In this study, discussions with women with a history of GDM through focus groups and informant interviews provide information about barriers and facilitators to healthy lifestyle choices in the year postpartum, and they discuss their preferred design and components of a lifestyle intervention program. Women in this population found it difficult to attend a single focus group, and most stated that they would have difficulty with an intervention based exclusively on in-person group sessions or contacts by phone. This evidence of significant time constraints suggests that implementing a face-to-face lifestyle intervention like the Diabetes Prevention Program  would be difficult in this population. An intervention that does not require face-to-face contact that women can utilize at their convenience may have more success. In our study, the majority of participants expressed interest in an internet-based lifestyle intervention that they could access on their own schedules. We found universal familiarity and high levels of daily internet use in this population, similar to rates found in a recent survey of this age group . This is a novel and important finding, as the cost of internet-based technologies has decreased significantly, making it more feasible and potentially cost-effective to implement internet-based interventions. Given the high risk for type 2 diabetes in the GDM population and the need for effective intervention strategies in the postpartum "window of opportunity," these data provide valuable information for the design of postpartum interventions with the goal of decreasing the incidence of type 2 diabetes.
A larger proportion of the women in our study considered themselves to be at moderate or high risk for type 2 diabetes than has been seen in previous studies [15, 24]. This may be due to the fact that the study took place at an academic medical center and may have included more women exposed to GDM education. Despite this higher level of perceived risk, participants nonetheless reported difficulties adopting healthy lifestyle changes, citing barriers similar to those found in other studies [24–26]. Similarly, Swan et. al. demonstrated that rural Australian women with a history of GDM had a low prevalence of healthy diet and physical activity behaviors despite increased perception of risk and knowledge about prevention strategies for type 2 diabetes . Previous studies identified lack of time and lack of childcare as the most important barriers to healthy lifestyle activities in women with prior GDM , findings which are echoed in our study. In addition, women in our study discussed how financial constraints, difficulties related to work, and feeling guilty about being away from their children all served as barriers to achieving a healthy lifestyle.
A potentially important novel finding was the frustration expressed by participants with feeling judged by clinicians about their lifestyle and choices, which may be important to address in this population. Respondents' reports of feeling judged by their health care providers may signify untoward effects of the "medical model," which assumes that if individuals are provided adequate information and skills and are motivated to change their behaviors, they will naturally make informed choices and modify their unhealthy behaviors . It is possible that more collaborative approaches to behavior change such as patient-centered counseling or motivational interviewing techniques, may be more acceptable in this population,  and that counseling delivered by non-clinicians may feel less judgmental and perhaps could be more successful. These findings should be considered when designing future intervention studies in this population. In our study, we found that women were enthusiastic about the idea of a lifestyle coach, whom one woman stated would seem "more like a partner."
We identified several key concepts that may be useful to integrate into future postpartum interventions for women with a history of GDM. We found that only ten of 38 potential participants who indicated their desire to attend a one-time focus group were able to make arrangements to attend, and nine of these 38 women could not be scheduled for an informant interview. This is consistent with the findings from a recent postpartum weight loss trial for overweight and obese women in which participants were enthusiastic about group meetings and exercise sessions, but on average attended only 3.8 classes out of the eighteen that were specified for the protocol, and 43% did not attend a single session . Given the difficulty with retention seen in previous studies employing face-to-face interventions, as well as the time constraints identified in our study, an intervention that women can utilize at their convenience may have more success. We identified substantial interest in an internet-based intervention, with some women expressing interest in face-to-face interactions to complement the internet-based format.
Both focus group participants and informants discussed financial barriers to the adoption of healthy lifestyle changes, including the high cost of healthy foods and gym memberships. In addition, many participants discussed difficulties convincing their children to eat healthy foods, and the majority of informants wanted their spouses/partners or other members of their family included in an intervention program. Informants discussed the importance of family awareness and support, and also mentioned that lifestyle changes would benefit the entire family. It has become increasingly clear that public health interventions that focus specifically on individual-level determinants (e.g., attitudes, beliefs, and skills) are less successful than those that take into account the social ecological contexts in which the interventions occur . Including an ecological perspective to design a postpartum intervention program would incorporate multiple levels of influence on health, including individual, interpersonal/familial, community, and public health.
This study is limited by its small sample size, and the fact that it was only conducted in English; therefore these women may not be representative of the general population of women with a history of GDM. Since we were asking women to recall barriers and facilitators in the postpartum year, the responses may be affected by recall bias given the varying amount of time since their last GDM pregnancy. In addition, some of the participants may have given answers they thought we wanted to hear, reflecting a social desirability bias . Finally, the quality of the informant interviews may have varied given that many women were multi-tasking while participating in the interview, including caring for their children or driving home from work.