Stage 1 evidence review, qualitative research and stakeholder consultations
Our systematic review of RCTs evaluating behavioural interventions that aimed to prevent T2DM in women with prior GDM [
21] identified 10 trials which examined the effect of the intervention on various metabolic indicators. The included studies were limited by small sample sizes and substantial heterogeneity in both intervention components and outcome measures. This complicated firm conclusions about the superiority of specific intervention content, duration or modes of delivery. Therefore, based on the included studies, it was not possible to identify one specific intervention type as superior, but meta-analysis of four trials assessing the effect on diabetes incidence showed that interventions in the first 2 years after delivery were superior to no intervention (pooled estimated of risk difference per 100: − 5.02 (− 9.24;-0.80)). Furthermore, there was a tendency for interventions that started during pregnancy or within the first 6 weeks after delivery to have poorer outcomes compared to interventions starting later. This informed our decision to initiate the Face-it intervention approximately 3 months after delivery, which would also allow the baseline data collection for the trial evaluation to align with the timing of the routinely recommended postpartum oral glucose tolerance test.
Our systematic review that explored determinants and barriers for GDM services, including healthy lifestyle after delivery and prevention of future T2DM [
13], identified risk perception, self-efficacy and social support as important determinants for engaging in healthy dietary and physical activity behaviours. Consequently, these constructs became determinants that we sought to promote through the intervention. The review also identified a number of barriers, such as lack of time, motivation and social support, and suggested that women with prior GDM may be facing emotional distress.
The review on determinants and barriers predominantly identified studies from the U.S., Canada and Australia. However, the qualitative studies we performed gave us evidence from the local Danish context in which our intervention would be carried out and evaluated. The first of these explored the experiences of five women with previous GDM within the first 3–4 months after the delivery [
23]. The women in the study described emotional distress as a consequence of the GDM diagnosis, which was similar to the findings of the systematic review. Danish women with prior GDM reported feelings of sadness, guilt and self-blame, and it was apparent that the intervention needed to be sensitive to these feelings and not to assign blame to the mother or induce medicalisation.
The women in our explorative qualitative study also emphasised the importance of social and emotional support in general, and particularly from their partners, to mobilise time and energy to follow a diet and physical activity regime [
23]. This coincided with a postpartum intervention study from the UK and Canada, which showed that not only did paternal weight correlate with maternal and offspring weight, but having a partner involved in the study was associated with successful study completion [
32]. This convinced us that we needed to include the partner in our study, both to address his/her cardio-metabolic risk and as a source of social support for the woman with prior GDM. To further examine how this might manifest in a Danish context, we interviewed five male partners of women with prior GDM [
24]. A key finding from these interviews was that the baby and the family have absolute first priority. Therefore, taking time to, e.g. exercise, was perceived as selfish and associated with feelings of guilt. However, being a good role model for one’s child by being physically active was also highlighted in the interviews. The challenge was thus to create an intervention which promoted healthy behaviours in the context of being a good role model rather than taking away precious time from the family.
Our review also indicated that a lack of knowledge about the risk of T2DM after the diagnosis of GDM and guidelines for health service support were a barrier to sustaining a healthy behaviour after delivery [
13]. Furthermore, women considered postpartum health services to be unsupportive and most women were not aware of postpartum services or did not know how to navigate them:
“I’ve been to my doctor and had my blood sugar tested, but then there are no more [follow-up] after the delivery. I just think it’s easy to fall back into the unhealthy lifestyle again when there isn’t anyone keeping an eye on you anymore [ … ] You are a bit abandoned and left on your own when you’ve delivered” (Woman with prior GDM, quote from Svensson et al [
23]
)
This finding suggested that poor health literacy and challenges in accessing the healthcare system required further exploration. Therefore, we conducted a third qualitative study; this time focusing on healthcare providers and the health system level [
25]. The study showed that health visitors, despite playing a key role in health promotion in families in the first years of the baby’s life, had limited knowledge about GDM and its implications. Often, the health visitors were not even aware whether a woman had been diagnosed with GDM or not. Findings also suggested that general practitioners (GPs) often omitted follow-up and long-term risk measurement after GDM. Moreover, we discovered that women received opposing messages from different healthcare providers, which could lead to women neglecting their long-term risk of diabetes. Thus, it was apparent that increasing health visitor skills and knowledge about GDM was required and that knowledge transfer and collaboration across sectors needed to be established to align knowledge about GDM and create a coherent preventive pathway.
From the scientific symposium [
22] a key recommendation was to apply a multi-determinant approach and structure the intervention on multiple levels. For example, it was agreed that barriers to healthy behaviour exist and should be addressed at the individual, family and health system levels. Further, it would be necessary to take on a broad and positive understanding of health in line with the WHO definition focusing on social, psychological and physical health [
33].
Our consultations and workshops with midwives and health visitors further strengthened health visitors’ potential as the most optimal group of intervention deliverers. In particular, health visitors provide counselling based on the broad WHO concept of health. However, the consultations and workshops also confirmed the qualitative research findings that the health visitors needed additional training. They were not particularly comfortable with addressing risk behaviours and disease prevention in our target group:
“We don't come into parents’ home with a raised finger. And if one can see that there are a lot of soft drinks on the table in a home, then we may address this in a broader way by paying attention to food and meals in general terms” (Health visitor)
At the symposium, experiences from the Australian MAGDA study demonstrated that a tele- or digital component might hold promise as an approach to improve engagement in the intervention among women with prior GDM [
34,
35].
In addition, other studies have identified app-based technology as a possible solution to support people at risk of diabetes [
36] and women with prior GDM in particular [
37,
38]. One argument is the flexibility that such eHealth solutions offer as they can be accessed in people’s own homes and at any time of the day. In this way, eHealth technologies can increase the availability of health promotion to populations that are usually difficult to reach [
39,
40]. We decided to further explore the potential for involving a digital solution and found a few digital platforms incorporating the interpersonal level, e.g. relying on social support and feedback, which was suggested by the women in our workshops. We identified the Liva app as the best e-solution for adaptation and tailoring to the families in the Face-it intervention [
41,
42]. The Liva app is an interactive eHealth lifestyle coaching program (long-term Lifestyle change InterVention and eHealth Application [LIVA 2.0]) [
43]. The app builds on a strong personal relationship between user and a health coach, who supports the user through individualised goal-setting and feedback [
43]. As digital support was suggested by the women themselves in initial interviews, we found that the combination of providing digital support as an addition to home visits aligned with the tailoring of intervention to meet the needs of the target group. The health visitors involved in this stage reported that they could take on the role as health coach as well.
From stage 1, we identified health visitors as intervention deliverers in the family and found a digital, interactive platform as part of the intervention content. As such, we left an exploratory phase and proceeded into co-production to identify practical solutions for intervention content.
Stage 2 co-production of the intervention
The second stage in the systematic development process was based on continuous development and adaptation of the knowledge gained in stage 1, and aimed to design an intervention prototype that was ready for testing [
15]. Together with health visitors, families and hospital-based healthcare professionals involved in GDM care, we co-produced the intervention content and delivery components: Specifically, we presented the findings from stage 1 through workshops as well as the intervention premises to co-develop potential intervention content to accommodate the families’ barriers and motivators for healthy behaviour. Further, we had health visitors suggesting their own available resources for adaptation and initiated role play exercises to tailor identified intervention tools. In particular, we wanted to ensure a smooth and coherent transition from hospital discharge after delivery to the health promotion intervention delivered by municipal health visitors.
The cross-sectoral preventive pathway
To ensure a coherent cross-sectoral preventive pathway for the women, both in the trial and in a possible future implementation, the local stakeholders from the three project hospitals, general practices and senior health visitors were invited to local workshops to discuss possible care pathways. The healthcare professionals in the hospitals were satisfied with systematic information flow across professions in the obstetric department. However, GPs and health visitors felt limited by the lack of information provided to them by the obstetric departments.
We interviewed women and their families about their experiences of GDM-related care among other topics. The women described a need to leave the GDM diagnosis behind due to the strict treatment regimen they experienced in pregnancy. However, the families also recognised the benefit of the health visitor taking on a health promotion role to motivate health behaviours in the family.
“It is very important that it does not become a raised index finger but becomes motivating. So, you think to yourself "that was a good idea". I think it depends a lot on how your relationship with the health visitor is” (Partner to woman with prior GDM)
We returned to the healthcare professionals with new insights from the families and considered the best ways to secure knowledge transfer from obstetric departments to municipal health visitors. They suggested providing a hospital discharge summary to the health visitor delivering the intervention. In order to create a coherent pathway for the families, the health visitors also suggested that they, by the end of the intervention, should encourage the women to book and attend the recommended glucose test and counselling with her GP. The idea was that this would strengthen communication and knowledge transfer to the GP and would increase the likelihood of the women being followed-up regularly by their GP as recommended. Thus, engaging closely with the health visitors and hospital staff allowed the identification of a possible solution for a coherent care pathway that lived up to the requests of all stakeholders.
Home visits and an interactive dialogue tool: ‘the family wheel’
We met with the health visitor management in one of the three municipalities that we planned to involve in a later trial and presented our current principles on how to promote health in families where the mother had GDM, e.g. focusing on the broad health concept of WHO, social support, motivation, self-efficacy, risk perception and health literacy [
26]. This led the health visitors to introduce a health pedagogic tool:
the family wheel. The
family wheel is an interactive dialogue tool developed by health visitors themselves to support socially vulnerable families in the transition to parenthood both during and after pregnancy. A prior evaluation of the family wheel found that health visitors used it to help structure and professionalise their dialogue with families. The family wheel originally contained relevant themes for a postpartum intervention, including social relations, breastfeeding, living situation etc. We saw great potential in modifying this conversation tool to uphold the health visitors’ usual practice and structure their new role as health promoters for the whole family. Earlier interviews with families had taught us that a close relationship between the families and the health visitor was critical to enable an open conversation about health, particularly as this often involved sensitive topics, such as overweight, future diabetes risk, partner support and specific food and physical activity habits. In workshops with health visitors, we discussed how the increased risk after delivery could be presented in a motivating way by using the family wheel. The health visitors were not used to addressing parents’ health behaviour and expressed concerns about unintendedly stigmatising the families. Yet, health visitors suggested that by adopting the family wheel as part of standard practice in the intervention, it legitimised conversations on health risk which led to the first thematic category on the modified family wheel: ‘GDM’. The main topic would be a debriefing of the experiences from the GDM-affected pregnancy and a discussion on the risk of T2DM. When asking the health visitors how to modify the wheel further, they specified the need to touch upon all themes relevant to health:
“When I set it [the family wheel] up, I usually ask them how much they need to talk about that theme. The area in question is pointed out. And I do not follow the manual slavishly. Because it may well be that they have no need to talk about gestational diabetes but have a huge need to talk about childbirth. It may be easier to articulate some topics and to get into some issues if they suggest it themselves” (Health visitor)
It was essential to the health visitors to make the families reflect on their health views and encourage already established health behaviours. We redesigned the wheel through continuous dialogue with the health visitors. The family wheel finally included five topics: 1) GDM, 2) everyday routines, 3) food and meals, 4) physical exercise, and 5) family, friends, and network. As such, health in the family was the main focus and GDM was only one in five themes of the family wheel to be addressed. When we presented the family wheel to the families, they were satisfied with the broad aspect of topics and did not feel that they were defined only by their GDM diagnosis. By making health comprise of multiple and interconnected areas, the families perceived this part of the intervention as welcoming a focus on their daily lives.
The choice to adopt the family wheel in the Face-it intervention helped facilitate a strong collaboration with health visitors. Health visitors expressed ownership across municipalities as they felt acknowledged in their profession by building on similar pedagogical non-directive and non-judgmental methods and gained new knowledge about this high-risk group. Moreover, it strengthened the methodological quality of the intervention by tailoring and qualifying the material to their profession. In this way, the adoption of new themes into the family wheel supported health visitors in taking on a new role as health promotors. They helped the families to navigate health information and services, thus increasing health literacy and facilitating and increasing positive family dynamics and social support around health behaviour change.
As a result of the findings from stage one, we wanted to introduce the Liva app as part of the intervention content to families and health visitors. The Liva app includes health behaviour features; however, it was clear that the content was shaped by other target groups e.g. those with diabetes or overweight who would report on medication use and blood sugar values [
43]. When introducing the app to the health visitors, they were less enthusiastic about the digital solution. The health visitors would usually spend time during home visits encouraging families to reduce their screen time and they felt ambivalent towards promoting an app. As such, the health visitors emphasised the need for the app to promote positive everyday activities:
“It [the app] should follow up on what succeeded for you and not what failed. Because I may have a goal to “run on Wednesday afternoon”, but it did not work out … And I do not think they would benefit from that at all. But look, I went Monday!” (Health visitor)
Through co-production with health visitors, we emphasised the role of the health coach to ensure that the families would set goals based on the families’ own wishes, preferences and circumstances. Thus, we decided that the built-in feature of ‘life goals’ should be highlighted in the digital support as a way to prompt individual and family-based health behaviours. A goal could be to ask a friend to go for a walk, read a book, create a shopping list, plan the snacks for the day, or to encourage your partner to go for a walk etc. In accordance with the families’ wishes for an app, a breastfeeding feature was developed, and to accommodate a broader understanding of exercise, the category of physical activity was expanded to include activities drawn from everyday life in a family i.e. activities such as ‘walking’, ‘vacuuming’, ‘exercises with baby’ or ‘gardening’.
The Liva app also helped counter some other challenges raised by health visitors at this stage. The health visitors were worried about their ability to provide specific advice on GDM, diet or exercise if requested by the families. By making specific health information available in the app, we wanted to assure the health visitors that they were not expected to be experts in all health-related topics. We tailored materials in the app to families of women with former GDM, such as physical activity and dietary recommendations, exercise charts and shopping lists etc. To finalise the content, we wrote manuals for the family wheel and the Liva app and started recruiting health visitors in the three municipalities.
The co-production phase was finalised as the intervention was now premised by health-visitor-led home visits guided by the family wheel and a tailored health app, and an intersectoral knowledge pathway. We ended the co-production phase when the Liva app and family wheel were approved by health visitors.
Stage 3 prototyping, feasibility and pilot testing
In stage 3, the core intervention components in the Face-it intervention were ready for modelling and testing as a whole in the municipalities. At this stage, we involved families, GDM experts and health visitors and health coaches (in two of three settings, this was the health visitor) as intervention deliverers aiming to 1) secure testing and tailoring of content, 2) strengthen ownership, 3) adapt intervention delivery to the local context and 4) ensure proper training and competences.
We tested the acceptability of the family wheel and Liva app with two families and two women with prior GDM. In these interviews, we addressed topics on the family wheel e.g., ‘food and meals’ and ‘exercise’ to test the acceptability by enquiring into how families experienced talking about these subjects and asked whether they felt comfortable with talking to a health visitor about this. The family wheel was assessed to be acceptable while relying on only a few contextual factors. Firstly, its aim to address sensitive subjects in the family depended on a trusting relationship between the family and the health visitor. Secondly, the fact that health visitors would come to the participants’ homes provided more flexibility for the families as they did not have to transport themselves. Thirdly, the families noted a concern regarding the Liva app about the time needed for data registration and the app potentially competing with other digital elements, e.g. watches with step counts. This concern about the app was balanced by the families’ positive attitude towards their ability to easily access a health visitor/health coach and the possibility of receiving tailored health information, e.g. in the form of recipes.
“I would think it would be a good idea that someone is pushing me to do it. But I don’t think my husband would use it at all. I think I would choose something like exercise, weight, or diet in the app. My milk production is not very good so it could be very nice to talk about what could help increase it [through the app]” (Woman with prior GDM)
We then held meetings with each municipality to tailor the structure of the intervention to local resources and preferences. The local municipalities decided themselves how to organise the staff delivering the intervention.
GDM experts (dieticians, nurses, endocrinologists and obstetricians) from the collaborating hospitals were invited to discuss the intervention components and adapt the cross-sectoral pathway with senior health visitors to ensure a coherent preventive pathway at the three intervention sites. The experts raised the issue that women with prior GDM and their partner often varied in their perception of GDM. In contrast, others emphasised the role of inactivity and poor diets and dealt with lack of motivation to change health behaviours. Further, the duration and frequency of the intervention with three home visits within 9 months was deemed appropriate by health visitors and experts as long as the health coaching was available between the visits to provide feedback and advice. Thus, the delivery of the intervention demanded continuous tailoring of communication to meet the needs of the families and ensure the intervention deliverers collaborated with the families to support the achievement of behavioural goals.
Lastly, we conducted four full training days for health visitors/health coaches to educate them in intervention delivery. At these training days, we presented the intervention manual consisting of a conversation guide for each theme on the family wheel. Thereafter, the education was problem-based, e.g. the health visitors pointed towards three challenges after the first day’s training with the family wheel: balancing the conversation of future risk in the family; engaging the partner in the home visit, and getting the families to act on their goals. These three themes became central to the following two training days. One training day was exclusively focused on using the Liva app. Throughout the training days, we pilot-tested the home visits in the intervention by using case descriptions of various families, probing communication strategies and adding suggestions for ‘good questions’ to start a conversation in the intervention manual. The health visitors found that the visual design of the family wheel, including the use of colours (green, yellow, red), helped them approach certain topics, but also helped the families to assess their own wishes for change within those topics. Some flexibility was allowed in terms of which theme to talk about when and in the approach to addressing topics and posing questions.
Stage 4 involvement in developing outcomes for evaluation
In the development of a core outcome set for health promotion in diabetes after pregnancy, the 115 key stakeholders agreed on 19 relevant themes during the final consensus meeting. Core outcomes for the specific intervention depended on the focus of the intervention. These included constructs from behavioural change theory (self-efficacy, motivation, barriers and perceived risk), health behaviour (dietary intake, physical activity, sleep and breastfeeding), cardio-metabolic- and adiposity measures (body mass index, weight, waist circumference, glucose, cholesterol, and blood pressure), offspring outcomes (growth, diabetes), quality of life, knowledge, social support, and program delivery (participation, engagement). The detailed description of the core outcome set has been published elsewhere [
27,
31,
44]. The involvement of different stakeholders in selecting the outcomes allowed for the inclusion of different perspectives on what was considered important to measure. Particularly, including women with GDM in the process meant that more ‘patient-oriented’ outcomes, e.g. social support and quality of life, were retained in the core outcome set [
27].
Based on the core outcome set, the qualitative interviews performed at stage two and the consensus meetings with core stakeholders, the research team made the final decisions on which outcomes to include in the evaluation of the Face-it trial. Data collection covered: biochemical measurements (blood samples), blood pressure, anthropometric measures and a self-administrated questionnaire to assess dimensions of health behaviour, social support, motivation and family dynamics. The questionnaire contained both validated scales and self-constructed questions building on qualitative evidence from the earlier stages of intervention development. The full list of measurements is available in the trial protocol [
26]. The findings from the previous stages informed the need to include various psychosocial outcomes. The finding that emotional distress was often present in the target group after delivery combined with the results from the core outcome set informed the decision to include questions on quality of life. The findings about the importance of motivation, risk perception, family dynamics and partner support for health behaviour in the child’s first year formed the decision to construct a set of questions on health behaviours in the family context. We also needed to investigate stigmatisation in relation to GDM diagnosis. Therefore, we developed and pilot-tested a new scale to investigate internalised stigmatisation related to GDM. After identifying the outcomes and finalising the intervention content and modes of delivery, we estimated sample size as well as recruitment and retention rates and finalised the study protocol. The details are available in the published study protocol [
26].