4

Exploring Opportunities for Innovation

There are a number of approaches that can be used to understand the weight-related issues around pregnancy and birth and to use this understanding to create interventions that aim to improve outcomes for mothers and infants. In this session, speakers discussed how to address perinatal health and associated weight-related issues using public health frameworks, community-oriented interventions, data aggregation and linkages, systems science, and clinical care.

PUBLIC HEALTH FRAMEWORK

The public health framework recognizes the upstream, midstream, and downstream contributors to health outcomes, said Leah Lipsky, National Institutes of Health. Upstream factors include economic, social, and physical environments; midstream factors are population-level behaviors; and downstream factors are individual-level health services and clinical care. Interventions can occur at any of these levels. Upstream interventions could include subsidies or taxes, food labeling regulations, or land-use policies. Midstream interventions target specific communities or populations, and could include school nutrition and physical activity policies, workplace wellness campaigns, or educational efforts. Downstream interventions, targeted at individuals, include initiatives like clinical practice guidelines, or increasing dietitians on staff in a clinic.

Individual-Level Interventions

Lipsky noted that while the public health framework provides a comprehensive structure for addressing weight and perinatal health at multiple levels, in practice most public health actions remain focused at the downstream, individual level. A review of public health approaches in this area found that most efforts were guidelines and recommendations aimed at supporting providers in advising patients on the importance of healthy gestational weight gain (GWG) (Kalantari et al., 2024). These guidelines generally give recommendations in multiple areas, including providing counseling, supporting individual strategies, addressing barriers, monitoring, and being sensitive to weight-related concerns. Similarly, the U.S. Preventive Services Task Force (USPSTF) recommendations advise health care providers to offer “effective behavioral counseling interventions aimed at promoting healthy weight gain and preventing excess gestational weight gain in pregnancy” (Davidson et al., 2021). USPSTF suggests advising patients through in-person appointments, refering to other resources, and using other methods such as technology-supported coaching.

Lipsky said that these types of individual-level GWG interventions have been shown to have health benefits, including reduced risk of emergency cesarean delivery, macrosomia, and large-for-gestational age (LGA) (Cantor et al., 2021). They are also cost-effective; one study found that every $1 spent on intervention saved $4.75 through reduced adverse events (Lloyd et al., 2022). Meta-analyses have attempted to isolate the components of interventions that yield the greatest benefit (Teede et al., 2022). When all interventions are taken together, the mean decrease in GWG is about 1 kilogram. When separated by behavioral target, the largest effect is found with diet-only interventions, with a decrease of 2.63 kilograms. Research has also found that diet-only interventions are most likely to produce GWG within the recommendations. The smallest effect is seen with interventions that offer unstructured support or provide only written information, with or without weight monitoring. Some evidence about intervention delivery methods suggests that individual, face-to-face delivery by allied health staff is most effective, said Lipsky; however, there is insufficient evidence on optimal intervention content.

Lipsky shared ways in which these individual-level public health approaches—mostly clinical practice guidelines—fall short. There is a need for more specific weight management guidelines, evidence-based intervention strategies, support for implementation, community-level policies, and training and capacity building for health care providers. Clinical practice guidelines need to be supported by evidence from rigorously tested interventions in multiple settings. Many interventions that have been studied are those delivered in early pregnancy, in a clinical setting, and which target

individual behavior change. The evidence is far more limited for interventions that occur before pregnancy. Further, there is a lack of consensus on the precise weight goals for individuals with obesity, and the optimal timing for interventions.

Community-Level Interventions

In addition to individual-level interventions, said Lipsky, there are broader public health efforts to target weight-related behaviors. For example, the Diabetes Prevention Program (DPP) began as a randomized clinical trial testing the effectiveness of a lifestyle intervention compared to metformin; it included a 16-lesson curriculum over 24 weeks, followed by monthly individual sessions and optional group sessions. Participants were encouraged to get at least 150 minutes of physical activity per week and to eat a healthy, low-calorie, low-fat diet. After the lifestyle intervention showed benefits to both weight loss and a decrease in risk of diabetes, the Centers for Disease Control and Prevention (CDC) launched the National DPP (NDPP) to bring the benefits to a wider population. There are now 1,500 organizations that deliver the intervention, reaching 1 million Americans across all 50 states. This large-scale, community-based public health initiative has demonstrated health benefits and cost-effectiveness; enrollment in the program has been estimated to have an 88 percent chance of saving costs for adults with prediabetes, with a savings of $4,600 per person and $160,000 per case of diabetes prevented.

There have been trials that have used elements of the DPP to limit excessive GWG, said Lipsky, and they demonstrated effectiveness in this population as well (Brokaw et al., 2018; Peaceman et al., 2018; Ritchie et al., 2023). There is a study under way that is testing the usefulness of the NDPP as a preconception intervention for women with overweight and obesity (Sauder et al., 2023). Measured outcomes will include postconception body mass index (BMI) and fasting glucose and neonatal adiposity. In addition to these uses of the NDPP in the preconception and pregnancy periods, the American Journal of Obstetrics and Gynecology published a clinical opinion recommending that all postpartum women with a history of gestational diabetes participate in a DPP intervention (Henderson et al., 2023). Lipsky noted that if postpartum women lose weight through the DPP, they will be in a healthier position for future pregnancies. Given that the DPP is already a successful public health approach for targeting weight behaviors and reduction in adults, she said, it is likely a promising option as a public health approach to improve pregnancy-related weight outcomes.

Environmental Interventions

Lipsky said that in her review of the literature, she found no environmental interventions targeting prepregnancy BMI or GWG. There are multiple challenges involved in implementing environmental interventions, she said, including influencing multiple environments, collaboration among multiple stakeholders, and a long timeline to measure its effect. There is one child obesity intervention—Shape Up, Somerville—that has achieved this level of complexity in a research setting, but it has not been replicated elsewhere; this experience highlights the real-world difficulty of environmental approaches. Another environmental approach to target obesity is food labeling, said Lipsky. For example, a Chilean food labeling initiative required front-of-package labels for foods high in sugars, saturated fat, sodium, and calories; prohibited marketing with cartoon characters and ads on TV for these foods; and banned these foods in schools. Evidence shows that the law has resulted in reformulation of products, a reduction in child and adolescent exposure to unhealthy food advertising, and a reduction in purchases of unhealthy foods. This is an example, said Lipsky, of an upstream environmental approach for changing the whole food environment in multiple contexts, with a measurable effect on individual behaviors.

In summary, said Lipsky, despite advances in clinical and public health approaches, improving prepregnancy BMI and achieving healthy GWG continues to be challenging. Current public health efforts largely focus on clinical guidelines for health care providers, with limited community and environmental interventions. The evidence supports the effectiveness of individual behavioral interventions, but gaps remain in optimal implementation, specificity, and reach. While there are no existing environmental interventions specifically aimed at GWG, promising models like the National DPP and the Chilean food labeling and marketing regulations demonstrate the potential for scalable public health approaches that could address weight and metabolic health before, during, and after pregnancy.

COMMUNITY: FOOD AS MEDICINE

The Ceres Community Project is a “community-based, food-as-medicine organization,” said Cathryn Couch, Ceres Community Project. It serves four counties in Northern California, and provides medically tailored meals, medically tailored groceries, and produce “prescriptions.” The people served by Ceres are primarily female (63.1 percent), over 60 years old (65.2 percent), and have a low income (81.6 percent). In 2021, Ceres implemented a pilot project focused on perinatal health with a goal to demonstrate cost-effectiveness for the local Medicaid Managed Care Plan. Couch said that this was a great opportunity to implement an upstream

intervention that via improvements in birth outcomes would also show a short-term return-on-investment for a health care payer. The project involved partnerships with two federally qualified health centers (FQHCs), one in an urban/suburban area and one in a very rural part of the county. Participants were low-income pregnant women who were eligible for the Supplemental Nutrition Assistance Program (SNAP). The pilot project was integrated into an existing Comprehensive Perinatal Services Program, said Couch, which provides wraparound support during pregnancy and the postpartum period, including lactation support and nutrition education. Women were enrolled when they first accessed prenatal care, which for most women was in the 2nd trimester.

A total of 227 participants were enrolled; around two-thirds were Hispanic/Latina, and they ranged in age from 18 to over 40. The intervention consisted of home delivery of medically tailored meals for all family members for 4 weeks, followed by a weekly produce bag until birth; this component began as soon as women enrolled in the program. After birth, the family received another 5 weeks of medically tailored meals for all family members. Couch said that medically tailored meals are a “really exceptional teaching tool.” When food is already prepared and is attractively presented, people are more likely to eat it. Prepared meals also expose people to new and different foods and give them a sense of healthy serving sizes and balanced meals. Couch added that the meals incorporated culturally appropriate foods in both the meals and produce bags, client care staff were bicultural and bilingual, and all materials were available in English and Spanish.

Early data found that the intervention was beneficial on multiple levels. Around 38 percent of participants reported an improvement in their health, and the share of participants who rated their health as “fair” or “poor” declined 41 percent. Increased vegetable consumption was reported by 41 percent, and measures of food security improved. There were also improvements in birth outcomes, with 19 percent of participants less likely to have infants with low birth weight compared to Medi-Cal members.¹ Participants were also 35 percent less likely to have preterm births compared to Medi-Cal members. Couch noted that 40 participants dropped out of the intervention early; reasons provided included not liking the food, the delivery time did not work, lost contact, or pregnancy complications.

There were several important lessons learned from this pilot project, said Couch. First, partnering with FQHCs worked well, particularly given the existing structure of the Comprehensive Perinatal Services Program. Second, they learned that pregnant women vary widely in their needs and circumstances. Some moms were dealing with nausea or bedrest, while

___________________

¹ Medi-Cal is California’s Medicaid program.

others were feeling healthy and working; some wanted prepared meals every week, while others were happy with just the produce bags. It is important to acknowledge these differences and give participants a choice in how an intervention is delivered. One challenge in the project was getting people to come to in-person classes; they found that written information was preferable for some participants. Couch said that this pilot project demonstrated the potential for building sustainable interventions into the health care system and showing a positive return-on-investment for Medicaid and other health plans.

She closed with a story of one participant named Leticia, a 40-year-old woman with an at-risk pregnancy and diabetes. Leticia was “incredibly grateful” to be enrolled in the program and reported that the access to high-quality food, tailored to address her diabetes, helped ensure her health and that of her family.

DATA AGGREGATION AND LINKAGES

There are opportunities to study prepregnancy BMI and GWG using big data sources such as vital statistics data, medical claims data, and electronic health record (EHR) data, said Stephanie Leonard, Stanford University. However, there are also limitations to these approaches; Leonard walked workshop participants through the advantages and disadvantages to using each type of data.

Vital Statistics Data

Vital statistics data are data collected on Certificates of Live Birth and Fetal Death. Typically, these forms are completed during the birth hospitalization and then reported to state health departments and the U.S. National Vital Statistics System, part of CDC. Leonard noted that CDC revised forms in 2003 to include fields for height, prepregnancy weight, and weight at delivery; these changes were fully implemented in the United States in 2016. There are many strengths to vital statistics data. There is both a large sample size and high generalizability; in the context of weight, these data can be used to make estimates and monitor trends in prepregnancy BMI and GWG. Another strength, said Leonard, is that these data can be linked to other robust data sources, such as hospital records and birth and death certificate data. By linking these data, a researcher can create a linked birth cohort of an entire state. Some have linked multiple states together, which creates a very large, robust data system. These large datasets can be used to assess serious but relatively rare outcomes. For example, the risks of severe maternal morbidity are higher for women with severe obesity, but the incidence of severe maternal morbidity is only around 1 percent, and

the prevalence of severe obesity is about 3 percent. Detecting and assessing these outcomes, said Leonard, requires a huge sample size. The final strength of vital statistics data is that they contain a number of useful data fields with high validity.

In the context of pregnancy weight, said Leonard, there are two major disadvantages of vital statistics data. First, the validity of weight measures is mixed. Several studies have found that prepregnancy BMI and GWG measures are largely consistent with medical records, but the validity varies by patient characteristic; specifically, there is lower validity for Black women compared to White women (Bodnar et al., 2014; Deputy et al., 2019). Further, a study found that people tend to underreport their prepregnancy weight and weight at delivery, while overreporting GWG (Headen et al., 2017). This leads to moderate misclassification of BMI and GWG categories; however, Leonard noted that these misclassifications do not bias the association between weight and outcomes. The second disadvantage of vital statistics data is that it is sometimes challenging or impossible to link the data to other datasets. For example, a state may prohibit this use, or it may be prohibitively costly. Leonard noted that when linkage is possible, it is probabilistic linkage rather than deterministic linkage because of the lack of a universal form of identification for people in the United States.

Medical Claims Data

Medical claims data contain patient encounter information, including diagnoses and procedures, said Leonard. They may also provide information on billing and costs, demographics, enrollment in an insurance program, medications, and laboratory values. Commonly used nationwide claims datasets include Medicaid research files, Merative MarketScan, Optum Clinformatics, Veterans Affairs, National Inpatient Sample, and others. Leonard listed the five main advantages of medical claims data:

1. They have large, generalizable patient groups.
2. Data are often longitudinal.
3. Detailed information on diagnoses, procedures, and often medications is included.
4. Multiple types of pregnancy outcomes are included (not limited to live births).
5. Data can be linked to other datasets such as vital statistics.

There are, of course, disadvantages of medical claims data as well, she noted. There is limited information on weight, height, and demographic details; the data often only include ICD-10-CM codes for BMI group. There are barriers to linking mother–child pairs in medical claims data, and there

is lag time between data collection and research time. Further, medical claims data can be expensive, or it can be difficult to obtain access. Finally, managing, storing, and analyzing the huge amount of data can be challenging, given that these records capture every encounter for every patient.

Electronic Health Records

EHRs capture a number of important data points for pregnancies, said Leonard. At most prenatal visits, the patient’s weight is measured by a health care professional and recorded in the EHR, giving the data high validity. Many EHR systems aggregate and share their data for research purposes, and some systems have developed easy-to-use interfaces for researchers to access the data and search for specific information. Another advantage of EHRs is that they can have a large sample size and high generalizability if there are data from multiple sites. There are several ways to accomplish this combining of data, said Leonard. A data coordinating center can gather data from multiple sites and combine the data into a robust dataset. A new study approach involves distributed data networks that use common data models to combine data from multiple institutions. Each institution converts its EHR data into a common data model format, runs an analysis, and sends the aggregated data to one site that combines aggregated results into a meta-analysis. This approach makes it possible for institutions to contribute data to research without sharing any patient-level data.

Finally, said Leonard, centralized EHR datasets have been developed; for example, Epic Cosmos combines data from participating health systems into one massive dataset with 300 million patient records. There are downsides to EHR data as well. Height and weight data can be missing and are more likely to be missing in the records of healthier people. It can be challenging to identify and correct erroneous weight values and may require looking at multiple records to identify mistakes. While weight is usually measured at the first prenatal visit, and this value is used as a proxy of prepregnancy weight, the timing of this visit does not always happen at the beginning of pregnancy. Similarly, the weight at the last prenatal visit is used as a proxy for delivery weight but may not be accurate. Depending on the timing of these visits, using prenatal visit weight measures as proxies could result in a higher measure of prepregnancy BMI and a lower measure of total GWG.

In conclusion, Leonard said that big data sources have advanced considerably in recent years, and they have the important strengths of large size and generalizability. However, researchers and policymakers need to consider the specific strengths and weaknesses of each of the big data sources for studying prepregnancy BMI and gestational weight gain. Moving forward, there is a need to improve the validity of measures in big data

sources. Specifically, there is a need to validate measures for pregnancy trajectory of weight and to develop methods to improve the internal validity of research using maternal weight, height, and BMI values from big data sources.

SYSTEMS SCIENCE

As other speakers have noted, there are a wide array of determinants that shape outcomes before and during pregnancy, said Irene Headen, Drexel University. These determinants are embedded within systems operating at multiple levels, and birthing people themselves are embedded in multilevel social environments. Further, factors across social environments are usually not independent but are instead interrelated, interdependent, and cumulative. Improving outcomes for this population, said Headen, requires understanding the multiple factors at different levels and how they relate. For example, housing quality, neighborhood characteristics, and the built environment may each be a factor that affects perinatal health; however, they are interconnected and move together in addition to their independent contributions. Traditional epidemiologic approaches do not support the ability to analyze these interconnections, but systems science provides a way to understand and analyze this interdependence between factors. Systems thinking is a paradigm that shifts thinking away from individual factors and toward a perspective of interconnectedness, circularity, feedback, and emergence. In other words, said Headen, “What happens when an outcome is more than just the sum of its parts?”

There are a range of different tools under the systems science umbrella. These include agent-based modeling, in which individual decisions interact to result in behavior of the system and an eventual outcome; social network analysis, in which dynamic relationships between actors in a network give rise to a trend in outcomes over time; and system dynamics (SD), in which feedback among multiple factors in a system create a trend over time. The formal definition of SD, said Headen, is the use of informal maps and formal models with computer simulation to uncover and understand sources of feedback in systems behavior (Hovmand, 2014). There are both qualitative and quantitative approaches within SD. Qualitative approaches use visual diagrams to represent interdependent relationships and feedback, while quantitative approaches use mathematical modeling in the form of systems of equations to simulate the system of interactions generating behavior over times. In addition, there are participatory and nonparticipatory versions of SD. Participatory approaches blend community-based participatory research and qualitative data collection methods in order to better understand community perspectives on a problem, and to identify ways to intervene in the problem. Nonparticipatory approaches can rely on

a wide array of secondary data including literature reviews, case studies, and expert input to create a qualitative or quantitative model.

Systems Science and Gestational Weight Gain

Turning to the application of systems science to GWG, Headen said that there is a small body of literature that has started to use SD approaches to better understand pregnancy-related weight. One study used quantitative SD to create a model of female obesity, age, and fertility (Sabounchi et al., 2015). The study considers the dynamic cost-benefit approach that individuals with obesity may navigate in choosing to either lose weight in order to improve their fertility, but because of the aging process, their fertility may decrease during the time that it takes to lose the weight or pursuing pregnancy at their current weight and potentially encountering costly fertility and other pregnancy complications down the road. The dynamic model contextualizes factors within both the biophysiological and clinical context in order to understand the fullness of those implications that amount to more than just a simple trade-off. In another study, said Headen, researchers used SD to create a qualitative understanding of the dynamic feedback of weight management during the childbearing window, looking at weight in the context of clinical, social, and other policy implications, as part of setting the strategic agenda of a research collective (Skouteris et al., 2015). While there is limited systems science work in the area of pregnancy and weight, Headen said that there has been a considerable amount of work on other maternal health outcomes, as well as on obesity trends in nonpregnant populations.

Systems science can be used to advance understanding, take action, and evaluate progress, said Headen. Once a model of a system is in place, one or more elements of the system can be chosen for intervention, and the quantitative approaches can be used to measure where effects are happening. Headen shared examples from her own work to demonstrate how systems science can be used to understand, act, and evaluate. Headen developed a conceptual framework of structural marginalization, dynamic neighborhood systems, and racialized maternal morbidity outcomes (Headen, 2025). The framework sought to explain how neighborhood factors such as public services, built environment factors, and social context are codependent on each other and create an emergent “neighborhood system” that dynamically interacts with a birthing person’s stress, access to health care, socioeconomic status, and exposure to processes of racialization to produce maternal morbidity over time. This model was then used to inform a community-engaged process with Black birthing people and service providers in Philadelphia to understand their experiences navigating neighborhood systems, and how these experiences affected pregnancy outcomes.

Another study (Headen et al., 2025) was not focused on pregnancy specifically but used SD mapping to explore synergy in an equity-focused obesity prevention framework. The core of this work, she explained, was a transition from a linear theory of change framework to a dynamic framework; a model of change was developed to show the factors that affect child obesity in Philadelphia and to unpack which interventions worked by what mechanism. The model mapped out the system on multiple levels—the proximal context in which an intervention happened, the unintended consequences of the intervention, the social and economic resources that shaped the larger context, and upstream contextual community activities that were happening at the same time as the intervention that synergistically supported ultimate success.

Systems science tools can be important bridging tools from understanding to action and action to evaluation, especially for structurally marginalized populations, said Headen. However, there are some key research gaps that need to be addressed. First, there is a need to better operationalize the social and structural drivers of GWG. This can give insight into how to tailor intervention approaches for different populations. Second, systems approaches could be used to develop “flight simulators” for intervention development. That is, models can be developed to test interventions and see the potential effect of making a change in the environment. There is also a need for a better understanding of how to use big data and large linked datasets to improve both qualitative and quantitative systems development.

CLINICAL APPROACHES

In this session of the workshop, a panel of experts discussed clinical approaches for addressing prepregnancy BMI and GWG. Esa M. Davis, University of Maryland School of Medicine, moderated the discussion. The following summary of the panel discussion represents the speakers’ experience and expertise. Davis asked the panelists—who included experts in nutrition, obstetrics, maternal fetal medicine, primary care, family medicine, and bariatric surgery—to paint a picture of what is happening on the front lines of clinical care, and to identify opportunities for innovation in this setting.

Current Environment

Davis asked panelists to describe the challenges associated with treating patients with obesity or underweight, and how the 2009 guidelines fit into their clinical practice. Laura Riley, a maternal-fetal medicine specialist at Weill Cornell Medical College, said that the guidelines have been “drummed into everyone’s education,” and they are “spit out to the patient.” The

patient is told to gain less than 25 pounds in 9 months, and there is little direction or follow-up on how to do that. There are a few people who can afford to get nutrition or exercise counseling, she said, but most people are simply given “a prescription that they actually can’t follow.” Nanette Santoro, a specialist in obstetrics and gynecology at the University of Colorado, agreed with Riley that the guidelines are not always useful or appropriate when caring for a patient. For example, if an older woman with obesity is trying to get pregnant, advising her to lose weight before pregnancy may do more harm than good. Asking her to lose weight and to postpone pregnancy in her narrowing window of fertility may cause the woman stress and worry. Santoro also noted that for women who struggle with infertility, the harm may be even greater; there is some evidence that behavioral weight loss is associated with more miscarriages. On the other side of the equation, Santoro said that there are a fair number of women in Colorado with the “athletic triad,”² and they can be resistant to gaining weight to optimize pregnancy outcomes. The mainstay of treatment in this population is cognitive behavior therapy; she said this is sometimes necessary simply to restore menstrual cycles.

Amoreena Howell, a full-spectrum family medicine physician at the University of Maryland School of Medicine, said that the guidelines have led women to think that they need to gain a minimum amount of weight, regardless of their starting point. Some women get very stressed if they are not gaining the weight they think they should, especially in the 1st trimester. In addition, the recommendations are not individualized, other than by BMI, and there is no differentiation for women with a BMI over 40. From a primary care standpoint, trying to address prepregnancy and interconception weight loss is challenging because Medicaid and private insurance do not cover glucagon-like peptide 1 agonists (GLP-1s) for weight loss at this point.

Judy Simon, a registered dietitian nutritionist specializing in reproductive nutrition at the University of Washington Medical Center, said that as a dietitian, she sees patients on both ends of the BMI spectrum. For all women, the assumption that weight equals health is a problem. Women of all sizes can have issues with metabolic health and may need help focusing on nutrition to have healthier pregnancy outcomes. Simon said that her clinic does not use weight cutoff points but instead looks at each woman individually and provides her with the care she needs; women with higher BMIs can have good outcomes with good care, she said. Rather than encouraging women to take time to lose weight prior to pregnancy—as they lose ovarian reserve—Simon said that they need accommodations that focus on health instead of weight. Instead of “throwing numbers” at patients, providers should put nutrition in the context of being healthy for

___________________

² Low energy availability, menstrual dysfunction, and low bone mineral density.

pregnancy. She added that weight bias can do a lot of harm, and that health care providers need to be trained to help all people live healthier lifestyles.

The current guidelines do not take into consideration whether a person has undergone bariatric surgery, said Lawrence Tabone, a specialist in bariatric surgery at West Virginia University Medicine. However, there is evidence that women with a history of bariatric surgery gain less during pregnancy than women who have not had bariatric surgery. This association holds true when matched for age, weight, and other health conditions, and it is most pronounced for women with a BMI under 25 kg/m². He noted that women with a history of bariatric surgery are also less likely to experience gestational diabetes, preeclampsia, large-for-gestational age, and need for cesarean section, when compared with same-weight peers who did not have bariatric surgery. Bariatric surgery seems to be protective against these negative perinatal outcomes, said Tabone, and the evidence suggests that gaining less than what is recommended by the guidelines is not necessarily detrimental.

Further, there may be other measures that are more important than BMI or GWG, such as lean body mass, macronutrients and micronutrients, or fetal growth. Using the metric of weight gain alone is not giving a good picture of the health of the pregnancy, he said. Women who have had weight loss surgery are often told to wait a year or two before getting pregnant; however, those who get pregnant earlier gain less weight during pregnancy and have better outcomes.

Tabone said there is a need for studies to challenge the idea that pregnancy needs to be delayed after bariatric surgery. As others have noted, the issue of waiting to conceive is particularly challenging for women who are older, and there is a need to balance the risks and benefits of pregnancy timing after bariatric surgery. When women get bariatric surgery, it changes the food environment for other family members, and it changes the intrauterine environment for pregnancies. Tabone suggested that because the intrauterine environment influences future obesity risk through epigenetic mechanisms, one of the greatest benefits of bariatric surgery may be its effect on unborn children by reducing their risk of childhood and adult obesity and related comorbidities yielding an intergenerational effect.

Focus on Individual Health

Davis asked panelists to expand on Simon’s comments about focusing on health rather than weight and treating each person as an individual regardless of weight. Simon began by saying that early in her practice, she developed a 6- to 12-week class called Food for Fertility for women with overweight who were struggling to conceive. The program went beyond recommendations to “eat healthy, lose weight…take a walk.” Together, the women cooked healthy foods, took walks, did yoga, and supported each

other. The program made a difference in women’s fertility; women started ovulating and many became pregnant. Simon noted that these changes happened after only 6 to 12 weeks of the program—women did not have to wait for their BMI to drop below some threshold number of “health.” Women with overweight have been told by the medical system that they are “broken” and that there is “something wrong” with them, said Simon. As a result, they do not go to the doctor as often as they should, and diagnoses like prediabetes or polycystic ovarian syndrome are missed. These conditions can be treated if diagnosed, and women can have healthier pregnancies.

Simon said it is critical that providers make women feel comfortable talking about their health and that providers do not focus all of their patient interactions on weight and weight gain. Providers need to understand the details of a patient’s individual health status and provide care based on what she needs, whether that is help with nutrition, lifestyle, metabolic health, or access to resources. Riley agreed that interventions would ideally be personalized to each individual’s circumstances and needs. However, she noted that providers often have a very brief period to talk with patients; prenatal visits are as short as 7 minutes. There is some opportunity for education because patients come in frequently during pregnancy, but the best time to intervene would be before the patient even gets pregnant. The GWG guidelines can make things easier because the messaging is simple, she said, but the messaging may not be clear enough, and it may not be focused on what it should be.

Santoro added that women can carry guilt for not weighing the “right” amount, or for not eating the “right” foods during pregnancy. She recalled that during a focus group, women who had a prior miscarriage were weeping because they felt responsible for their loss. While there may be some association between weight and miscarriage, the connection is not nearly as strong as the burden of guilt these women felt. Guidelines need to be thoughtfully developed and disseminated, said Santoro, to ensure that they will do good and not result in unintentional harms.

CLOSING REMARKS

To close, Rasmussen shared her own brief summary of the key points made by speakers over the 2 days of the workshop. This workshop was conducted with the goal of informing future gestational weight gain guidelines. These guidelines, she said, will need to incorporate new evidence and new methods, and acknowledge the strengths and weaknesses of various approaches (for key considerations raised during the workshop panel discussions, see Box 4-1). New attention was brought to the needs of underweight women. Speakers focused on women in the highest BMI categories,

as this group is growing more quickly than others, and they have the most severe clinical challenges. In 2009, evidence was lacking to differentiate guidelines for women with BMIs over 40. There is still a need for further research in this area, said Rasmussen. The clinical challenges of providing medical care for pregnant women rise with their BMI values, she added. And some treatments currently in use, such as bariatric surgery and use of GLP-1s require further study in this population. Another research gap, she said, is the availability of data that relate weight changes in pregnant women to the short- and long-term outcomes for these mothers and their children in the same cohort (for research gaps raised during the workshop panel discussions, see Box 4-2). Additionally, she noted that data on the association between GWG and pregnancy outcomes that are stratified by BMI is also needed, as demonstrated in the scoping review (see Appendix C).

Rasmussen acknowledged that this is a challenging topic A future committee to develop guidelines will need to consider providing more specific recommendations for dietary patterns and energy intake, especially in light of the new Dietary Reference Intakes for energy. High-quality data on dietary intake and dietary patterns of pregnant women remain remarkably limited; this is an area for further work. Some speakers suggested that the recommendations for weight gain could be lower, at least for some women. Rasmussen said that given the low adherence rate to current recommendations, making future recommendations for weight gain that are even lower than the current ones would require more work on the part of providers to support women in meeting these goals. The development of percentile-based charts by BMI category for monitoring weight gain in pregnant individuals could assist with the task.

On the second day of the workshop, said Rasmussen, presenters noted that social, economic, and environmental factors all affect women’s health, prepregnancy BMI, and GWG. These factors act across the life course and can be the subject of interventions to improve health outcomes. Several speakers noted that community-engaged research is needed, particularly in low SES communities, but it can be challenging and time-consuming. Speakers distinguished between interventions at the upstream, midstream, and downstream levels, and noted that most interventions for GWG are at the downstream level in the form of clinical practice guidelines. Speakers gave examples of dietary and lifestyle interventions that have been successful and cost-effective, and of successful upstream environmental-level interventions that were not specific to pregnant women. However, gaps remain in the implementation, specificity, and reach of such interventions.

Speakers discussed various available big data sources, including vital statistics, medical claims, and EHRs; each has its strengths and weaknesses. Speakers highlighted that birthing women represent a complex

system embedded in multilevel social systems. Systems thinking is a way to understand and analyze interdependence between factors and to integrate quantitative and qualitative data. This approach is beginning to be used to study GWG and health outcomes, but more research is needed to identify the social and structural drivers of GWG and to use this information to develop interventions.

The panel on clinical care made a number of important points, said Rasmussen. Several panelists suggested that GWG guidelines be considered in the social, environmental, and economic context of people’s lives. A few speakers noted that the guidelines themselves are not enough; there is a need for providers to have detailed information about how to support patients in meeting the guidelines. Based on their experiences, some panelists highlighted that GWG guidelines should not be “spit out” to patients with little guidance and inadequate referrals; patients cannot be expected to gain weight within the guidelines if they are given a “prescription they can’t follow.” Further, many individual speakers noted that patients need individualized treatment that is based on their health status, not simply their weight. Speakers also emphasized the importance of reaching patients before they become pregnant but noted the challenges in doing so. A key message across the workshop, said Rasmussen, was that the guidelines should “do no harm.” To avoid doing harm, she added, it is necessary to understand all of the potential ways that guidelines could harm patients. In summary, this workshop provided wide-ranging insights into what needs to be considered in the development of future guidelines. Rasmussen thanked the speakers; the sponsors; and National Academies of Sciences, Engineering, and Medicine staff and adjourned the workshop.