Gestational diabetes mellitus (GDM) is glucose intolerance that initiates or is first diagnosed during pregnancy, excluding overt diabetes, and affects from 2% to 10% of all pregnant women annually in the United States (1,2). Approximately one third of women with GDM will be identified as having diabetes or glucose intolerance at their 6-week postpartum visit (3–5). Women with diagnosed GDM are at as much as an 84% increased risk for GDM during subsequent pregnancies (6), and the risk for developing type 2 diabetes mellitus during their lifetime is approximately 7-fold greater than that for women with normoglycemic pregnancies (7). Prior analyses suggest that cumulative incidence of type 2 diabetes ranges from 20% to 60% within 10 years of GDM diagnosis, plateauing thereafter (8,9). Without appropriate screening, type 2 diabetes often remains undiagnosed as a result of insidious and asymptomatic clinical progression until onset of secondary complications, including hypertension, heart disease, stroke, retinopathy, and renal failure. Furthermore, uncontrolled type 2 diabetes complicates subsequent pregnancies and is associated with birth defects and birth-related illnesses in the child (10).

The American Diabetes Association (ADA) recommends that women with prior GDM be screened for type 2 diabetes at 6 to 12 weeks postpartum and at least every 3 years thereafter by testing fasting plasma glucose, oral glucose tolerance, or hemoglobin A1c (2). The American College of Obstetricians and Gynecologists (ACOG) recommends postpartum screening for women who had GDM, and refers to ADA recommendation for subsequent screening (11).

In the United States, prevalence of postpartum screening is reported to be suboptimal for women with prior GDM (12–14). Women receive preventive medical services from various practitioners, and most prior work documenting screening has focused on care provided by obstetrician/gynecologists (OB/GYNs) and midwives. In Ohio, 70% of OB/GYNs and 50% of midwives reported almost always screening for type 2 diabetes mellitus during the postpartum visit (15,16). In 2011, Blatt et al analyzed data from a laboratory database and found that only 19% of 23,299 women with GDM-affected pregnancies were tested for diabetes within 6 months after delivery (17). Follow-up screening may also be suboptimal; one study reported that during the 5 years after the postpartum period, 41% of women with prior GDM had not been tested for diabetes by an OB/GYN, family practitioner, or internal medicine physician (18). Few studies have focused exclusively on general health practitioners; therefore, we surveyed Ohio primary health care providers to understand their practices and attitudes related to ascertaining GDM history and screening for subsequent type 2 diabetes as part of a public health effort to improve type 2 diabetes screening rates for women with a GDM history.

We developed a questionnaire with 25 questions related to clinical specialization; patient demographics; and knowledge, attitudes, and practices regarding GDM, screening for GDM history, and subsequent screening for type 2 diabetes; other findings from this survey were published previously (15,16,19). After pilot-testing the questionnaire for clarity with 4 family physicians, we used data from the Ohio eLicense Center to randomly sample 700 of 2,253 physicians licensed in Ohio with “family medicine” or “family practice” self-reported as specialties and 700 of 4,726 with “internal medicine” or “general practice” reported (20). We mailed paper questionnaires up to 2 times during 2010, preceded by a postcard announcement and interposed by a postcard reminder, with an option to respond online. Nonrespondents were sent a final request by e-mail. We subsequently deemed respondents ineligible if they were not treating female patients, worked primarily in a nursing home or long-term care facility, did not routinely deliver primary care, or if they were hospitalists, retired, or primarily practicing outside Ohio, as indicated by notes on returned survey reminders, direct communication from respondents, or survey responses. Respondents were also deemed ineligible if survey responses indicated that questions regarding screening for glucose intolerance do not apply to their practice.

The identification of women with prior GDM, and subsequent lifelong screening for glucose intolerance, is a critical step in preventing type 2 diabetes or identifying the disease early. Screening is particularly important because many women with prior GDM lack additional risk factors (such as obesity) and would otherwise not be screened on the basis of age alone. This study sought to characterize the practices and attitudes of primary care physicians regarding long-term screening for type 2 diabetes among women with prior GDM to aid in the development of public health interventions to increase screening rates by primary care physicians.

Rates for screening were suboptimal, with approximately half (57%) of respondents indicating that all new female patients in their practices are screened for a history of GDM, and fewer than two-thirds (62%) reporting that, every 3 years, they screen all women aged 45 years or younger with prior GDM for glucose intolerance. Similarly, Stuebe et al found that 44% of surveyed primary care providers reported asking women of reproductive age about GDM history during at least half of office visits (21). These low rates reflect substantial missed opportunities, not just for early identification of type 2 diabetes, but also for ensuring healthy future pregnancies, because women with GDM during their first pregnancy have a nearly tenfold increased risk for GDM during their second pregnancy (22). Recurrent GDM increases the risk for adverse newborn outcomes beyond those observed during an initial GDM-complicated pregnancy (23). Early ascertainment of GDM history is therefore an important first step in providing preventive care for women between pregnancies (interconception care).

One limitation of this study is that practices and attitudes related to screening for GDM history and type 2 diabetes might be overestimated if health care providers most interested in the topic responded more frequently to the survey than did providers for whom GDM and type 2 diabetes are of less interest. Our response rate was only 34%; however, surveys of primary care physicians are often completed at a rate of 40% or lower (12,21). Additionally, some nonrespondents might have been ineligible for the survey, and if so, the response rate among eligible survey recipients would have been higher. Furthermore, respondents might have selected responses they perceived to be more socially desirable. Thus, the true screening rates may be lower than reported in this survey, and these data may not be generalizable beyond clinicians practicing in Ohio.

This study assessed the care provided to women with prior GDM. Lifestyle modifications aimed at weight loss and physical activity are effective for decreasing progression to type 2 diabetes (24), and both the ADA and ACOG recommend diet, exercise, and weight management counseling for women with high risk of developing type 2 diabetes (2,11). Accordingly, most respondents reported providing counseling to improve nutrition (79%) or increase physical activity (85%) when they knew a woman had prior GDM. We also found an association between screening and physician care, with those reporting screening women for prior GDM more frequently reporting a role in helping women improve nutrition, increase physical activity, and better understand their risk for type 2 diabetes and the need for lifelong screening. Although a causal relationship between these attitudes and practices cannot be inferred from this cross-sectional survey, these findings are consistent with the Theory of Planned Behavior, which describes behavioral practices as partially dependent on a person’s attitudes (25). Attitudes that could contribute to screening practices include providers’ perceptions of patient care responsibilities. Primary care physicians and internal medicine physicians indicated that they have a role in caring for women with prior GDM (19). However, in the United States, nonpregnant women younger than 50 years of age see OB/GYNs for most preventive visits (26), and some primary care providers may rely on such specialists for ascertainment of GDM history, particularly if the clinical relevance of GDM is perceived to primarily concern pregnancy.

Transition to primary care providers after pregnancy introduces potential for discontinuity of care for women with prior GDM, particularly since a decreasing proportion of family physicians in the United States provide obstetric and maternity care (27). Health information does not always transfer between practices or health care systems. One study of providers within a single health care system found that among 772 women with prior GDM, 46% had the condition documented in the network’s tool used to manage patient care, and only 22% of primary care respondents’ intake forms included assessment of GDM history (21). The same study reported that 55% of surveyed providers cited communication gaps between primary care providers and obstetrics and gynecology care providers as a barrier to providing post-GDM follow-up care, which further underscores the importance of proactive ascertainment of GDM history (21).

Collectively, these results indicate a series of missed opportunities for preventing the progression to or early identification of type 2 diabetes. Opportunities for improving women’s health and birth outcomes include periodic blood glucose screening and equipping patients to make lifestyle changes, but these opportunities begin with early identification of GDM history. Opportunities to improve screening exist at multiple levels, including individual primary care providers (asking patients about GDM), practices/clinics (including GDM-related questions on intake forms and implementing policies to ensure screening for GDM history and type 2 diabetes), health care networks (improving documentation of GDM history; health information exchanges), and insurers (improving access to lifestyle interventions and counseling). These results also highlight a need for public health agencies to identify and address barriers that hinder comprehensive follow-up for women with prior GDM. Providers identified a need for increased resources for improving care of women with prior GDM, including local nutrition specialists, local resources for physical activity, and patient education materials (19). Additional challenges that hinder weight-reduction interventions include insurers’ unwillingness to pay for obesity mitigation programs, physicians’ limited time for patient education and monitoring, physicians’ skepticism or uncertainty about the effectiveness of diet and exercise counseling, and patients’ failure to comply (28). Public health agencies can reduce progression of GDM to type 2 diabetes by supporting interventions that address these barriers, thereby empowering physicians to reduce the long-term impact of GDM. Understanding factors that contribute to low screening rates is critical for public health interventions aimed at increasing lifelong type 2 diabetes screening during primary care visits.

Public health leaders can improve patient outcomes by fostering better integration of primary care, obstetric, maternal and child health, and chronic disease systems. In Ohio, a collaborative was formed at the outset of this study to improve lifelong health outcomes related to GDM. This group is using these findings to create patient education and other resources for women and health care providers to better understand and provide care for GDM and to improve care through a quality improvement approach. For example, a standard letter is being piloted for obstetric providers to send to primary care providers after a woman is given a diagnosis of GDM. The consequences of GDM after pregnancy are well-documented, and the findings of our study reinforce the need for public health and primary health care to work together to improve identification and screening of women with prior GDM.