP) launched an initiative to spread DULCE using a Breakthrough Series Collaborative, a quality improvement (QI) model that can facilitate implementation of evidence-based practices.26 It has been used in clinics to improve delivery of Bright Futures preventive services27–29 and in cross-sector service systems to improve child and family well-being.30
This initiative spread DULCE to 5 sites between January 2017 and July 2018, with 3 aims for 6-month-old infants and their families: 75% receive all 5 recommended WCVs on time, 95% are screened for 7 HRSNs, and 90% of families with concrete supports needs and 75% of families with maternal depression or IPV receive information about available resources.
Methods
CSSP contacted its Early Childhood Learning and Innovation Network for Communities, a national network of 14 communities that are early childhood systems innovators. Three communities volunteered for this initiative and recruited clinics serving predominantly Medicaid-insured patients and local public interest law organizations to form local DULCE teams (Table 1).
Cross-Sector DULCE Team Members and Participating Communities
| Early Childhood System Lead Agencies | Clinic Partners | Legal Partners | |
|---|---|---|---|
| Unique contribution | Accountable for a local system of services for families with young children | Offer universal reach and longitudinal relationships with families | Offer a professional orientation toward problem-solving and advocacy |
| Expertise | Well-versed in community resources for families and training opportunities for FS | Well-versed in the use of standard protocols to improve quality of care | Well-versed in family rights and system responsibilities |
| Role on team | Inform team of available community resources, champion evidence-informed practices, influence policy | Provide ongoing monitoring of families’ status and coaching of the FS to respond to unique infant and family circumstances | Lend a policy lens and expertise, offer ongoing identification of supports and strategies to address family needs |
| Communities | |||
| Alameda County, CA | First 5 Alameda County | Highland Pediatric Clinic (Oakland, CA) | East Bay Community Law Center |
| Lamoille Valley, VT | Lamoille Family Center | Appleseed Pediatrics | Vermont Legal Aid |
| Los Angeles County, CA | First 5 Los Angeles | The Children’s Clinic (Long Beach, CA) | Legal Aid Foundation of Los Angeles |
| Northeast Valley Health Corporation, Sun Valley | Legal Aid Foundation of Los Angeles | ||
| St. John’s Well Child and Family Center | Neighborhood Legal Services of Los Angeles County |
| Early Childhood System Lead Agencies | Clinic Partners | Legal Partners | |
|---|---|---|---|
| Unique contribution | Accountable for a local system of services for families with young children | Offer universal reach and longitudinal relationships with families | Offer a professional orientation toward problem-solving and advocacy |
| Expertise | Well-versed in community resources for families and training opportunities for FS | Well-versed in the use of standard protocols to improve quality of care | Well-versed in family rights and system responsibilities |
| Role on team | Inform team of available community resources, champion evidence-informed practices, influence policy | Provide ongoing monitoring of families’ status and coaching of the FS to respond to unique infant and family circumstances | Lend a policy lens and expertise, offer ongoing identification of supports and strategies to address family needs |
| Communities | |||
| Alameda County, CA | First 5 Alameda County | Highland Pediatric Clinic (Oakland, CA) | East Bay Community Law Center |
| Lamoille Valley, VT | Lamoille Family Center | Appleseed Pediatrics | Vermont Legal Aid |
| Los Angeles County, CA | First 5 Los Angeles | The Children’s Clinic (Long Beach, CA) | Legal Aid Foundation of Los Angeles |
| Northeast Valley Health Corporation, Sun Valley | Legal Aid Foundation of Los Angeles | ||
| St. John’s Well Child and Family Center | Neighborhood Legal Services of Los Angeles County |
CSSP recruited faculty (CSSP staff, DULCE model developers, a QI expert, practicing pediatrician, and infant mental health specialists) who formed the DULCE National team that defined aims, a theory of change, and measures. DULCE participants met 4 times over 19 months. Each month, they tested interventions using Plan-Do-Study-Act (PDSA) cycles, reported measures, and participated in webinars. DULCE National provided monthly QI coaching calls, 2 site visits, and Brazelton Touchpoints Center training.31
Intervention
DULCE National taught DULCE’s aims, key drivers, and interventions (Fig 1). The first driver focuses on comprehensive care enriched by a FS who attends WCVs, reinforces protective factors, offers developmental guidance, and is families’ most frequent point of contact. The second driver concentrates on identification of families’ strengths and HRSN and implementation of family-led problem-solving across 7 evidence-based HRSN screening domains: maternal depression, IPV, food insecurity, housing instability, housing conditions, utilities, and employment and financial supports.
DULCE Key driver diagram DULCE’S key driver diagram (theory of change) with aims, primary drivers and associated process measures, and interventions.
DULCE Key driver diagram DULCE’S key driver diagram (theory of change) with aims, primary drivers and associated process measures, and interventions.
The third driver emphasizes the cross-sector team that includes the FS, an early childhood system representative, legal partner, clinic administrator, and pediatric and behavioral health clinicians. This team conducts weekly case reviews; collaborates to support families’ access to benefits, services, and legal protections; and identifies opportunities to effect policy and systems improvements.32 The fourth driver prioritizes families as partners via diverse strategies (eg, episodic input) and as DULCE QI team members. The fifth driver concentrates on QI to overcome implementation challenges using PDSA cycles.33
Team members communicated in person at weekly case reviews and monthly QI meetings and by e-mail, telephone, and text. Providers and FS communicated via the electronic health record and conversations in clinic. DULCE behavioral health specialists provided FSs weekly reflective supervision.
Enrollment Procedures
Families with newborns up to 8 weeks of age were enrolled at their first office visit, excluding newborns hospitalized for >7 days after birth because they may warrant specialized services. At sites with more newborns than 1 FS could serve, DULCE was offered to a randomly selected subset. Newborn enrollment was ongoing and continued beyond the study period. This report includes infants born between January 2017 and July 2018 and followed through their 6-month WCV.
Clinics introduced DULCE as part of routine care, included information about DULCE in welcome packets, and introduced the FS as a care team member at the first WCV. Families were given the opportunity to opt out.
Evaluation Approach
Data and Measures
The FSs entered individual-level demographic, program participation, encounter, and HRSN data into an online, custom-built registry (Tables 2 and 3).
Description of DULCE Data and Measures
| Type of Data | Data Item |
|---|---|
| Demographic data | Infant sex and date of birth |
| Primary caregiver role, marital status, age, race and ethnicity | |
| Secondary caregiver role, age | |
| Household number of adults, number of children | |
| Primary language spoken at home | |
| Program participation data | Enrollment date |
| Termination date: date the infant received a 6-mo WCV or dropped out | |
| Termination reason (if dropped out before 6-mo WCV) | |
| No. weeks in program: number of days between the date an infant enrolled in DULCE and the date the infant completed or terminated DULCE participation | |
| Encounter data | Date of Encounter |
| Contact time: No. minutes FS spent conducting encounter | |
| Type of encounter. Encounters included WCVs, sick visits, FS contacts (telephone calls, text messages, and e-mail messages with or on behalf of the family; face-to-face meetings not associated with clinic visits), and case reviews where family was discussed | |
| HRSN Data | Date screening was conducted |
| Screening result (positive or negative) | |
| Resource information provided (yes or no) |
| Type of Data | Data Item |
|---|---|
| Demographic data | Infant sex and date of birth |
| Primary caregiver role, marital status, age, race and ethnicity | |
| Secondary caregiver role, age | |
| Household number of adults, number of children | |
| Primary language spoken at home | |
| Program participation data | Enrollment date |
| Termination date: date the infant received a 6-mo WCV or dropped out | |
| Termination reason (if dropped out before 6-mo WCV) | |
| No. weeks in program: number of days between the date an infant enrolled in DULCE and the date the infant completed or terminated DULCE participation | |
| Encounter data | Date of Encounter |
| Contact time: No. minutes FS spent conducting encounter | |
| Type of encounter. Encounters included WCVs, sick visits, FS contacts (telephone calls, text messages, and e-mail messages with or on behalf of the family; face-to-face meetings not associated with clinic visits), and case reviews where family was discussed | |
| HRSN Data | Date screening was conducted |
| Screening result (positive or negative) | |
| Resource information provided (yes or no) |
DULCE Process and Outcome Measures
| Process Measure | Numerator | Denominator |
|---|---|---|
| Percent of WCVs attended by the FS | Of the recommended WCVs that occurred each month, the number that were attended by the FS | No. recommended WCVs for DULCE-enrolled infants that occurred each month |
| Percent of families that were screened for 7 HRSN | Among all families enrolled in DULCE, number of families that were screened for 7 HRSN using validated, standardized screening questions | No. families enrolled in DULCE |
| Percent of families with identified HRSN that were provided information about available resources | Among all families enrolled in DULCE that screened positive for at least 1 HRSN, number of families that were provided information about available resources by the FS and/or other members of the cross-sector team | No. families enrolled in DULCE that screened positive for at least 1 HRSN via validated, standardized screening questions |
| Outcome measure | Numerator | Denominator |
| Percent of infants that received alla recommended WCVs on timeb | Among all infants enrolled in DULCE, number of infants that received alla recommended WCVs on timeb | No. infants enrolled in DULCE |
| Process Measure | Numerator | Denominator |
|---|---|---|
| Percent of WCVs attended by the FS | Of the recommended WCVs that occurred each month, the number that were attended by the FS | No. recommended WCVs for DULCE-enrolled infants that occurred each month |
| Percent of families that were screened for 7 HRSN | Among all families enrolled in DULCE, number of families that were screened for 7 HRSN using validated, standardized screening questions | No. families enrolled in DULCE |
| Percent of families with identified HRSN that were provided information about available resources | Among all families enrolled in DULCE that screened positive for at least 1 HRSN, number of families that were provided information about available resources by the FS and/or other members of the cross-sector team | No. families enrolled in DULCE that screened positive for at least 1 HRSN via validated, standardized screening questions |
| Outcome measure | Numerator | Denominator |
| Percent of infants that received alla recommended WCVs on timeb | Among all infants enrolled in DULCE, number of infants that received alla recommended WCVs on timeb | No. infants enrolled in DULCE |
For infants who graduated from the 6-mo intervention, these are the infants that received all 5 WCVs on time. For infants that dropped out of DULCE before 6 mo of life, this includes infants who had received all recommended WCVs up to the date of dropout.
DULCE National defined “on time” as follows: first WCV between nursery discharge and day 9 of life, a 1-mo WCV between days 9 and 42 d, a 2-mo WCV between days 43 and 81, a 4-mo WCV between days 105 and 141, and a 6-mo WCV between days 165 to 201.
Process measures aligned with the primary drivers (PDs): PD1: percentage of WCVs attended by the FS; PD2: percent of enrolled families that were screened for 7 HRSN using validated, standardized screening questions; and PD3: percentage of families with identified HRSN that received resource information.
The main outcome was the percentage of 6-month-old infants who received all recommended WCVs on time. It includes infants who completed the intervention and received 5 WCVs on time and infants who dropped out and received all recommended WCVs on time up to the date of dropout. DULCE National defined “on time” on the basis of precedent34 : first WCV between nursery discharge and day 9 of life, 1-month WCV between days 9 and 42, 2-month WCV between days 43 and 81, 4-month WCV between days 105 and 141, and 6-month WCV between days 165 and 201.
Analysis
Descriptive statistics were calculated for patient demographics, enrollment and completion rates, number of weeks enrolled, total number of encounters per family, and FS–family contact time.
The process and outcome measures were analyzed in time series as run charts.35 Subjects were counted in the denominator of each measure once (ie, denominators are independent of each other) and placed in the month they enrolled. Thus, the run charts end in July 2018, and all participants were followed through their 6-month WCV.
Because no preintervention data were available, the first 10 data points generated the baseline. Means were used instead of medians because these were noncontinuous data, and some measures’ medians were extreme values because of high baseline performance (eg, baseline median screening rates of 100%) and small denominators at the site level (eg, months with few or no families identified with IPV). Criteria for applying probability-based rules for identifying improvements were met: denominators were roughly equal over time, and at the aggregate level, data were appropriately dispersed.36
Two probability-based rules were used to identify changes in the data that have <5% probability of occurring by chance: a “shift” of 6 or more points in a row above or below the mean, and a “trend” of 5 consecutive increasing or decreasing points.37 When a shift occurred, the average of the 6 shifted points became the new mean, from which subsequent shifts were identified.
At monthly QI meetings, DULCE National provided site-level and aggregate data reports. Teams reviewed data and discussed PDSAs, which were recorded and annotated on run charts, allowing teams to draw inferences from the temporal relationships of interventions and results. Analyses were conducted by using Stata 14.2.
Ethical Considerations
This study was approved by University of Chicago School of Social Administration’s Institutional Review Board (IRB17-0414).
Results
Five local DULCE site teams and 692 families with infants born between January 1, 2017, and July 31, 2018, participated. One site (St. John’s) withdrew on June 30, 2018; the analytic sample included infants from that site born January 1 to December 31, 2017. All participants were followed through their 6-month WCV.
Families represented the demographics of the clinic populations (Table 4). Overall, 97% of primary caregivers were mothers whose median age was 28 years. Half were single; 25% identified as white, 16% as Black, and 56% as Hispanic/Latino. Three-quarters of families reported a second caregiver: 90% were fathers, whose median age was 30 years. Families mainly spoke English (66%) or Spanish (24%) at home.
Demographic Characteristics of DULCE Families by Site
| Total, n (%) | Appleseed: Lamoille County, VT, n (%) | Highland: Alameda County, CA, n (%) | Northeast Valley: LA County, CA, n (%) | Children’s Clinic: LA County, CA, n (%) | St. John’s: LA County, CA, n (%) | |
|---|---|---|---|---|---|---|
| Full sample | 692 (100) | 139 (100) | 148 (100) | 139 (100) | 161 (100) | 105 (100) |
| Child sex | ||||||
| Male | 341 (49) | 74 (53) | 76 (51) | 63 (45) | 71 (44) | 57 (54) |
| Female | 351 (51) | 65 (47) | 72 (49) | 76 (55) | 90 (56) | 48 (46) |
| Primary caregiver | ||||||
| Mother | 666 (97) | 125 (90) | 146 (100) | 138 (99) | 153 (95) | 104 (99) |
| Father | 15 (2) | 12 (9) | 0 (0) | 1 (1) | 1 (1) | 1 (1) |
| Othera | 9 (1) | 2 (1) | 0 (0) | 0 (0) | 7 (4) | 0 (0) |
| Primary caregiver marital status | ||||||
| Single | 321 (52) | 21 (15) | 72 (51) | 16 (20) | 125 (80) | 85 (87) |
| Married | 212 (34) | 77 (56) | 51 (36) | 38 (48) | 31 (20) | 15 (15) |
| Domestic partner | 83 (13) | 39 (28) | 17 (12) | 25 (32) | 1 (1) | 1 (1) |
| Divorced | 1 (0) | 1 (1) | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Primary caregiver age, median (range) | 28 (14–52) | 30 (17–52) | 29 (15–47) | 27 (16–43) | 28 (14–44) | 26 (16–42) |
| Primary caregiver race | ||||||
| White | 164 (25) | 133 (97) | 10 (8) | 1 (1) | 19 (12) | 1 (1) |
| Hispanic/Latino | 374 (56) | 0 (0) | 67 (51) | 131 (96) | 86 (56) | 90 (86) |
| Black | 104 (16) | 3 (2) | 46 (35) | 3 (2) | 39 (25) | 13 (12) |
| Asian American | 13 (2) | 1 (1) | 6 (5) | 2 (2) | 4 (3) | 0 (0) |
| Pacific Islander | 8 (1) | 0 (0) | 2 (2) | 0 (0) | (5 (3) | 1 (1) |
| American Indian | 2 (0) | 0 (0) | 1 (1) | 0 (0) | 1 (1) | 0 (0) |
| Secondary caregiver | ||||||
| Father | 470 (90) | 115 (90) | 125 (98) | 35 (100) | 129 (95) | 66 (69) |
| Grandparent | 28 (5) | 1 (1) | 1 (1) | 0 (0) | 4 (3) | 22 (23) |
| Mother | 13 (3) | 12 (9) | 0 (0) | 0 (0) | 1 (1) | 0 (0) |
| Other caregiver | 11 (2) | 0 (0) | 1 (1) | 0 (0) | 2 (2) | 8 (8) |
| Secondary caregiver age, median (range) | 30 (15–-66) | 32 (18–53) | 30 (16–66) | 29.5 (21–33) | 30 (15–57) | 27.5 (16–40) |
| No. adults in home | ||||||
| 1 | 32 (5) | 8 (5) | 10 (7) | (4 (3) | 8 (5) | 1 (1) |
| 2 | 463 (70) | 105 (68) | 76 (56) | 91 (67) | 105 (68) | 72 (71) |
| 3 | 89 (13) | 17 (11) | 26 (19) | 25 (18) | 17 (11) | 15 (15) |
| 4 or more | 82 (12) | 24 (16) | 23 (17) | 16 (12) | 16 (24) | 14 (14) |
| No. children in home | ||||||
| 1 | 204 (31) | 45 (33) | 44 (35) | 32 (24) | 40 (26) | 43 (42) |
| 2 | 232 (36) | 56 (40) | 45 (36) | 44 (33) | 58 (38) | 29 (28) |
| 3 | 130 (20) | 30 (22) | 15 (12) | 35 (26) | 35 (23) | 15 (15) |
| 4 or more | 85 (13) | 7 (5) | 21 (17) | 22 (17) | (13) | 15 (15) |
| Primary language spoken at home | ||||||
| English | 449 (66) | 137 (99) | 67 (47) | 69 (53) | 121 (77) | 55 (52) |
| Spanish | 162 (24) | 0 (0) | 38 (26) | 55 (42) | 19 (12) | 50 (48) |
| English and Spanish | 28 (4) | 0 (0) | 10 (7) | 5 (4) | 13 (8) | 0 (0) |
| Otherb | 38 (6) | 2 (1) | 29 (20) | 2 (2) | 5 (4) | 0 (0) |
| Total, n (%) | Appleseed: Lamoille County, VT, n (%) | Highland: Alameda County, CA, n (%) | Northeast Valley: LA County, CA, n (%) | Children’s Clinic: LA County, CA, n (%) | St. John’s: LA County, CA, n (%) | |
|---|---|---|---|---|---|---|
| Full sample | 692 (100) | 139 (100) | 148 (100) | 139 (100) | 161 (100) | 105 (100) |
| Child sex | ||||||
| Male | 341 (49) | 74 (53) | 76 (51) | 63 (45) | 71 (44) | 57 (54) |
| Female | 351 (51) | 65 (47) | 72 (49) | 76 (55) | 90 (56) | 48 (46) |
| Primary caregiver | ||||||
| Mother | 666 (97) | 125 (90) | 146 (100) | 138 (99) | 153 (95) | 104 (99) |
| Father | 15 (2) | 12 (9) | 0 (0) | 1 (1) | 1 (1) | 1 (1) |
| Othera | 9 (1) | 2 (1) | 0 (0) | 0 (0) | 7 (4) | 0 (0) |
| Primary caregiver marital status | ||||||
| Single | 321 (52) | 21 (15) | 72 (51) | 16 (20) | 125 (80) | 85 (87) |
| Married | 212 (34) | 77 (56) | 51 (36) | 38 (48) | 31 (20) | 15 (15) |
| Domestic partner | 83 (13) | 39 (28) | 17 (12) | 25 (32) | 1 (1) | 1 (1) |
| Divorced | 1 (0) | 1 (1) | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Primary caregiver age, median (range) | 28 (14–52) | 30 (17–52) | 29 (15–47) | 27 (16–43) | 28 (14–44) | 26 (16–42) |
| Primary caregiver race | ||||||
| White | 164 (25) | 133 (97) | 10 (8) | 1 (1) | 19 (12) | 1 (1) |
| Hispanic/Latino | 374 (56) | 0 (0) | 67 (51) | 131 (96) | 86 (56) | 90 (86) |
| Black | 104 (16) | 3 (2) | 46 (35) | 3 (2) | 39 (25) | 13 (12) |
| Asian American | 13 (2) | 1 (1) | 6 (5) | 2 (2) | 4 (3) | 0 (0) |
| Pacific Islander | 8 (1) | 0 (0) | 2 (2) | 0 (0) | (5 (3) | 1 (1) |
| American Indian | 2 (0) | 0 (0) | 1 (1) | 0 (0) | 1 (1) | 0 (0) |
| Secondary caregiver | ||||||
| Father | 470 (90) | 115 (90) | 125 (98) | 35 (100) | 129 (95) | 66 (69) |
| Grandparent | 28 (5) | 1 (1) | 1 (1) | 0 (0) | 4 (3) | 22 (23) |
| Mother | 13 (3) | 12 (9) | 0 (0) | 0 (0) | 1 (1) | 0 (0) |
| Other caregiver | 11 (2) | 0 (0) | 1 (1) | 0 (0) | 2 (2) | 8 (8) |
| Secondary caregiver age, median (range) | 30 (15–-66) | 32 (18–53) | 30 (16–66) | 29.5 (21–33) | 30 (15–57) | 27.5 (16–40) |
| No. adults in home | ||||||
| 1 | 32 (5) | 8 (5) | 10 (7) | (4 (3) | 8 (5) | 1 (1) |
| 2 | 463 (70) | 105 (68) | 76 (56) | 91 (67) | 105 (68) | 72 (71) |
| 3 | 89 (13) | 17 (11) | 26 (19) | 25 (18) | 17 (11) | 15 (15) |
| 4 or more | 82 (12) | 24 (16) | 23 (17) | 16 (12) | 16 (24) | 14 (14) |
| No. children in home | ||||||
| 1 | 204 (31) | 45 (33) | 44 (35) | 32 (24) | 40 (26) | 43 (42) |
| 2 | 232 (36) | 56 (40) | 45 (36) | 44 (33) | 58 (38) | 29 (28) |
| 3 | 130 (20) | 30 (22) | 15 (12) | 35 (26) | 35 (23) | 15 (15) |
| 4 or more | 85 (13) | 7 (5) | 21 (17) | 22 (17) | (13) | 15 (15) |
| Primary language spoken at home | ||||||
| English | 449 (66) | 137 (99) | 67 (47) | 69 (53) | 121 (77) | 55 (52) |
| Spanish | 162 (24) | 0 (0) | 38 (26) | 55 (42) | 19 (12) | 50 (48) |
| English and Spanish | 28 (4) | 0 (0) | 10 (7) | 5 (4) | 13 (8) | 0 (0) |
| Otherb | 38 (6) | 2 (1) | 29 (20) | 2 (2) | 5 (4) | 0 (0) |
LA, Los Angeles.
6 Foster parents, 2 Legal guardians, 1 Grandparent.
Amharic, Arabic, Bengali, French, Igbo, Mam, Pashto, Portuguese, Punjabi, Samoan, Swahili, Tagalog, Tigrigna, Turkish, Yoruba, English and other, Spanish and other.
All families that were offered DULCE enrolled; 79% completed the 6-month program. Two-thirds of families that dropped out moved away or changed clinics (Table 5). Enrolled families averaged 24 weeks in the program (confidence interval [CI], 23.1–24.2), 11 encounters (CI, 10.3–11.1; range, 1–49), and 280 minutes of FS contact time (CI, 265–294).
DULCE Enrollment, Completion and Reasons for Early Termination
| Total, n (%) | Appleseed: Lamoille County, VT, n (%) | Highland: Alameda County, CA, n (%) | Northeast Valley: LA County, CA, n (%) | Children’s Clinic: LA County, CA, n (%) | St. John’s: LA County, CA, n (%) | |
|---|---|---|---|---|---|---|
| Families offered DULCE | 692 | 139 | 161 | 148 | 139 | 105 |
| Families enrolled in DULCE | 692 (100) | 139 (100) | 161 (100) | 148 (100) | 139 (100) | 105 (100) |
| Families completed DULCE | 549 (79) | 129 (93) | 124 (77) | 102 (69) | 123 (88) | 71 (68) |
| Families that terminated DULCE participation earlya | 143 (21) | 10 (7) | 37 (23) | 46 (31) | 16 (12) | 34 (32) |
| Reasons for early termination | ||||||
| Moved home | 51 (36) | 2 (20) | 10 (27) | 17 (37) | 7 (44) | 15 (44) |
| Change clinic or provider | 44 (30) | 5 (50) | 19 (52) | 15 (33) | 4 (26) | 1 (3) |
| Lost to follow-up | 24 (17) | 3 (30) | 3 (8) | 12 (26) | 5 (31) | 1 (3) |
| Infant removed from home | 2 (1) | 0 (0) | 1 (3) | 1 (2) | 0 (0) | 0 (0) |
| Family requested | 3 (2) | 0 (0) | 2 (5) | 0 (0) | 0 (0) | 1 (3) |
| Missing | 19 (13) | 0 (0) | 2 (5) | 1 (2) | 0 (0) | 16 (47) |
| Total, n (%) | Appleseed: Lamoille County, VT, n (%) | Highland: Alameda County, CA, n (%) | Northeast Valley: LA County, CA, n (%) | Children’s Clinic: LA County, CA, n (%) | St. John’s: LA County, CA, n (%) | |
|---|---|---|---|---|---|---|
| Families offered DULCE | 692 | 139 | 161 | 148 | 139 | 105 |
| Families enrolled in DULCE | 692 (100) | 139 (100) | 161 (100) | 148 (100) | 139 (100) | 105 (100) |
| Families completed DULCE | 549 (79) | 129 (93) | 124 (77) | 102 (69) | 123 (88) | 71 (68) |
| Families that terminated DULCE participation earlya | 143 (21) | 10 (7) | 37 (23) | 46 (31) | 16 (12) | 34 (32) |
| Reasons for early termination | ||||||
| Moved home | 51 (36) | 2 (20) | 10 (27) | 17 (37) | 7 (44) | 15 (44) |
| Change clinic or provider | 44 (30) | 5 (50) | 19 (52) | 15 (33) | 4 (26) | 1 (3) |
| Lost to follow-up | 24 (17) | 3 (30) | 3 (8) | 12 (26) | 5 (31) | 1 (3) |
| Infant removed from home | 2 (1) | 0 (0) | 1 (3) | 1 (2) | 0 (0) | 0 (0) |
| Family requested | 3 (2) | 0 (0) | 2 (5) | 0 (0) | 0 (0) | 1 (3) |
| Missing | 19 (13) | 0 (0) | 2 (5) | 1 (2) | 0 (0) | 16 (47) |
Early termination refers to families that dropped out of DULCE before completing their 6-mo WCV.
The percentage of WCVs attended by the FS increased over time, from 0% before DULCE to 66% during the first 10 months of DULCE implementation (Fig 2). A shift to 70% occurred (August 2017 to June 2018) as providers gained comfort with FSs, teams adapted schedules to reduce simultaneous infant appointments and improved workflows to maximize families’ time with FS in clinic.
Percentage of WCVs attended by the FS. DULCE embeds a FS within a cross-sector team to support families with infants from birth through 6 months. Attending the WCV offers the FS an opportunity to build relationships with families, encourage timely WCV attendance, screen for HRSN, and engage in family-led problem solving for identified HRSN. LA, Los Angeles.
Percentage of WCVs attended by the FS. DULCE embeds a FS within a cross-sector team to support families with infants from birth through 6 months. Attending the WCV offers the FS an opportunity to build relationships with families, encourage timely WCV attendance, screen for HRSN, and engage in family-led problem solving for identified HRSN. LA, Los Angeles.
Figure 3 reveals the process measures for HRSN screening. More than 95% of families were screened for each HRSN: maternal depression (95.9%), IPV (96.3%), food insecurity (97.2%), employment and financial needs (98.6%), utilities (96.8%), and housing instability (97.2%). Housing conditions screening improved from 94.5% to 95.8%. Ninety-two percent of families were screened for all 7 HRSN, and 3 sites demonstrated shifts from 72% to 79%, 92% to 100%, and 95% to 100%, respectively.
A, Percentage of families screened for each of the HRSNs. B, Percentage of families screened for 7 HRSNs. Percentage of families screened for HRSN. DULCE FSs screen all families for 7 HRSN (B) using validated screening questions. These evidence-based screening domains are maternal depression, IPV, food insecurity, housing instability, housing conditions, utilities, and employment and financial supports (A). LA, Los Angeles.
A, Percentage of families screened for each of the HRSNs. B, Percentage of families screened for 7 HRSNs. Percentage of families screened for HRSN. DULCE FSs screen all families for 7 HRSN (B) using validated screening questions. These evidence-based screening domains are maternal depression, IPV, food insecurity, housing instability, housing conditions, utilities, and employment and financial supports (A). LA, Los Angeles.
Seventy percent of families had at least 1 positive screen: 25% had 2 and 16% had 3 or more. Fifty-one percent of families screened positive for financial-employment needs, 46.1% for food insecurity, 14.3% for maternal depression, 13.2% for housing insecurity, 5.1% for IPV, 3.5% for unhealthy housing conditions, and 2.2% for utility needs.
The percentage of families with identified HRSN that received information about available resources varied (Fig 4). For maternal depression and IPV, 70.7% of families that screened positive received resource information. One site demonstrated a shift from 60.0% to 86.7% when the FS built a relationship and process for directly referring mothers with depression to the behavioral health clinician in an affiliated women’s health clinic.
A, Maternal depression and intimate partner violence needs. B, Concrete supports needs. Percentage of families with identified HRSN that received information about available resources. DULCE's cross-sector team (composed of a FS, early childhood system representative, legal partner, clinic administrator, and pediatric and behavioral health clinicians) leverages up-to-date knowledge of and relationships with clinic- and community-based resources to provide families resource information for maternal depression, IPV (A) and concrete supports needs (B). LA, Los Angeles.
A, Maternal depression and intimate partner violence needs. B, Concrete supports needs. Percentage of families with identified HRSN that received information about available resources. DULCE's cross-sector team (composed of a FS, early childhood system representative, legal partner, clinic administrator, and pediatric and behavioral health clinicians) leverages up-to-date knowledge of and relationships with clinic- and community-based resources to provide families resource information for maternal depression, IPV (A) and concrete supports needs (B). LA, Los Angeles.
Ninety-five percent of families with concrete support needs received resource information at baseline. There was a downward shift to 86.4% overall (January to July 2018), but no downward shifts for individual sites. Three of 4 sites provided resource information to >90% of families throughout, and 1 demonstrated a shift from 90% to 100%.
The main outcome, the percentage of 6-month-old infants that completed all 5 recommended WCVs on time, improved from 45.5% to 64.6%. Three sites showed shifts; 1 met the 75% aim with a shift from 50% to 83% (Fig 5).
A, Percentage of 6-month-old infants who received all 5 recommended WCVs on time. B, Percentage of infants who received 1-month WCV on time. Percentage of infants that received recommended WCVs on time. Sites improved rates of 6-month-old infants receiving all 5 recommended WCVs on time (A), partly, because of improved 1-month WCV timeliness (B). To meet the 75% aim for (A), each WCV category had an 80% timely attendance aim. The annotations describe PDSAs sites tested to make these improvements. LA, Los Angeles.
A, Percentage of 6-month-old infants who received all 5 recommended WCVs on time. B, Percentage of infants who received 1-month WCV on time. Percentage of infants that received recommended WCVs on time. Sites improved rates of 6-month-old infants receiving all 5 recommended WCVs on time (A), partly, because of improved 1-month WCV timeliness (B). To meet the 75% aim for (A), each WCV category had an 80% timely attendance aim. The annotations describe PDSAs sites tested to make these improvements. LA, Los Angeles.
Improvement in on-time 1-month WCVs from 62.5% to 79.5% contributed. During monthly QI meetings, sites discovered that some team members were unaware that BF4 added a 1-month WCV reimbursable by Medicaid.38 They tested new clinic protocols and schedule templates via PDSAs, and 4 of 5 sites exhibited shifts.
Improvements in different WCVs at different sites also contributed (Figs 6–9). Two sites improved on-time 3-to-5-day visits (from 96.7% to 100% and 86.5% to 94.1%) by changing call centers’ protocols to schedule newborns’ first WCV at the clinic closest to home. Two clinics increased on-time 2-month WCVs (from 77.3% to 93.7% and 84.4% to 96.0%) by scheduling 1- and 2-month WCVs at checkout for newborn visits. Finally, 2 clinics improved on-time 4-month WCVs (from 86.4% to 95.1% and 70.0% to 89.4%) by working with DULCE legal partners to address inappropriate lapses in infants’ Medicaid.39 No improvements were seen for 6-month WCVs, which had the second-lowest on-time visit rate after 1-month visits.
Percentage of who that received the 3-to-5-day WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Appleseed and The Children’s Clinic improved on-time 3-to-5-day visits by scheduling newborns’ first WCV at the clinic closest to home. LA, Los Angeles.
Percentage of who that received the 3-to-5-day WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Appleseed and The Children’s Clinic improved on-time 3-to-5-day visits by scheduling newborns’ first WCV at the clinic closest to home. LA, Los Angeles.
Percentage of infants who received the 2-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Highland and Northeast Valley improved on-time 2-month visits by scheduling 1- and 2-month WCVs at checkout for newborn visits. LA, Los Angeles.
Percentage of infants who received the 2-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Highland and Northeast Valley improved on-time 2-month visits by scheduling 1- and 2-month WCVs at checkout for newborn visits. LA, Los Angeles.
Percentage of infants who received the 4-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Northeast Valley and The Children’s Clinic improved on-time 4-month visits by working with DULCE legal partners to address lapses in infants’ Medicaid. LA, Los Angeles.
Percentage of infants who received the 4-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. Northeast Valley and The Children’s Clinic improved on-time 4-month visits by working with DULCE legal partners to address lapses in infants’ Medicaid. LA, Los Angeles.
Percentage of infants who received the 6-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. No improvements were seen for 6-month WCVs, which had the lowest on-time visit rate after 1-month visits. LA, Los Angeles.
Percentage of infants who received the 6-month WCV on time. To meet the 75% aim of all infants receiving 5 recommended WCVs in the first 6 months of life, each WCV category had an 80% timely attendance aim. No improvements were seen for 6-month WCVs, which had the lowest on-time visit rate after 1-month visits. LA, Los Angeles.
Discussion
QI-supported DULCE expansion to 5 diverse US sites was associated with a 50% increase in the proportion of 6-month-old infants who received all recommended WCVs on time, and with high rates of HRSN screening and provision of information about available resources.
These results mirror those of the previously reported RCT.25 That study revealed showed significant improvements in immunizations, a proxy for WCV, the main outcome reported here. This suggests that the QI process retained outcome fidelity while facilitating intentional adaptations to meet the diverse needs of local sites.
This extends results of a previous QI initiative conducted by an early childhood system that increased the proportion of families who received 3 WCVs in the first 6 months.40 This initiative, in contrast, increased the proportion of families receiving all 5 BF4-recommended WCVs in the first 6 months.
Ensuring that families participate in routine health care early in life establishes families’ relationship with their medical home, a source of support and access point for resources. Universal approaches to promoting healthy development via pediatric medical homes offer a convenient, less stigmatizing way to engage families that many early childhood systems struggle to reach.9 For example, evidence-based home visiting programs that target high-risk families with young children typically enroll 75% of families offered services and retain 39% to 82% at 6 months.41 In this study and the previously reported randomized controlled trial, >95% and 75% of families enrolled and completed the 6 month intervention, respectively.25 Most dropouts were due to families no longer receiving care at intervention clinics; high family mobility may limit the ability of health care-based approaches to address HRSNs.
Medical homes face challenges when addressing HRSN: most clinics screen for 0 or 1 HRSN21 and two-thirds of clinic-based HRSN interventions address a single HRSN.42,43 For the clinics in this study, universal HRSN screening was new, and maternal depression screening was refined.44 DULCE implementation resulted in highly reliable screening for more HRSN than most extant examples.45,46
The high prevalence of HRSN among DULCE families is consistent with previous studies of families living in poverty47 and Medicaid-insured populations.48,49 Effective linkage to resources appears to improve screening rates and accuracy.50 DULCE integrates screening and linkage by integrating early childhood system, legal community, and pediatric health colleagues in its cross-sector team, leveraging up-to-date information about and relationships with clinic- and community-based resources. Further study is needed to validate the observation that engagement with a near-peer who screens conversationally prompts greater willingness to disclose needs,44 possibly increasing the sensitivity of standardized screening tools. Lower-than-expected HRSN prevalence and a propensity to forego assistance are emerging challenges in real-world, broad-based HRSN screening.51
Clinic-based HRSN interventions often struggle to link patients to resources: a recent QI collaborative in 19 pediatric clinics increased HRSN screening from 19% to 73% but did not increase HRSN referrals.52 Even well-established interventions successfully link 48% to 78% of families with HRSN to resources.53–56 DULCE’s comprehensive, systems-based approach exceeded those rates, despite a decrease in the families with concrete supports needs that received resource information from 95% to 86% in January 2018. This decrease was driven by a drop at 3 California-based sites that serve mostly immigrant families. It coincided with the US State Department’s revision of the public charge definition to include the use of noncash health benefits (including Medicaid) in issuing visas,57 and local actions by the US Immigration and Customs Enforcement agency, including parking a marked van in 1 Supplemental Nutrition Program for Women, Infants, and Children clinic’s parking lot.
Like other successful interventions to address HRSNs, DULCE relies on ancillary personnel.56,58 Trained FSs engage families over time, offer repeated opportunities to disclose concerns to the same trusted provider, and anticipate that challenges evolve. The low drop-out rate (for reasons other than moving or changing clinic sites) suggests that families were satisfied with this approach. We hypothesize that the additional personal connection and perceived value received through DULCE motivated families to attend WCVs. Further research can explore family-level drivers for obtaining timely routine health care for infants. At a minimum, these results demonstrate that universal HRSN screening does not drive families away from routine health care.
The selection of volunteer sites for implementing DULCE limits generalizability. One site ceased implementing DULCE at the end of the study period because of unrelated administrative changes. This is the only site that did not demonstrate improvement in any measure, and its performance decreased in the last 3 months. This would bias findings toward the null and does not alter our interpretation; it does reflect the importance of stable leadership to DULCE and any QI initiative.
Our analyses identify improvements that are unlikely due to chance alone but lack causal inference. For this study, we relied on data reported by FSs for QI purposes which did not include balancing measures; ongoing work will incorporate data about contextual factors and stakeholders’ perspectives, electronic health record and claims data for participants and a comparison group. For this cohort, we lack information about how many families provided resources successfully accessed them. This has been incorporated into data collection subsequently and is a focus of future research. Last, despite the 20-percentage-point improvement overall, only 1 site met the aim of 75% of infants receiving all 5 WCVs on time; sites are engaged in further improvement activities to reach this goal.
Conclusions
QI-supported DULCE expansion improved health care use and strengthened the health care system’s capacity to deliver care aligned with BF4 guidelines, including identifying and addressing families’ HRSN. By integrating existing resources and meeting families where they are, this cross-sector approach holds promise for helping families address early-life adversity at scale.
Acknowledgments
This manuscript was completed with the support of Azieb Ermias, Placidina Fico, Jang Lee, Alison Muckle, and Jayne Singer. We express our deep gratitude and appreciation for the DULCE families and teams: Family Specialists (Laura Lopez, Ana DeJesus, Cynthia Garcia, Jennifer Chittick, Monica Martinez); early childhood system leaders (Page Tomblin, Leticia Sanchez, Joselyn Ramirez, Barbara Dubransky, Scott Johnson); legal partners (Erin Le, Yvonne Jimenez, Chris Curtis, Jean Murray, Lauren Holzer, Eliza Schafler, Ja’nai Aubry); pediatric providers (Sam Singer, Mykie Pidor, David Bolour, Adrienne Pahl, Alice Brinkman); clinic administrators (Itta Aswad, Leslie Larsen, Christina Park, Michelle Espiritu, Deborah Rosen, Gina Johnson, Felix Tunador, Helen Duplesiss, Maria Chandler, Elisa Nicholas); and behavioral health specialists (Alma Membreno, Andrea Sanserino, Yolanda Cespedes, Elena Fernandez, Carol Lang-Godin).
FUNDING: DULCE expansion was funded by The JPB Foundation through the Center for the Study of Social Policy.
Dr Arbour conceptualized and designed the study, designed the data collection instruments, coordinated and supervised data collection, conducted the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Floyd and Ms Morton conceptualized and designed the study, designed the data collection instruments, and reviewed and revised the manuscript; Mr Atwood conducted the initial analyses and reviewed and revised the manuscript; Ms Hampton, Ms Murphy Sims, and Ms Doyle supported acquisition and interpretation of data and reviewed and revised the manuscript; Dr Sege conceptualized and designed the study and critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
- BF4
Bright Futures 4 Guidelines for Health Supervision of Infants, Children, and Adolescents
- CI
confidence interval
- CSSP
Center for the Study of Social Policy
- DULCE
Developmental Understanding and Legal Collaboration for Everyone
- FS
family specialist
- HRSN
health-related social needs
- IPV
intimate partner violence
- PD
primary driver
- PDSA
Plan-Do-Study-Act
- QI
quality improvement
- WCV
well-child visit
Comments