Understanding Gestational Diabetes, Future Diabetes Risk, and Diabetes Prevention: A Qualitative Study of Patient, Provider, and Staff Perspectives

Women with a history of GDM were recruited from the St. Louis, MO, area using flyers at local FQHCs and community organizations, a local radio advertisement (on station WHHL FM 104.1), public transportation (MetroTransit) postings, and notification after delivery at an academic hospital. Women were eligible if they were 18–40 years of age and had had a pregnancy complicated by GDM within the past 10 years (defined by Carpenter and Coustan criteria [17] or chart documentation) and either on Medicaid, Gateway to Better Health (St. Louis safety-net coverage), an Affordable Care Plan marketplace insurance plan, or uninsured during the GDM-affected pregnancy. To ensure representation in the sample, women were stratified by time since GDM-affected pregnancy (<1 or ≥1 year since delivery) and postpartum screening status. Receipt of recommended postpartum screening was defined as a fasting plasma glucose or 2-hour oral glucose tolerance test within 4–12 weeks, postpartum or an A1C from 12 weeks to 1 year postpartum, confirmed by medical record review. HCP and staff participants were stratified by their health care role and recruited from member FQHCs of the St. Louis Integrated Health Network (IHN). The IHN coordinates leadership and transitions of care among area FQHCs and hospital systems.

After obtaining informed consent, two trained research team members conducted semi-structured interviews with 36 women who had had a pregnancy complicated by GDM. Nine clinic staff (nurses and dietitians) who provide patient education were interviewed individually or in pairs. The team conducted four focus groups with 21 prescribing HCPs (physicians, nurse practitioners [NPs], and physician assistants [PAs]). HCP focus groups were completed within specialty (two for obstetrics and gynecology [OB/GYN], one for family practice, and one for internal medicine).

Community stakeholders provided input on interview guides. Interviews explored patient understanding, attitudes, and self-efficacy using multiple questions. These included: “What did it mean for you to have high blood sugar during your last pregnancy?,” “Please describe the advice you remember receiving during your pregnancy,” “Please describe what you know about how high blood sugar during pregnancy affects your health in the future,” and “Imagine you are telling your friends about ways to lower your risk of getting diabetes. What would you say?” Full prompts are provided in Supplementary Appendix S1.

OB/GYN and family practice HCPs and maternal-child clinic staff were asked to imagine that the interviewer was a patient just diagnosed with GDM and provide counseling. Internal medicine providers were asked, “Do you routinely ask your female patients about complications of pregnancy, and, if so, how does this alter your management?” They were also asked, “How important do you feel it is for your patients to be screened for type 2 diabetes after a pregnancy with gestational diabetes” (HCPs) or “What do you view as your role in helping patients get screened for diabetes after pregnancy?” (staff) and “What do you think your patients understand about gestational diabetes and the way it affects their future risk of diabetes?” (HCPs and staff).

Demographic information was collected, and interviews and focus groups were audiotaped and transcribed verbatim. Social cognitive theory underpinned our exploration of the interrelationships between personal factors (patient-level understanding, self-efficacy, and attitudes in each theme), environmental factors (in this context, education provided by staff and HCPs on each theme), and patient behavior around lifestyle changes during and after pregnancy and postpartum screening for diabetes (Figure 1). In particular, we noted times when HCP and staff perceptions of patient understanding and attitudes may affect both patient understanding, attitudes, and self-efficacy and the education provided.

Using content analysis, two independent raters (R.C.W. and C.J.H.) developed one codebook for interviews with patient participants, and two independent raters (R.C.W. and L.N.H.) developed one codebook for HCP and staff focus groups and interviews. Each code had a definition, inclusion and exclusion criteria, and sample quotes. Two independent raters then coded all transcripts independently using the codebooks (R.C.W. and L.N.H. for patient participants and H.R.L. and R.C.W. for HCPs and staff). There was >80% agreement before discussion between the two raters, and discussion resolved remaining discrepant codes. A third rater (C.J.H.) was available but was not required to reconcile codes.

Using NVivo12 qualitative data analysis software (QSR International, Burlington, MA), the researchers compiled and analyzed codes related to understanding among patient participants and HCP and staff provision of education and perceptions of women’s understanding to ascertain similarities and differences. The Washington University Human Research Protection Office approved the study on 6 June 2016. Interviews and focus groups occurred between May 2017 and April 2018.

Table 1 presents population demographics. Doctors and dietitians were the most common sources of information cited by women with GDM. However, many women also received information from their mother or other family members. A minority of women mentioned online sources as a prominent resource. Five main themes, each with several subthemes, were identified. These included 1) general understanding of GDM diagnosis with focus on neonatal complications; 2) variable recall of diet, exercise, and weight recommendations; 3) overwhelming medication and self-monitoring routines; 4) short-term focus of type 2 diabetes risk and screening; and 5) limited understanding of all options for diabetes prevention. These themes are discussed in detail below with associated illustrative quotes. (Quotes with participant numbers ending in “PS” were from providers or staff, as opposed to patients.) Supplementary Appendices S2 and S3 provide additional quotes.

A few women used specific terminology, such as placenta and pancreas, in describing their GDM diagnosis, but a majority described that their blood glucose had been too high. Some were concerned that they had diabetes permanently. The majority of women recalled that the baby could be too large or could have other severe complications.

Additionally, many women recalled diagnosis later in pregnancy with some areas of confusion. Several women thought that the baby could have diabetes at birth, although others correctly remembered the immediate concern for the baby was low blood glucose at delivery.

Regarding diagnosis, some women were skeptical about test timing, test method, and the advice they received.

HCPs and staff generally felt that too much information was overwhelming and chose to focus more on immediate treatment goals and complications.

Many HCPs and staff tailored their initial counseling based on patients’ reactions to the diagnosis. Whereas some patients became very anxious and needed reassurance, others were not concerned and needed to focus on complications. Immediate pregnancy complications were more motivating than long-term risks.

Overall, commonalities arose among HCPs, staff, and patients around the recognition that the disease is complex and that the most motivating factor is the baby’s health. It is difficult to know whether the patients’ focus on neonatal complications derived from the education they were provided or whether the HCP and staff perceptions of the patients’ immediate concerns directed their focus in counseling. Patients’ skepticism also highlighted potential missed opportunities for enhancing trust. For example, women needed more information on why the testing was done with a sugary drink, why it was often done later in pregnancy, and whether there were indications for earlier testing.

Women accurately recalled the main points of counseling on diet, exercise, and weight gain during pregnancy. Many knew they should be eating more vegetables and fewer carbohydrate-heavy foods and limiting portion sizes. However, portion definitions varied, and serving size was mentioned infrequently. Many expressed surprise about less obvious sources of carbohydrates such as fruits, starchy vegetables, and sweetened beverages, especially juices and teas. There was a stronger focus on avoidance of certain foods than on inclusion of healthy foods. One participant thought the restricted diet contributed to low birth weight.

Almost all HCPs and staff focused on the basics when counseling their patients about diet and exercise: eliminating sodas and juices and getting patients to recognize and replace carbohydrate-heavy foods. Most HCPs also perceived that their patients lacked understanding of the basic nutrition concepts needed to manage their GDM.

Patients recalled exercise recommendations with limited detail, with a focus on walking and the positive impact it could have on lowering blood glucose after a meal. Recall of weight recommendations was also general in most cases, and where specific numbers were recalled, they were often higher than the Institute of Medicine recommendations (18).

HCPs and staff recognized that patients did not clearly understand recommendations for weight gain in pregnancy, but they also found it difficult to discuss weight.

In summary, with regard to understanding of diet, exercise, and weight recommendations, the strongest focus among patients, HCPs, and staff was on dietary change, and specifically carbohydrate restriction. Some carbohydrate sources were a surprise to patients (e.g., fruit, juice, sweet tea, and starchy vegetables), and HCPs and staff recognized that education may need to start with very practical information with less focus on label reading. Exercise and weight recommendations were less specific, with patients recalling little about recommended exercise types and time outside of walking after meals. Patients also recalled little about the amount of weight gain recommended, with acknowledgment among HCPs and staff that weight can be difficult to discuss, and societal expectations of appropriate weight gain in pregnancy often exceed the recommendations.

Although many women expressed annoyance at having to change their diet during pregnancy, they were most frustrated with blood glucose monitoring and medication administration. Women were worried that they might hurt their baby, and almost every patient had negative responses to either the frequency of or discomfort associated with checking blood glucose and/or administering insulin.

Despite these concerns, the majority of patients had an accurate recall of the self-monitoring process. Many women also discussed keeping logs of blood glucose (and sometimes food) and discussing this with their HCPs. At least some felt that seeing the connection between blood glucose and food was helpful, but many felt that self-monitoring and charting were cumbersome. Several women reported that they felt anxious or scared when first learning to manage GDM, but many also reported that they grew accustomed to the process and recognized they needed to take this diagnosis seriously for their child.

HCPs and staff heavily emphasized blood glucose monitoring and medication administration, recognizing that time and pain were barriers to adherence.

Almost universally, patients recalled the process of monitoring their blood glucose multiple times per day and taking medication during pregnancy in a negative way, despite recognizing that it was a necessity and displaying resilience in making this work for their baby’s health. Although a small number of patients minimized the impact on daily life, these negative reactions included pain and discomfort with needles, fears of harming the baby, anxiety related to future diabetes, and the inconvenience of incorporating checking and logging into daily routines. HCPs and staff acknowledged the fear and challenge associated with the multiple tasks required with this diagnosis.

Almost all women in the screened group had an accurate understanding of their postpartum type 2 diabetes risk and the need for follow-up, whereas fewer than half of unscreened patients demonstrated this same understanding. Unscreened patients who had delivered <1 year before their interview understood their risk better than participants ≥1 year out from their GDM-affected pregnancy.

In general, most women understood that they were at a higher risk for developing type 2 diabetes and needed screening in the immediate postpartum period.

However, whereas women understood the need for screening, there was less awareness of how screening needed to continue after the first postpartum screen. A few women felt strongly that diabetes was only a risk immediately after pregnancy and, if it was gone, they were okay.

There was less universal understanding of what an increased risk for type 2 diabetes meant for patients. Many already perceived their risk to be high because of family history, so an additional increase in risk associated with GDM was not always clear. Related to this, women were unsure they could change their risk.

The majority of HCPs and staff did not prioritize counseling on long-term type 2 diabetes risk unless patients were concerned. Counseling on risk typically occurred briefly at GDM diagnosis and in the context of scheduling patients for their follow-up appointments, without additional counseling about future risk and testing.

Just as HCPs and staff felt that patients’ concern regarding the GDM diagnosis seemed to be at one extreme or the other, a similar dichotomy existed with regard to future diabetes risk. Some HCPs and staff felt that patients were motivated to avoid diabetes, but many found women unmotivated by future risk.

Similar to the theme of GDM understanding and focus on neonatal complications, there was a tendency among patients, HCPs, and staff to emphasize the short-term risks and immediate postpartum diabetes screening rather than examine long-term risks and the need for continued screening. Again, it cannot be determined whether the short-term focus of HCP and staff counseling influenced patients’ attitudes or whether the HCPs’ and staff’s perceptions of patients’ motivation constrained their patient education. Additionally, HCPs and staff perceived extremes in motivation to be the norm; patients were either highly concerned with or indifferent to risk without much middle ground. Finally, a lack of clear understanding of how lifestyle change could interface with family history to alter risk may have influenced motivation and engagement for some patients.

Patients generally had an incomplete understanding of postpartum prevention of type 2 diabetes after their GDM-affected pregnancy. The majority of patients understood that maintaining the diet and exercise changes made during pregnancy could reduce risk for postpartum type 2 diabetes. Some women recalled broad guidance, whereas others were more specific.

However, women also recognized the challenges in maintaining these changes.

Almost all of the HCPs and staff emphasized the maintenance of healthy diet changes in the postpartum period, both for type 2 diabetes prevention and weight loss, and they talked with their patients about the importance of setting achievable goals.

Although many patients recalled general counseling about breastfeeding for the health of the baby, few women knew that this may reduce their type 2 diabetes risk. Additionally, most women reported no counseling about metformin as an option for prevention and expressed concerns about side effects.

HCPs and staff also reported that most patients did not want to take a pill for prevention, and although the majority of HCPs and staff counseled on breastfeeding, they usually focused on the benefits for the baby or weight loss as opposed to type 2 diabetes prevention.

Similar to recall of diet, exercise, and weight recommendations during a pregnancy complicated by GDM, the focus of diabetes prevention was mostly on diet, with some recognition of staying active and losing weight (without specific details). Regarding dietary changes recommended for diabetes prevention, some patients remembered very specific recommendations, whereas others recalled the importance of a healthy diet more generally. Notably, most patients did not understand the roles of breastfeeding and metformin in diabetes prevention, and there was a strong bias against using a pill for prevention. HCPs and staff noted this desire to avoid medications and, similar to the disease-understanding theme, focused most of their breastfeeding counseling on the immediate benefits for the baby because that was perceived to be the best motivator.

Our data illustrate unique perspectives among urban women without private insurance who have had GDM and HCPs and staff caring for them. This population has been under-represented in many prior qualitative studies. Our approach allowed for an in-depth exploration of women’s understanding about GDM both during and after pregnancy. By recruiting HCPs and staff with a wide range of training, roles, and experience, we were also able to provide a breadth of perspectives from health care personnel caring for women who live in a midsized city and do not have private insurance.

Previous qualitative investigations have primarily focused on barriers to follow-up and included mostly perspectives from patients (10–14,19–25). This study explored women’s understanding of GDM and future risks in detail and compared this understanding with the education they received, as reported by HCPs and staff.

To our knowledge, this is the first qualitative study on this topic in the United States to include the perspectives of nurses and dietitians, as a prior study in a similar population limited professional interviews to physicians and nurse midwives (16). Nurses and dietitians provide an important perspective, as they spend the most time counseling patients about diet, exercise, and blood glucose monitoring. Triangulating the perspectives of women with GDM, HCPs, and clinic staff contextualizes information perception and provision, highlighting opportunities for intervention. It also raises awareness of practice variation, patient education, and patient understanding across the pregnancy-to-postpartum transition.

Our study found many similar themes to those highlighted in prior literature. Women and their care providers focused more on immediate neonatal complications than on long-term maternal risk (16). Regarding the initial diagnosis of GDM and the tasks of monitoring and self-management during pregnancy, women expressed feeling anxious and overwhelmed, and HCPs and staff acknowledged these common sentiments (11,12,16,21). Women remained motivated by the health of the baby during pregnancy, but they often desired to return to old habits afterward and were relieved to learn their initial postpartum type 2 diabetes screening was negative (11,13,16). Most women remembered that they should get an initial postpartum test, but few remembered goals for continued screening (15,20,25). Some women also felt that there was little focus on the postpartum period among HCPs (16,26), and the actual likelihood of developing diabetes was poorly understood and sometimes underestimated (10,11). A recent systematic review of multiple qualitative studies performed throughout the world noted this variation in future risk perception as well (13). In our analysis, women who were screened and those who were within 1 year of delivery had the most accurate understanding of postpartum risk and screening guidelines, highlighting an opportunity for HCPs and staff to continue to reinforce the importance of screening over time for women with a history of GDM.

Unique points raised in our interviews included the limited knowledge that women had about the potential benefits of breastfeeding or metformin for diabetes prevention. This was consistent with messaging from HCPs and staff. Additionally, most women were hesitant to consider preventive medication, preferring to focus on diet and exercise, despite noting that maintaining these lifestyle changes was difficult.

Some women in our study questioned whether lifestyle change would substantially change risk, attributing their likelihood of getting diabetes to genetics. In this way, our findings were similar to a study conducted among Native American women, who also experience high rates of GDM and type 2 diabetes (27). This is concerning in that, although genetics contributes to diabetes risk, there are multiple factors that affect development of type 2 diabetes, including potential environmental and social determinants. A perception of prevention efforts as futile could lower motivation to sustain difficult lifestyle changes, and HCP focus on genetic predisposition may perpetuate false attribution of racial differences to biology.

Regarding diet, some information was recalled accurately (e.g., the need to increase nonstarchy vegetables and limit high-carbohydrate foods), but there was less focus on the details of food label reading and serving size, likely because elimination of sugar-sweetened beverages was a predominant recommendation. Patients recalled exercise counseling and weight management instruction with less detail, and HCPs and staff may have felt less comfortable providing this education.

This analysis highlighted areas of accurate and inaccurate understanding that may improve counseling during and after a GDM diagnosis. Commonly, women expressed limited awareness of the need for continued lifelong diabetes screening, and even if they were aware of the risk, they focused on diet and exercise as their only options for prevention. Although it is understandable to emphasize the immediate risks to a pregnancy, our results underscore the importance of continued emphasis on diabetes screening, concrete and tailored recommendations for diet and exercise postpartum, and increased counseling on the roles of metformin and breastfeeding in diabetes prevention. Given that metformin was found to be as effective as intensive lifestyle intervention for type 2 diabetes prevention among women with a history of GDM (4), it is important to further explore the reticence to use metformin and to counsel about the role metformin can play in prevention.

Pregnancy is a vulnerable time for women, but it also represents a crucial window for diabetes prevention. In our study, women reported, and HCPs and staff perceived, more motivation for lifestyle changes during pregnancy, when maternal health directly affects the baby. Women may receive or retain less information on type 2 diabetes screening and prevention postpartum.

Scheduling challenges necessitated the use of focus groups instead of individual interviews with HCP participants. This strategy allowed us to include the perspectives of more HCPs, but we acknowledge that a group setting may have influenced the responses collected. HCP and staff participants came from just two FQHC systems in the St. Louis area with whom we were able to establish a partnership. Although this recruitment strategy will allow for continued collaboration with these FQHCs, it limited generalizability.

Notably, HCP and staff participants in this study were predominantly White (86.7%), whereas women with a history of GDM were predominantly Black (72.2%). Though our participants are representative of the patient population served by St. Louis-area FQHCs (73.7% Black) (28), we suspect that current and historical racial dynamics in the United States and in St. Louis could have affected our data in multiple ways. First, the interviews did not explicitly explore the role of structural racism in distrust of the health care system that may have influenced patient, HCP, and staff experiences. Second, although our interviews were conducted by both White and Black study personnel, racial discordance between interviewer and interviewee and the fact that the White interviewer was also a physician may have complicated the already difficult task of establishing trust within a 90-minute interview and limited what participants were comfortable sharing.

Despite these limitations, there are multiple important implications for clinical practice. First, it is important for women to understand that diabetes risk continues in the long term after an affected pregnancy and that screening should continue every 1–3 years throughout life. It is crucial for primary care providers to know patients’ pregnancy history and incorporate it in their prevention and screening counseling. Next, counseling, beginning in pregnancy and continuing to the postpartum period, must focus on building agency and self-efficacy in diabetes prevention to help women understand that diabetes is preventable even in the context of family history and GDM history and equip them to sustain lifestyle changes. Additionally, messaging about the diabetes prevention benefits of both metformin and breastfeeding must be clearer, with additional discussion of the barriers patients perceive to using medication for prevention. Finally, given the importance of baby’s health as a motivator for maternal action, interventions both during and after pregnancy should support both maternal and child health, extending beyond the clinical encounter to address social determinants of health that may play a role.

Our findings provide important insights into women’s understanding of GDM and future diabetes risk and prevention. Additionally, we present further information on the education provided by HCPs and clinic staff across disciplines and the way these individuals perceive women’s understanding of these topics. Future interventions that capitalize on motivation during pregnancy may help women manage their GDM and transition to healthy postpartum behaviors for type 2 diabetes prevention.

The authors thank all of our study participants for their time and invaluable insights. They acknowledge Rana Rahaman and Reeti Pal for their role in transcript deidentification and Samantha Williams for her role in early interviews, recruitment, and transcript deidentification. They thank Kendra Copanas and the Making Change Happen Leadership Group at Generate Health for their review and feedback on the interview guide and the Center for Community Health Research and Partnership Community Advisory Board for advice on recruitment materials. Finally, they acknowledge Karen Bradshaw, the Network Community Academic Partnership group, and the Transitions of Care Task Force with the St. Louis Integrated Health Network for providing important input on the development of the study and the interpretation of its results.

This work was funded by the National Center for Advancing Translational Sciences (grant number KL2TR002336) and a Barnes Jewish Hospital Foundation/WU Institute of Clinical and Translational Sciences pilot grant. H.R.L. received funding from the National Heart, Lung, and Blood Institute Short-term Training Grant program (grant number T35HL007815). L.N.H. received funding from a National Institute of Diabetes and Digestive and Kidney Diseases training grant (grant number T32DK007120). C.J.H. is funded through the National Institute for Child Health and Development (grant number K23HD096204). A.S.J. and G.A.C. are supported in part through funding from the Siteman Cancer Center and the Barnes Jewish Hospital Foundation.

No potential conflicts of interest relevant to this article were reported.

H.R.L. was involved in codebook creation, transcript coding and reconciliation, data interpretation, and manuscript preparation. R.C.W. recruited and interviewed participants, created codebooks, coded and reconciled transcripts, and edited the manuscript. L.N.H. coded transcripts and edited the manuscript. B.J.-J. and H.B.M. were involved in study conception, interview guide design, community engagement, and manuscript review. A.S. facilitated community engagement with data interpretation and edited the manuscript. G.A.C. was involved in study conception and design, interpretation of findings, and manuscript editing. A.S.J. mentored C.J.H. in study design and implementation, assisted in interpretation of the results, and edited the manuscript. C.J.H. conceived of and designed the study, conducted interviews and analysis, and edited the manuscript. C.J.H. is the guarantor of this work and, as such, had full access to all of the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.