Healthcare-community partnership to improve nutrition insecurity for optimal glycaemic control and pregnancy outcomes in women with pregestational diabetes: clinical trial protocol for the NOURISH study

Abstract

Introduction

Pregestational diabetes is one of the most frequent medical conditions in pregnancy. Nutrition insecurity is a non-medical, health-related social need and affects glycaemic management and pregnancy outcomes for women with both type 1 and type 2 pregestational diabetes. Collaborative healthcare-community partnerships to enhance food access, facilitate culinary medicine, diabetes self-management education and support (DSMES) and address unmet social needs for pregnant women with pregestational diabetes remain to be evaluated.

Methods and analysis

In a two-arm randomised controlled trial, we will examine the combined effects of a tripartite NOURISH intervention ((1) produce home delivery, (2) culinary medicine with DSMES and (3) community health worker-led social needs assessment and support) versus the current standard of diabetes and prenatal care. We will recruit and enrol 174 pregnant women (87 NOURISH, 87 standard care) ≤22+6 weeks of gestation with pregestational diabetes, inadequate glycaemic management (haemoglobin A1c≥6.5%) and self-reported risk of food insecurity from a diabetes and prenatal care programme at a tertiary care academic health system located in the USA. We will measure the primary outcome of glycaemic management by delivery (haemoglobin A1c<6.5%), and secondarily, adverse pregnancy outcomes (gestational weight gain and composite neonatal morbidity) and patient-reported outcomes (nutrition security). We will also use cost-effectiveness analysis to assess the intervention’s value relative to standard care and use process evaluation to characterise intervention and contextual factors influencing uptake, effectiveness and sustainability.

Ethics and dissemination

The Institutional Review Board at The Ohio State University approved this study (IRB: STUDY20260307; date: 23 April 2026). We plan to submit results of the trial for publication in peer-reviewed journals and presentations at international scientific meetings.

Trial registration number

NCT07560813.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• NOURISH is a healthcare-community partnership that focuses on reducing nutrition and food insecurity through home delivery of produce from a community foodbank, culinary medicine with diabetes self-management education and support, and social needs assessment and support for pregnant women with pregestational type 1 and 2 diabetes.

• Strengths of the NOURISH trial include an intervention focused on women with pregestational diabetes, poor glycaemic management and risk of food insecurity who are at high risk of adverse pregnancy outcomes; and a trial that integrates cost-effectiveness analysis and a process evaluation framework to identify barriers and facilitators to programme stability over time.

• Limitations of the NOURISH trial include that it is powered for glycaemic management and some pregnancy outcomes, but not maternal morbidity or postpartum outcomes.

• Challenges of the NOURISH trial include engaging a high-risk population of pregnant women with inadequate glycaemic management and unmet social needs.

Introduction

Background

Pregestational diabetes (type 1 and 2) is one of the most frequent comorbid conditions in pregnancy.¹ Every year pregestational diabetes complicates pregnancies of >100 000 women in the USA alone, and its frequency is anticipated to double in the next 10 years, affecting at least 5% of pregnant women.² Pregestational diabetes increases the risk of adverse pregnancy outcomes for both the pregnant woman (including severe maternal morbidity, caesarean delivery and pre-eclampsia) and the infant(s) (including large-for-gestational-age (LGA) at birth, preterm birth and neonatal hypoglycaemia).³ The risk of adverse pregnancy outcomes is over twofold higher with inadequate glycaemic management.^4–6 And, pregestational diabetes disproportionately affects pregnant women with unmet non-medical, health-related social needs (hereafter, social needs).^7–9

Nutrition insecurity, a concept at the intersection of food insecurity and poor diet quality, is a social need of critical concern for pregnant women with pregestational diabetes in the USA.^{10 11} Food insecurity affects one in five pregnant women,¹⁰ and poor diet quality affects one in two pregnant women.¹² Both of these factors increase the risk of inadequate glycaemic management as measured by haemoglobin A1c (A1c) and adverse pregnancy outcomes (figure 1).^13–17 Effective diabetes management requires a healthy diet and lifestyle in addition to medication.¹⁸ Nutrition and food insecurity are therefore critical intervention targets to achieve glycaemic management and improve pregnancy outcomes.^{11 19–21}

Figure 1. NOURISH conceptual diagram. APOs, adverse pregnancy outcomes; DKA, diabetic ketoacidosis; GWG, gestational weight gain; HDP, hypertensive disorders of pregnancy; LGA, large for gestational age; NICU, neonatal intensive care unit; PTB, preterm birth; SMM, severe maternal morbidity.

Pregnant women affected by food and nutrition insecurity are more likely to encounter multiple challenges, including transportation barriers, suboptimal food-related norms and values, unstable housing and less employment, all of which undermine their capacity to eat nutritious diets.²² Pregnant women can currently enrol in programmes that address food and nutrition security risk, such as the Special Supplemental Nutrition Programme for Women, Infants and Children (WIC) and Supplemental Nutrition Assistance Programme (SNAP). However, enrolment in such programmes has decreased over the past decade.^{23 24} Of these two programmes, SNAP reaches far more households across the USA but is not designed to encourage a healthy dietary pattern²⁵ and does not necessitate that participants engage in culinary or nutrition education or be assessed for co-occurring unmet social needs.²⁶

Healthcare-community partnerships are recognised as a feasible, acceptable and potentially cost-saving approach for addressing nutrition security among individuals living with diabetes.^{27 28} These partnerships are becoming more common through food-is-medicine initiatives in response to the growing recognition of the public health nutrition crisis; however, robust evidence of clinical impact during pregnancy is lacking.^27–30

A growing consensus recognises the feasibility, acceptability and potential cost-savings of healthcare-community partnerships to address nutrition security for individuals living with diabetes.^{27 28} These relationships are becoming more common through food-is-medicine initiatives; however, robust evidence of the clinical impact of these initiatives during pregnancy is lacking.^27–30

Given that unmet social needs are estimated to contribute to at least 50% of adverse pregnancy outcomes and that women with pregestational diabetes have some of the highest rates of adverse pregnancy outcomes,^{6 7 31} interventions that address food and nutrition security and related social needs in pregnancy are needed.²⁸ To date, there is a lack of adequately powered randomised controlled trials (RCTs) that examine the association of interventions that address food and nutrition insecurity on health outcomes in pregnant women with diabetes.^{29 32}

To address these gaps, NOURISH was designed to integrate three healthcare-community partnerships that already exist, and further modifies them for the specific context of diabetes and pregnancy. First, The Mid-Ohio Food Collective, a produce-distributing US foodbank that delivers >170 000 meals-worth of food daily through 680 agency partnerships, will coordinate produce delivery as frequently as every week.³³ Second, a culinary medicine curriculum that integrates the Ohio State University Cooperative Extension’s Dining with Diabetes programme (which is active across all 88 Ohio counties and delivered nationally via community-based extension network),^{34 35} and the more hands-on Cooking Matters for Diabetes programme, will be curated to build food preparation knowledge and skills along with diabetes self-management education and support (DSMES).^36–38 Third, Health Impact Ohio’s Central Ohio Pathways Hub (Hub), consistent with the Agency for Healthcare Research and Quality Pathways Community Hub model, will address unmet social needs via community health worker (CHW)-led referrals.³⁹

The theory-based NOURISH intervention is a healthcare-community partnership that focuses on food and nutrition insecurity through produce home delivery, culinary medicine with DSMES and social needs assessment and support for pregnant women with pregestational diabetes.

Aim and hypotheses

The objective of NOURISH is to test whether a healthcare-community partnership that includes three intervention components (produce home delivery; culinary medicine with DSMES; and social needs assessment and support) versus the current standard of care for prenatal and diabetes care results in improved glycaemic management and fewer adverse pregnancy outcomes for pregnant women with pregestational diabetes, inadequate glycaemic management and self-reported risk of food insecurity (table 1).

Table 1. NOURISH objectives and endpoints by study aim.

Objectives	Endpoints
Aim 1
Use an RCT to test the effect of NOURISH on improving glycaemic control by delivery (haemoglobin A1c) and secondarily, pregnancy outcomes (gestational weight gain, neonatal morbidity) and nutrition security	Haemoglobin A1c: <6.5% vs A1c≥6.5% will be assessed as the value most proximate to the time of delivery Composite neonatal outcome: includes one (or more) of the following: large-for-gestational-age at birth, hypoglycaemia, hyperbilirubinaemia, preterm birth <37 weeks and neonatal intensive care unit admission Nutrition Security Scale: continuous measure from 0 to 4 and as a binary variable classifying individuals as ‘nutrition secure’ (average score >2) or ‘nutrition insecure’ (average score ≤2)
Aim 2
Use cost-effectiveness analysis to assess the NOURISH intervention’s value relative to standard care	Direct costs: personnel and non-personnel costs of the intervention Indirect costs: costs that are not directly associated with the intervention but that support their delivery
Aim 3
Use process evaluation to characterise intervention and contextual factors influencing uptake, effectiveness and sustainability of the NOURISH intervention	Reach: representation of the target population within the study region Dose: degree of engagement in study activities Fidelity: adherence to intervention protocols Context: individual values and circumstances, and household and programmatic environment, sustainability and coordination from community stakeholders, and potential for scalability to broader population

RCT, randomised controlled trial.

Primary hypothesis

We hypothesise a relative risk that is 1.80-fold higher in the proportion of participants in the NOURISH intervention group with an A1c<6.5% by the time of delivery compared with the standard care group.

Secondary hypotheses

We hypothesise that individuals in the NOURISH intervention group will have fewer adverse pregnancy outcomes (maternal gestational weight gain within pregnancy guidelines and composite neonatal morbidity (ie, any of the following: LGA, hypoglycaemia, hyperbilirubinaemia, preterm birth <37 weeks and intensive care unit admission) and superior patient-reported outcomes (ie, nutrition security)).

Methods and analysis

Design

NOURISH is a randomised, controlled, single-centre superiority trial designed to determine whether pregnant women with pregestational diabetes, inadequate glycaemic management and risk of food insecurity enrolled in a healthcare-community partnership that includes three intervention components (produce home delivery from the foodbank; culinary medicine with DSMES; and social needs assessment and support) will improve glycaemic management by delivery and pregnancy outcomes compared with those receiving standard prenatal and diabetes care. In addition, NOURISH will employ cost-effectiveness analysis using a micro-costing approach to assess the intervention’s value relative to standard care with respect to study outcomes, and use process evaluation to characterise factors influencing uptake, effectiveness and sustainability of the intervention through implementation records, interviews with participants and partners, and a per protocol analysis. The overall approach of NOURISH is guided by the Prioritising health, Engaging the community, Targeting health disparities, Acting on the data and Learning and improving (PETAL) framework (figure 2). Each component of the proposed intervention is aligned with social cognitive theory (SCT).⁴⁰

Figure 2. NOURISH Petal framework. SA, specific aim.

Data management and coordination will be performed by an independent team at the clinical site of the NOURISH trial and will be led by the study statistician (GNB). Participant data will be collected, stored and maintained using Research Electronic Data Capture, a secure and confidential data management system. Randomisation by computer-generated random numbers, as directed by the study statistician, will be performed with a 1:1 non-blinded allocation between intervention:standard care groups. The allocation will be released to research staff electronically at the time of enrolling a participant to a study group.

Population

Women will be recruited as part of The Ohio State University Diabetes in Pregnancy programme, which provides integrated diabetes and prenatal care at a tertiary academic medical centre located in the Midwestern USA. Some women continue to receive routine prenatal care in their local community and then receive high-risk prenatal and diabetes care from the programme. These women will also be eligible for study participation. NOURISH builds on programmes that exist but are currently not combined into a multi-level intervention for pregnant women. In addition, these community partner-led programmes are not provided as part of routine diabetes and prenatal care.

Participation in the NOURISH trial will end at 6 weeks after delivery, at which time participants in both the intervention and standard care group will be provided with a referral to a primary care provider for diabetes care. After the trial completion, participants in either study arm can opt to access the existing version of the unintegrated intervention components as part of routine care through already existing care pathways.

Inclusion criteria

Pregnant women with an ultrasound-confirmed intrauterine pregnancy; between 8+0 weeks and 22+6 weeks gestation; aged ≥18 years; with a diagnosis of type 1 or 2 diabetes; with inadequate glycaemic management (if not taking a glucagon-like peptide-1 or sodium-glucose co-transporter 2 medication within 12 months of enrolling in prenatal care, haemoglobin A1c≥6.5%; or if taking a glucagon-like peptide-1 or sodium-glucose co-transporter 2 medication within 12 months of enrolling in prenatal care, haemoglobin A1c≥6.5% during the 12 months prior to initiation of these medications); screen positive for food insecurity risk based on answering ‘Often’ or ‘Sometimes’ true to either of the two questions on the US Department of Agriculture Hunger Vital Sign screening tool contained within the Accountable Health Communities Health-Related Social Needs Screening Tool^{41 42}; English or Spanish speaking; and willing to participate and provide a home address to which can be delivered.

Exclusion criteria

We will exclude women who have diminished decision-making capacity (ie, are unable to consent for themselves) or who are involuntarily confined or detained.

Recruitment

Pregnant women will be recruited over a 3-year period. Potentially eligible women will be identified via the electronic health record (EHR). In addition, individuals will be recruited using flyers placed in clinics and direct provider referrals from the community.

Baseline visit

A research assistant or healthcare provider will ask patients if they are interested in joining the study. Those who are will be referred for an orientation visit. Eligibility will be confirmed and informed consent will be obtained in person in English or Spanish (online supplemental file). Baseline data, including demographics and medical history, will be collected by electronic data capture. Participants will complete surveys at enrolment that assess baseline data about nutrition and food security, diet quality, home food environment, health literacy, social needs and diabetes distress and self-efficacy. If a haemoglobin A1c from the last 12 weeks is not already available, participants will complete a blood draw for measurement of haemoglobin A1c using a standardised point-of-care assay. Enrolled participants will be randomised in a 1:1 ratio to the intervention: standard care (figure 3).

Figure 3. NOURISH flow diagram.

Follow-up visits

Standard care includes referral to existing nutrition assistance programmes such as WIC and SNAP for eligible women. Standard clinical practice for diabetes management in pregnancy includes participant encounters every 2–4 weeks. Study staff will collect data generated from routine prenatal care, including glucose level, medication doses and healthcare utilisation (antepartum admissions, OB triage visits that have occurred). Maternal and infant EHR data will be collected through the final postpartum visit. Maternal blood will be collected for a point-of-care haemoglobin A1c measurement approximately once per trimester (randomisation, second trimester, third trimester), and post partum when possible, and will be analysed by personnel blinded to treatment assignment. Haemoglobin A1c will be measured using automated high-pressure liquid chromatography (VARIANT II TURBO HgbA1c Kit, Bio-Rad Laboratories, Hercules, CA, USA).

At follow-up timepoints (second trimester, third trimester and post partum), participants will complete surveys to assess: food security, nutrition security, social needs and diabetes distress and self-efficacy; at the postpartum visit they will additionally provide information regarding infant feeding practices.

Partial patient and public involvement

Pregnant women with pregestational diabetes and their healthcare providers, including physicians and nurses, registered dietitians, certified diabetes care and education specialists, clinical social workers and CHWs, have collaboratively worked with the study team in designing the NOURISH intervention by providing feedback via structured interviews.⁴³

Measures

The primary clinical outcome is the proportion of participants with a haemoglobin A1c<6.5% by delivery (see table 2 for all primary and secondary clinical and patient-reported outcomes). For those participants with more than one haemoglobin A1c assessment in late pregnancy, the value most proximate to delivery will be used.⁶ For most participants without a preterm delivery, this haemoglobin A1c assessment will be in the late third trimester. Haemoglobin A1c will be assessed once per trimester consistent with current guidance.⁴⁴ Of note, we selected a haemoglobin A1c threshold of <6.5% as opposed to <6.0%, as the latter may result in more frequent episodes of hypoglycaemia and the risk of adverse pregnancy outcomes is similar at both thresholds.^{3 6}

Table 2. Primary and secondary clinical and patient-reported outcomes.

	Randomisation	Second trimester	Third trimester	Delivery	6 weeks post partum
Demographics	X
Medical history	X
Glucose data abstraction	X	X	X		X
Point-of-care A1c	X	X	X
Qualitative interviews (subset)		X			X
Maternal and infant outcomes		X	X	X	X
EHR data abstraction	X	X	X	X	X
Patient-reported measures
Adherence, fidelity and responsiveness assessment (intervention group only)	X	X	X		X
Accountable Health Communities Health-Related Social Needs Screening Tool	X		X
Rapid Estimate of Adult Literacy in Medicine-Short Form	X
Nutrition Security Scale	X	X	X		X
Diabetes Distress Scale	X		X
Self-Efficacy for Diabetes	X		X
US Adult Food Security Survey Module	X	X	X		X
Mediterranean Eating Patterns for Americans (MEPA)-III	X		X		X
Infant Feeding Practices					X

Prespecified secondary outcomes for which this trial will be powered include maternal gestational weight gain, composite neonatal morbidity and patient-reported outcomes (ie, a standardised measure of nutrition insecurity). Gestational weight gain will be defined as the z-score of less than and in excess total pregnancy weight gain compared with within pregnancy recommendations (reference) per the Institute of Medicine guidelines.³¹ The composite neonatal outcome is defined to occur if a participant has at least one of the following: LGA birth weight >90th percentile, hypoglycaemia, hyperbilirubinaemia, any preterm birth <37 weeks and intensive care unit admission. This composite measure of neonatal morbidity is causally associated with maternal glycaemic management,^{6 45} consistent with measures used in recent diabetes pregnancy trials⁴⁶ and inclusive of outcomes recommended by a core-outcome set.⁴⁷

The prespecified patient-reported outcome for which this trial will be powered is the Nutrition Security Scale, which will be assessed as both a continuous (0–4) and categorical (>2 vs <2) measure.⁴⁸

Secondary clinical outcomes include maternal measures (ie, haemoglobin A1c assessed as a continuous percentage and <6%, hypertensive disorder of pregnancy, mode of delivery and obstetric anal sphincter injuries; individual neonatal measures: mechanical ventilation, small-for-gestational age at birth, respiratory distress syndrome); and the individual components of the composite neonatal outcome. Secondary measures of glycaemic control will include the percentage of time in range between 63 and 140 mg/dL, mean glucose and time spent above and below target range, all of which will be assessed by continuous glucose monitoring.

Secondary patient-reported outcomes include diabetes distress by the Diabetes Distress Scale-17,⁴⁹ unmet social needs by the Center for Medicare Services Accountable Health Communities Health-Related Social Needs Screening Tool,⁵⁰ severity of food insecurity per the US Adult Food Security Survey Module (HFSSM),⁵¹ diet quality per the Mediterranean Eating Patterns for Americans-III,⁵² health literacy per the Rapid Estimate of Adult Literacy in Medicine-Short Form⁵³ and diabetes self-efficacy.⁵⁴

For the intervention group, we will employ a mixed method process evaluation to characterise intervention reach, dose and fidelity separately for each of the three NOURISH components (table 1, aim 3). Sources of data will include implementation records, qualitative interviews and participant outcome data. Quantitative metrics will include representativeness of the intervention participants relative to the local population of pregnant women with diabetes (ie, reach); number and portion of culinary sessions attended, food boxes received and CHW engagements (ie, dose); and the team’s adherence to the study protocol in terms of the frequency, duration and composition of the food deliveries, culinary medicine sessions and CHW touchpoints. We will also assess intervention satisfaction via net promoter score. Qualitative interviews with cross-sector partners and participants will provide contextual data on factors influencing intervention enrolment (reach), engagement (dose) and fidelity of implementation. Dose data will be transformed and integrated with the full dataset of clinical and patient reported outcomes to afford a per protocol analysis and determine whether there was a dose–response relationship with primary and secondary study outcomes.

Procedures

Intervention Work Group

Because NOURISH is a collaborative and multi-stakeholder healthcare-community partnership, we will draw on user-centred design to coordinate and implement the intervention.⁵⁵ We will embrace the expertise of our Intervention Work Group members (ie, patients, clinical leads and community partner representatives). This collaborative approach contributes to the sustainability and scalability of NOURISH.⁵⁶ These sessions will leverage elements of user-centred design, a methodological process for exploring the root causes of a major challenge from multiple ‘user’ or stakeholder perspectives. These sessions will be guided by the Collaborative Intervention Planning Framework.⁵⁵ We have designed our Intervention Work Group to include members with prior cross-sector relationships that support healthcare delivery and to assist with interpretation of findings in a manner that supports: (1) strengthened and sustained relationships between our healthcare system and the community partners and (2) translation to other regions and sets of partners.

Interventions components

We describe each of the three healthcare-community partnerships of the NOURISH intervention in the following sections.

Produce delivery

The Mid-Ohio Food Collective is a regional foodbank that delivers >170 000 meals-worth of food daily through 680 agency partnerships. Existing partnerships exist between the Mid-Ohio Food Collective and clinics in our healthcare system through the Farmacy Ohio programme, but referrals to the food pantry system are not routinely provided as part of prenatal care.⁵⁷ Prior data have identified institutional support, efficient screening and referral processes and positive programme expectations as important facilitators supporting implementation of the Farmacy Ohio as part of care delivery.^{27 58 59} Farmacy Ohio uses a referral from a patient’s healthcare provider to connect the patient to a network of over 20 affiliated food pantries in the metropolitan area. In NOURISH, the goal of Farmacy Ohio is to address food and nutrition insecurity by connecting participants with fresh produce via home delivery. Farmacy Ohio’s services may not be optimally effective at addressing the social need of transportation, which frequently co-occurs with food insecurity in pregnancy.⁶⁰ In NOURISH, we will extend the existing Farmacy Ohio programme by providing up to weekly home delivery of produce from the Mid-Ohio Food Collective.

Weekly Farmacy Ohio deliveries will include 15–20 pounds of low-to-medium glycaemic index fruits and vegetables, as recommended by the USDA Thrifty Food Plan, from available inventory. An example box would include 8–9 pounds of vegetables (~3 cup equivalents per day, including but not limited to tomatoes, cabbage, onions, peppers, sweet corn, carrots and zucchini) and 6–8 pounds of fruit (~2 cup equivalents per day, including but not limited to peaches, plums, apricots, nectarines, citrus fruits, apples, pears, grapes and berries).

In NOURISH, the produce delivery will continue through 6 weeks post partum, after which participants may use the current Farmacy Ohio programme wherein individuals visit a food pantry to receive their fresh produce and other food items.⁶¹ These deliveries are not meant to replace normal grocery shopping or use of other benefit programmes (ie, WIC, SNAP), but rather to complement those food sources and support the feasibility of healthy eating.

Culinary medicine with DSMES

Culinary medicine sessions will focus on food preparation, nutrition facts, meal planning, budgeting and shopping, as well as developing skills for produce preparation and integration into meals. These sessions will be delivered at least monthly using virtual and in-person participant engagement. These educational modules will integrate key elements from two evidence-based, currently available programmes: Dining with Diabetes, which is offered by OSU Extension’s community-based education programme, and Cooking Matters for Diabetes, an evidence-based adaptation of the standard Cooking Matters curriculum developed by Share Our Strength that focuses on how to make nutritious meals on a budget.^36–38

Prior data from outside pregnancy suggest that such programmes result in improved health behaviours, patient-reported outcomes and glycaemic management,^{34 62} as well as high participant uptake and acceptability.^{36 37} In NOURISH, we will further adapt these culinary medicine curricula from prior programmes to primarily be delivered virtually. The final, integrated curriculum will include nutrition education, culinary education and DSMES.

Social needs referrals

Social needs assessments will be done in partnership with Health Impact Ohio’s Central Ohio Pathways Hub, which operates using three features: (1) the Hub, a regional coordination entity, which employs CHWs to assess the social needs of patients and connect them to community resources; (2) CHWs, who initiate a ‘social need care pathway’, a defined action plan to address each patient’s unique needs, which is recorded and tracked in an electronic database; and (3) linkage of completion of each pathway to payment from insurance companies based on specific performance benchmarks.^{63 64} The CHWs are embedded within the local communities in which participants live. Participants will be screened at enrolment and follow-up for social needs.⁵⁰ Once an unmet social need (eg, food insecurity, unstable housing, unemployment) is identified, participants will be referred to the Hub through the provider dashboard to address it. CHWs will perform a comprehensive social needs assessment and connect the participant to community resources through Hub pathways.

Standard care arm

Participants randomised to the standard care group will receive the current standard of diabetes and prenatal care at our centre as part of the integrated diabetes and prenatal care programme.^{5 6} Participants will complete weekly glucose monitoring, which will be electronically sent to the EHR for provider review. Haemoglobin A1c will be assessed, and patient-reported outcomes will be collected at prespecified intervals similar to the intervention arm (ie, approximately once per trimester).

Statistical analysis plan

Baseline characteristics

We will provide descriptive summaries and examine any potential differences in sociodemographic and clinical characteristics between the intervention and standard care groups using χ² or Fisher’s exact tests for categorical variables and independent two-sample t-tests for continuous variables.

Primary and secondary hypothesis

Primary analyses will follow the intention-to-treat principle in which pregnant women will be analysed in the group to which they were randomised, regardless of whether they receive the assigned intervention. As an RCT, we will not adjust for baseline participant characteristics or engagement measures in the primary intention-to-treat analysis unless we identify statistically significant differences between the intervention and standard care groups. No interim analyses are planned. Multiple imputation will be considered for missing covariates.

Preplanned post hoc analyses

We will conduct interaction and, if indicated, subgroup analyses to determine whether and how the effect of the intervention differs among subgroups. Variables to be evaluated include type of diabetes (type 1 vs type 2 diabetes), severity of baseline food insecurity per the HFSSM (very low vs low or marginal), body mass index at randomisation (≥30 vs <30 kg/m²) and gestational age at randomisation (>14 vs <14 weeks).

Power calculations and sample size

The sample size of 174 participants total (87 per study arm) was based on the ability to detect a relative risk increase of at least 1.80 in the proportion of participants in the intervention group with a haemoglobin A1c<6.5% by delivery compared with the proportion (32%) in the standard care group, with a power of 0.90 and a two-sided alpha of 0.05, and a 10% loss-to-follow-up. There is no planned interim analysis given the intervention involves a nutrition intervention, education and social needs referral without significant risk of patient harm.

Compensation

Participants in both study arms will receive compensation of up to US$200 for completion of study activities, including surveys and attending data collection sessions. Cross-sector partners and participants who complete interviews will receive US$30 for participation.

Ethics and dissemination

The OSU Institutional Review Board (IRB) has approved this protocol. All protocol amendments will be communicated for approval to the OSU IRB. We will follow CONSORT (ie, Consolidated Standards of Reporting Trials) guidelines. We will submit study results for publication in peer-reviewed journals and presentation at international meetings. We will attempt to publish all findings in open-access journals when possible, or in other journals with a concurrent uploading of the manuscript content into PubMed Central for public access. Curated technical appendices and statistical code will be made available from the corresponding authors on request.

Discussion

In this RCT, we will examine the effect of a multi-component health system-community partnership and intervention (produce home delivery from the foodbank, culinary medicine with DSMES and social needs assessment and support) versus the current standard of care to improve glycaemic management and pregnancy outcomes among women living with pregestational diabetes and experiencing food insecurity.

Improving glycaemic management and pregnancy outcomes through addressing nutrition and food insecurity

Evidence-based diabetes and prenatal care can improve glycaemic control and pregnancy outcomes, but current strategies do not directly address social needs. The NOURISH intervention aims to improve patient self-efficacy through a primary focus on nutrition and food insecurity. First, Farmacy Ohio will provide diabetes-appropriate food with an aim to mitigate the negative effects of nutrition insecurity on glycaemic control in pregnancy, and will seek to mitigate transportation as a major barrier during the trial by providing home delivery.⁵⁷ Second, culinary medicine sessions will focus on theory- and evidence-based education critical to diet-related self-care behaviours. Third, Health Impact Ohio’s Central Ohio Pathways Hub will provide CHW referrals to address unmet social needs. In combination, this tripartite and integrated intervention may exceed the efficacy of clinical care and treatment alone, and may provide a sustainable method for improving diabetes self-management in pregnancy and beyond.

Theory-driven intervention that includes an RCT, cost-effectiveness analysis and process evaluation framework

NOURISH is an SCT-aligned pragmatic trial with scalability and potential to reach the broader population of pregnant women with unmet social needs. We will evaluate the integration of existing community-based interventions that have been implemented by our partner organisations (for enhanced feasibility) and that are increasingly available across the USA (for enhanced scalability).

In addition, NOURISH includes a cost-effectiveness analysis of the three complementary components. Evidence on the cost-effectiveness of community interventions for diabetes in pregnancy is lacking.⁶⁵ While some studies have examined the cost-effectiveness of diabetes self-management and lifestyle interventions outside of pregnancy, they have not studied interventions that address nutrition insecurity in pregnancy.⁶⁶ We will collect cost data prospectively during the RCT and assess cost-effectiveness from the perspective of a programme implementer.⁶⁷ This analysis will elucidate resource needs to implement and replicate NOURISH.⁶⁸

Finally, NOURISH includes a mixed-method process evaluation to inform sustainability by identifying barriers and facilitators to programme implementation and stability over time. By leveraging diverse sources of data and perspectives, we will be able to study the potential replicability and scalability of NOURISH. The process evaluation will monitor and explore implementation of this multi-modal intervention to identify within- and cross-sector strategies for sustainable and mutually beneficial implementation and impact.³⁷

Limitations and strengths

Limitations

First, there is a possibility of contamination across study arms where those in the standard-of-care arm may be more likely to be referred to community-based organisations to address social needs because they participate in the trial. However, we believe that this contamination will be unlikely as the services offered as part of NOURISH are not offered routinely by providers as part of our diabetes and prenatal care programme. In addition, NOURISH will evaluate the actual services received on study outcomes to estimate the effects of any contamination. Second, many NOURISH participants, if eligible, may enrol in current programmes that address nutrition and food insecurity, such as WIC and SNAP, that may confound or modify the effect of the intervention.²³ We will collect WIC and SNAP data, and consider time-dependent adjustment and stratification by WIC and SNAP status in analyses. Currently, such existing programmes neither provide the NOURISH intervention (produce delivery, intensive education, and assessment and referral for social needs) nor fully mitigate adverse outcomes and disparities. Third, individual and household cultural beliefs, preferences, and practices may influence eating patterns in, and beyond, pregnancy. We will assess food preferences by survey and qualitative interviews and culinary medicine sessions will reflect diverse cultural and food preferences.

Strengths

Nutrition and food insecurity are a critical public health issue. Pregnant women with pregestational diabetes, poor glycaemic management and food insecurity have some of the highest rates of adverse pregnancy outcomes and may benefit from an innovative and sustainable healthcare-community partnership that addresses nutrition and food insecurity along with related social needs. NOURISH will test a model for a scalable partnership with enhanced scalability to address nutrition and food insecurity among high-risk women with pregestational diabetes to improve pregnancy outcomes.