Comparing models that integrate obstetric care and WIC on improved program enrollment during pregnancy: a protocol for a randomized controlled trial

Abstract

Background

Low-income, rural pregnant women are at disproportionate risk for adverse pregnancy outcomes as well as future cardiovascular risk. Currently, less than half of eligible women enroll in the Women, Infants, and Children’s (WIC) Program. This study aims to evaluate whether integrating clinical care and social care may advance health equity and reduce health disparities by directly linking women receiving obstetric care to the Special Supplemental Nutrition Program for WIC and/or a Registered Dietitian/Nutritionist (RDN).

Methods

This pragmatic study is situated in real-world care and utilizes a randomized controlled trial design. A total of 240 low-income, rural, pregnant patients will be recruited from Geisinger (Pennsylvania, USA) obstetric clinics and randomized to receive one of four models: (1) Clinic; (2) Clinic-WIC; (3) Clinic-RDN, or (4) Clinic-WIC-RDN. Participants provide consent for electronic referrals that directly link their contact information from the electronic health record to WIC and/or RDN. Patients in the Clinic model receive standard prenatal care, which includes provision of basic information about WIC. The Clinic-WIC model includes a clinical decision alert to queue clinical staff to ask about WIC interest and place a referral to WIC using a social health access referral platform. In turn, WIC staff contact the pregnant woman about enrollment. The Clinic-RDN model includes a referral to an RDN for telehealth counseling to promote heart healthy eating and food resource management. The Clinic-WIC-RDN model includes referrals to both WIC and RDN. The primary outcome is difference in WIC enrollment between the Clinic and Clinic-RDN models versus the Clinic-WIC and Clinic-WIC-RDN arms at 6-months post-baseline. Secondary endpoints include WIC retention and adherence, change in participant behavior, skills, and food security, preterm delivery, birthweight, and maternal and child health outcomes. Implementation outcome measures include acceptability, appropriateness, and feasibility from the perspective of clinic and WIC staff.

Discussion

Study findings will inform system models that integrate clinic care and social care to improve health equity among a high-risk population. Specifically, these findings will advance implementation of strategies to increase enrollment in a widely available but underutilized food provision program during pregnancy.

Trial registration

ClinicalTrials.gov identifier (NCT06311799). Registered 3/13/2024.

Background

The importance of food and nutrition security during pregnancy and early childhood has been long appreciated and well supported in the United States. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) was established 50 years ago as a food provision program for lower-income mothers and young children. Several systematic reviews have demonstrated associations with decreased adverse pregnancy outcomes, improved maternal and child diet quality, and reduced risk for obesity [1–3]. There is strong evidence that WIC improves food insecurity to protect against low-birth weight infants. Conversely, the risk of low-birth weight infants directly increases with rising food insecurity [4–6].

Alarmingly, WIC program enrollment is at a 30-year low [7]. In fact, less than half of eligible pregnant women are enrolled. Food insecurity rates among pregnant women are high nationally, and rural women have some of the highest rates of food insecurity [8]. Further, rural women are disproportionately at risk for adverse pregnancy outcomes and cardiovascular disease factors that persist following pregnancy. Adverse pregnancy outcomes such as gestational diabetes, preeclampsia, and pre-term delivery occur at significantly higher rates among rural women in comparison to their urban counterparts [9] and have been associated with increased risk of cardiovascular disorders [10, 11] and mortality [12].

One method to address these health disparities is by leveraging social care programs like WIC. Participation in WIC is associated with improved food security, birth outcomes, and maternal and child health outcomes [1–3, 13] and efforts to increase enrollment are warranted. Accordingly, the recent White House National Strategy on Hunger, Nutrition, and Health calls for strategies to increase access to federal food programs such as WIC [14].

The Food is Health movement calls for strategies to integrate food provision and health care, and parallel strategies are endorsed in health equity models [15, 16]. Integrating social care (e.g., WIC) into the delivery of healthcare can improve health outcomes, and in turn, reduce health disparities [17]. Several initiatives have highlighted this need, specifically calling for the integration of services to address social determinants of health into standard health care delivery and prioritize health equity [18–21]. Evidence is building to support these upstream multi-setting strategies. For example, a recent non-randomized study found clinic-social integration at the system-level to be effective [22], above and beyond what is typically seen with single setting, or siloed, interventions [23]. In specific, incorporating standard screening for food insecurity and embedding WIC referrals into the electronic health record (EHR) in a pediatric primary care clinic was associated with a 42% increase in WIC enrollment [22]. This increase is particularly compelling given that EHRs are widely used in clinical care delivery [24], which in turn supports the dissemination and implementation of this intervention model.

Yet, implementation and evidence gaps remain. Food and nutrition security discussions are novel and far from standard of care in obstetrics leaving questions about how to effectively engage clinical teams and patients in screening, preventive counseling, and referrals [25]. Although digital data sharing is a recommended feature of clinical-social integration frameworks [21], privacy and interoperability must be addressed to advance scalable solutions. In consideration of the intent to improve food and nutrition security, food provision may be enhanced by other health system resources. Registered dietitian/nutritionists (RDNs) may boost the impact of the Clinic-WIC model by improving nutrition outcomes and skill development. For example, food resource management, which includes skills like budgeting and meal planning, strengthens food security beyond WIC food provision in randomized [26] and non-randomized studies [27]. Food resource management has also been shown to promote healthy gestational weight gain [28] and is associated with reduced adverse pregnancy outcomes [29].

Pragmatic efforts to integrate clinical and community resources to reduce health disparities among a pregnant population are needed. Consequently, the primary purpose of the current study is to examine the effectiveness of a clinical and social care model, specifically integrating obstetrics clinical care and WIC, on WIC enrollment among rural pregnant women, a high-risk population that is underrepresented in research [30]. The study also aims to explore the feasibility and effectiveness of enhancing this model with food resource skill development led by a RDN. A third aim focuses on gaining a better understanding of patient, clinical, and WIC staff lived experiences to understand the acceptability of implementing these models.

Methods

Study overview

The present study is a pragmatic, randomized controlled trial being conducted in Geisinger prenatal clinics in Pennsylvania (USA). The clinics are located in regions with high rates of preterm birth, low birth weight, and infant mortality [31], supporting the need for upstream efforts to strengthen food security. Study participants are randomized to one of four intervention arms: Clinic Model (Arm 1), Clinic-WIC Model (Arm 2), Clinic-RDN Model (Arm 3), or Clinic-WIC-RDN Model (Arm 4). Randomization occurs at the participant level. Therefore, clinical teams may care for participants in any arm and will see varied alerts and features in the EHR, per randomization arm. The Clinic Model is consistent with standard prenatal care, which includes provision of basic information about WIC. In lieu of basic information about WIC, the Clinic-WIC Model involves a digital referral placed directly to WIC using basic features in the EHR. The Clinic-RDN Model includes provision of basic information about WIC and a digital referral using the EHR to a study RDN for heart healthy eating and food resource management counseling. In the Clinic-WIC-RDN Model, participants receive digital referrals to both WIC and RDN counseling. Including the WIC Model in the study design allows for evaluation of the impact of directly connecting patients to WIC on WIC enrollment rates and related secondary outcomes, such as food security. Inclusion of the RDN Model offers the added benefit of evaluating the independent and added impact of heart healthy eating counseling on food intake behavior [28, 29] and food resource management on food security [26, 27]. This study was approved by Geisinger’s Institutional Review Board and is registered with ClinicalTrials.gov (NCT06311799).

Sample size and participants

Approximately 240 low-income pregnant people will be recruited and randomized 1:1:1:1 to each arm. See Fig. 1 for CONSORT diagram with participants enrolled to date. Participants are blind to their randomized condition, and those assessing study outcomes will also be blinded. Sample size was determined assuming a Type I error rate of 5% and 80% power to assess a rate of WIC enrollment of 25% in the WIC information only group (Arm 1 and Arm 3) to 45% in the EHR referral group (Arm 2 and Arm 4). Participant eligibility criteria include the following: age 18 years or older, English speaking, confirmed pregnancy, intent to deliver at a Geisinger facility, and public or no insurance. The insurance criterion conveniently serves as a proxy for lower income status, as WIC serves families with household incomes up to 185% federal poverty level. Potential participants are excluded if they are not eligible for WIC, have pre-existing enrollment in WIC as a pregnant person, have private insurance, and/or are unwilling to participate for up to 12 months.

Fig. 1.

CONSORT Diagram

Providers, including Geisinger clinical teams delivering prenatal care (e.g., nurse care coordinators, clinicians) and WIC staff, are also invited to participate in the study by completing surveys and semi-structured interviews.

Recruitment and informed consent

Potential participants are pre-screened for inclusion criteria via the Epic^® EHR using established algorithms to identify lower income [32, 33], pregnant people [32, 34, 35]. All potentially eligible participants are sent a mailed recruitment letter, email, text message, and/or a patient portal message, depending upon patient documented preference, that provides information about the study, participation, how to opt-out, and a URL and/or QR code linked to a REDCap survey. The REDCap survey contains several screening questions, and if eligible, an electronic consent form. A reminder email, text, and/or portal message is sent 3–5 days after the initial contact attempt. Recruitment early in pregnancy is prioritized to offer greatest potential benefit of food provision and nutrition interventions [36]. Due to the limited recruitment window and timing of scheduled prenatal appointments, the study team contacts potential participants by phone who have not responded to assess interest. Once contacted by phone, the study team member discusses the study, and if interested, obtains verbal informed consent.

Study flow

Patient participants are enrolled in the study for no more than 12 months with active participation for approximately 6 months. Upon enrollment in the study, patient participants complete a baseline questionnaire and are randomized into one of the four intervention arms (Clinic Model, Clinic-WIC Model, Clinic-RDN Model, or Clinic-WIC-RDN Model). Upon randomization, the study team creates EHR encounters that launch varied alerts and features corresponding with the randomized condition. Based on clinical stakeholder input on design, these alerts are intended to cue action, making it easier for the clinical teams to act accordingly. Patient participants complete questionnaires again 6 months after baseline. Approximately 10% of participants will also participate in a semi-structured qualitative interview at this time to describe their lived experience. During the 12-months of enrollment in the study, clinical (i.e., EHR) and community (i.e., WIC) data will be accessed to allow for analysis of study outcomes. Providers (both Geisinger clinical team and WIC staff) are also enrolled in the study. Participation includes completion of a baseline questionnaire and intervention delivery. A subset (10%) of the Geisinger clinical team and WIC staff will be asked to complete a semi-structured interview.

Intervention components

Participants are exposed to two intervention components: WIC (information only or direct digital referral) and RDN heart healthy eating and food resource management counseling (yes or no). Each intervention component is described in more detail below.

WIC information only (clinic model)

The provision of basic information about WIC is a component of usual prenatal care at Geisinger (Arm 1 and Arm 3). Most (81% in 2023 per historic data) of the targeted study population will receive a bundled care program called Healthy Beginnings Plus (HBP). In Pennsylvania, lower income pregnant people are offered HBP, a program with the goal of meeting psychosocial needs in addition to traditional medical and obstetric services. At Geisinger, patients receiving HBP meet with a nurse at least 8 times for education and referrals relevant to the stage of pregnancy. During the first appointment, the nurse conducts a screening for social determinants of health and makes corresponding referrals. Referrals to services within Geisinger (e.g., nutrition, behavioral health) are digitized in the EHR; however, referrals outside of the healthcare system, specifically WIC, can be faxed, emailed, or telephoned. However, clinical staff who were consulted during the planning stage indicated a common practice of providing patients with verbal or written information only. The patient is expected to act on their own to pursue WIC enrollment. If a patient decides not to enroll in HBP, social determinants of health screening, referrals, and information sharing remain part of the prenatal workflow.

WIC digital referral (WIC models)

The WIC Model (Arm 2 and a component of Arm 4) is characterized by an external digital referral in the EHR to directly connect the patient with WIC. The intent of this component is to make it easier for patients to voluntarily enroll in the WIC program. In the current study, digital EHR-based referrals to WIC are facilitated using Neighborly (FindHelp^(C) 2023, Aunt Bertha, a Public Benefit Corp.), a social health access referral platform. Neighborly is HIPAA compliant, HITRUST certified, and has been implemented across the Geisinger enterprise for use by clinical staff. Clinical staff can access Neighborly within the EHR to identify social care programs, and with the patient’s verbal consent, make direct referrals to address unmet health-related social needs. Neighborly can be searched by zip code, social need (e.g., food security), and program name (e.g., WIC). Community-based social programs formally enroll in Neighborly (termed ‘claiming the card’) to become active and can receive referrals. Each active community-based social program has a password-protected dashboard that displays referred individuals. In turn, the community-based social program contacts the referred patient to evaluate needs, eligibility, and enroll in program. All local WIC agencies in the current study are active in Neighborly. Though clinical staff could access and submit digital referrals through Neighborly prior to study inception, feedback from the clinical staff indicated a lack of awareness of how to access Neighborly within the EHR and adoption was negligible prior to study inception.

Clinical team staff are prompted to submit a digital referral to WIC through a clinical decision alert in the EHR. Clinical decision alerts are a standard part of clinical workflow throughout the Geisinger clinical enterprise, and prenatal clinical team staff were familiar with their purpose and format prior to the study start. Design of the alert for the current study was informed by input from the HBP nurse supervisor and prenatal clinical team members were trained on their use for the current study.

For participants enrolled in Arm 2 or Arm 4, the alert is intended to nudge the clinician to make a referral to WIC. The alert in the current study is programmed to the patient’s zip code, thereby reducing the number of required clicks and offering efficiency with Neighborly. From 4/29/2024 to 6/19/2024, the study team monitored and provided feedback to the HBP nurse supervisor about adoption in clinical care and the HBP nurse supervisor coached nurses with no or low adoption to use the alert. However, due to low adoption (and therefore, low intervention exposure), the alerts were transitioned on 6/20/2024 to become hard stops, in that nurses are required to address the alert prior to the end of the clinical encounter. Due to the low adoption, a second region of obstetric clinics in northeastern Pennsylvania were added to the study to evaluate direct connection of participants to WIC as a population health approach to ensure participants have access to interventions after consent and randomization.

Data sharing

To facilitate WIC referrals, some information is shared externally with the local agencies that administer WIC services. WIC services at administered at the county-level by Family Health Council of Central PA in the Central region and by Maternal Family Health Services in the Northeast region. Information, such as the patient’s name and contact information, is shared with WIC through FindHelp as part of standard care to coordinate referrals. However, Epic is limited in that it cannot receive input data from FindHelp during standard clinical care. In other words, once a referral to WIC is submitted through Neighborly during standard clinical care, clinicians do not obtain confirmation the referral has been received or acted upon by WIC. Therefore, a data sharing agreement was pursued for the current study to allow WIC to share information with the study team. The following information is collected by and shared from Family Health Council of Central PA (i.e., WIC) to Geisinger to evaluate study outcomes: name, date of WIC enrollment, dates of WIC care (retention), and dates benefits were distributed and redeemed (adherence). The later addition of the northeast region to the study was a barrier to establishing a second data sharing agreement within the study period. Study participants will be asked about WIC enrollment in the 6-month questionnaire, but these data will not be verified by the local WIC agency in the Northeast region due to privacy protections.

Heart healthy eating and food resource management counseling (RDN model)

The RDN Model (Arm 3 and a component of Arm 4) is characterized by an in-house digital referral using the EHR to connect patients with a study RDN for heart healthy eating and food resource management telehealth counseling. The Arm 3 and Arm 4 referrals to RDNs are prompted using a clinical decision alert that is programmed to generate an in-basket message to the study RDNs. In turn, the study RDNs contact the participants to schedule intervention sessions. Similar to the process for the digital referrals to WIC, the alert process was informed by input from clinical team leadership and clinical staff were trained on the process prior to study inception. Due to low adoption, the clinical decision alerts were similarly transitioned on 6/20/2024 to become hard stops, and participants were directly connected to the RDNs by study team members if alerts were not acted upon. Study RDNs are assigned to Arm 3 or Arm 4, but do not cross-over to reduce contamination risk.

The goal of counseling is to make it easier for patients to modify their eating behavior with the support of counseling and skill development. Counseling is conducted via telemedicine once a month (up to 6 sessions) and consists of the following components: food resource management skills [26, 37], heart-healthy eating counseling [38], and provision of cooking utensils [39]. Study RDNs have prior experience delivering telemedicine interventions focused on food resource management and nutrition behavior [39–43]. The RDN intervention is innovative and distinct from billable nutrition services as food resource management is not recognized as medical nutrition therapy. Participants do not incur any costs for study-related activities and do receive renumeration for study questionnaires.

Measures

Table 1 provides an overview of study measures, the data source, and study time point. Measure selection was guided by study stakeholders, including the Geisinger Patient Advisory Council on Obesity (PACO) and the American Heart Association’s Health Care X Food initiative. PACO, which consists of a group of patients, provides feedback on research studies and initiatives at Geisinger Health System and has experience consulting on other research projects related to obesity and associated diseases [33]. The American Heart Association’s Health Care X Food initiative (the funding source of the current study) provided a set of common measures designed to facilitate the pooling and comparison of data across their funded studies [44].

Table 1.

Summary of Study measures

Outcome	Measure	Timepoint	Data Source
WIC Enrollment	Date of WIC enrollment	Intervention Duration	WIC Certification Dataa
WIC Retention	Dates of WIC appointments	Intervention Duration	WIC Certification Datab
WIC Adherence	Dates WIC benefits distribution and redemption	Intervention Duration	WIC Certification Datab
Social Determinants of Health	Food security, housing, transportation, childcare, employment, utilities, clothing, financial strain, connections, safety	As Available	EHR
Weight	Body mass index (kg/m2)	As Available	EHR
Hypertension	Systolic and diastolic blood pressure and/or diagnosis	As Available	EHR
Diabetes	HbA1c and/or diagnosis	As Available	EHR
Birth Outcomes	Preterm delivery, birthweight	As available	EHR
Healthcare Interactions	Adherence to prenatal and RDN appointments, medication orders, clinical referrals to specialty care or community programs	As Available	EHR
Food Security	6-item Household Food Security Survey	Baseline and 6-Mo	Patient Survey
Food Resource Management Skills	Self-efficacy in food resource management	Baseline and 6-Mo	Patient Survey
Diet Quality	Mediterranean Eating Pattern for Americans	Baseline and 6-Mo	Patient Survey
Binge Eating Frequency	Eating Disorder Examination Questionnaire	Baseline and 6-Mo	Patient Survey
Eating Competence	Satter Eating Competence Inventory	Baseline and 6-Mo	Patient Survey
General Health Status	EQ-5D-5 L	Baseline and 6-Mo	Patient Survey
Patient Experience	Net promoter score, perceptions of intervention feasibility and acceptability	6-Mo	Patient Survey and Interview
Provider Experience	Perceptions of intervention feasibility and acceptability	Baseline and Study Completion	Provider Survey and Interviewc

^a = Study participants enrolled in the Northeast region will be asked about this in the 6-month survey, but data will not be verified by the local WIC agency due to privacy protections

^b = These data will not be collected for participants enrolled in the Northeast region due to privacy protections

^c = The provider survey is administered at baseline after intervention training and qualitative interviews are conducted upon study completion

Patient measures

WIC Data

WIC certification data is obtained on a continuous basis from the Family Health Council of Central PA. This includes date of WIC enrollment, dates of WIC care (retention), and dates benefits were distributed and redeemed (adherence).

EHR Data

Patient health indicators and healthcare appointment adherence data will be extracted from the EHR. This includes social determinants of health screening data, vital signs (body mass index and blood pressure), labs (HbA1c), medications, and postpartum outcomes. Encounter data will also be obtained to evaluate prenatal and RDN appointment adherence.

Questionnaires

Patient participants complete questionnaires upon enrollment (i.e., baseline) and again at 6-months via REDCap. Sociodemographic characteristics are collected at baseline. Measures of food security [45], self-efficacy in food resource management [37], diet quality [46, 47], binge eating frequency [48], eating competence [49], and general health status [50] are collected at baseline and 6-months to assess changes throughout the intervention period. Patients also complete the net promoter score [51], a single item measuring the patient experience, at 6-months.

Semi-structured interviews

Semi-structured qualitative interviews will be conducted with a subset of patient participants following the 6-month intervention to describe their lived experience of participating in the trial and perceived challenges and benefits. Interviews will be conducted virtually by trained study team members.

Provider measures

Clinical and WIC staff are surveyed, and a subset will be invited to complete a semi-structured qualitative interview. A questionnaire is administered at baseline to assess sociodemographic factors and implementation-related outcomes such as acceptability, appropriateness, and feasibility of the intervention [52]. The survey was designed to assess factors that may moderate or mediate the interventions effect on primary outcomes, although the study is not fully powered to test these outcomes. A QR code and URL was displayed after training sessions to allow staff to complete the survey anonymously. At the end of the study, interested staff will complete a single semi-structured qualitative interview assessing their lived experience related to delivering the interventions.

Aim 1

The primary outcome is difference in WIC enrollment rates between the WIC information-only arms (Arm 1 and Arm 3) and the digital referral arms (Arm 2 and Arm 4) at 6-months. Differences in secondary outcomes (including WIC retention and adherence, food security, self-efficacy in food resource management, diet quality, binge eating frequency, eating competence, and indicators of health status) will also be examined at 6-months. Descriptive and inferential statistical analyses will be used to examine differences in these quantitative measures.

Aim 2

An evaluation of whether the primary or secondary outcomes differ between non-RDN (Arm 1 and Arm 2) and RDN models (Arm 3 and Arm 4) will be conducted. Similar descriptive and inferential statistical analysis approaches used to evaluate the primary aim will be employed. Sensitivity analyses will be employed to examine the impact of major adaptations to the intervention (e.g., implementation of hard stop alerts).

Aim 3

Factors contributing to intervention implementation will be assessed. Quantitative measures of acceptability, appropriateness, and feasibility will be evaluated using descriptive and inferential statistics. The deductive rapid analysis approach will be used to analyze qualitative data collected during semi-structured interviews. This approach shortens the timeline and resources needed for typical inductive analysis while maintaining rigor to evaluate intervention acceptability [53, 54].

Discussion

This protocol describes a pragmatic randomized controlled trial that aims to improve maternal health outcomes through the integration of WIC program referrals into standard prenatal care at a large integrated healthcare system. The study also explores the feasibility and effectiveness of enhancing this model with heart healthy eating and food resource management skills as well as implementation-related factors. Study findings will inform care delivery models and the integration of social care into routine healthcare.

The current study directly responds to a need for evidence of the most effective clinical care models to improve maternal health outcomes and benefits from its system-level approach. Most extant interventions focus on individual-level change, not system-level approaches. The approach in the current study incorporates EHR features in prenatal care to make it easier for clinical teams to connect patients to WIC, a widely available, federally-funded food provision program that is regularly updated based on scientific data to optimize nutrition and strengthen food security, and in turn, improve maternal and fetal outcomes [1–3]. Additionally, the inclusion of food resource management counseling in the study design allows for evaluation of the independent and added effects of food resource management skills on outcomes. Taken together, the effects of these interventions will offer insights into the most effective care delivery models for a population that is a high-risk and underrepresented in research [30].

The study design also benefits from its evaluation of feasibility and implementation-related factors. Effective interventions for increasing food security and improving maternal outcomes already exist nationwide [22], but they are not widely disseminated and underutilized. By conducting research that is informed by the lived experiences of patients, clinicians and WIC staff [55] and assessing feasibility and implementation outcomes, findings from the current study will maximize the potential population health impact of this intervention. Results will inform dissemination and implementation efforts for which care models are the most effective in real-world settings.

Abbreviations

WIC

Special Supplemental Nutrition Program for Women, Infants and Children

EHR

Electronic health record

RDN

Registered dietitian/nutritionist

HBP

Healthy Beginnings Plus

PACO

Patient Advisory Council on Obesity

Author contributions

The study design and concept were conceived by LBD, ARC, and ADM. LBD, KB, ARC, SL, ADM, KAM, KMHN, LW, AN, and CJS designed the process evaluation and GCW wrote the statistical analysis plan and conducted the sample size calculation. KMHN prepared the first draft of the manuscript, and all authors provided edits and critique. All authors read and approved the final manuscript.

Funding

This work was supported by the American Heart Association Grant # 24FIM1266978/ Bailey-Davis, L./2024.

Data availability

The materials and data from the current study are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Geisinger Institutional Review Board (# 2024 − 0102). Participant recruitment is ongoing. Informed consent will be obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.