Introduction

Gestational diabetes mellitus (GDM) is a pregnancy complication characterised by a glucose intolerance that starts during pregnancy1. GDM poses health risks for the mother (e.g. preeclampsia, metabolic syndrome), the offspring (e.g. fetal macrosomia, childhood obesity), and childbirth (e.g. premature birth, C-section delivery)2,3,4,5. These risks occur in the presence of glycaemia values consistently out of the reference range1,4,5,6. The global prevalence of GDM in women aged between 20 and 49 years is 13.4% (16.9 million women), and it tends to increase with the growing prevalence of overweight and obesity7.

The frontline intervention in GDM focuses on dietary and physical activity changes, with pregnant women being promptly referred to a dietitian’s advice after the diagnosis1,5,6. In Portugal, nutritional intervention goals and recommendations align with international standards1,6. Specifically, dietary plans are prescribed considering gestational weight gain and corresponding energy requirements4,6, as well as blood glucose levels and related food intake1,4. Nutritional recommendations emphasize a balanced diet that includes at least five meals per day, alongside portion control. Physical activity advice typically entails a minimum of a 30-minute walk after meals4. Furthermore, the American Diabetes Association (ADA) underscores the critical role of patient involvement in care as a key factor in the success of behavioural interventions8.

Pregnancy is a crucial period for both the mother and child. Appropriate nutritional intake is required to avoid in utero foetus’ metabolic programming for chronic diseases9,10,11, particularly in GDM pregnancies12,13. Nutrition care for pregnant women with GDM has shown promising results regarding maternal glycaemia and birth weight14. Regular appointments with a dietitian during pregnancy have also been linked to a reduced risk of Neonatal Intensive Care Unit admissions15. Alongside dietitians, diabetes educators and endocrinologists are crucial in supporting women in glucose monitoring, physical activity, medical nutrition therapy and pharmacotherapy, and obtaining positive health outcomes16,17. The multidisciplinary approach requires healthcare professionals (HCPs) to maintain consistency in their recommendations to facilitate behaviour change concerning diet18. On the other hand, behavioural interventions in GDM enable pregnant women to be agents of self-care19. However, such interventions require women to rapidly learn and implement self-care practices and change habits, which can be challenging.

Barriers and facilitators to GDM management

Pregnancy represents a particular time in a woman’s life when she is often more motivated to change her habits20. Nonetheless, other factors influence women’s adherence to healthy behaviours during a GDM-affected pregnancy. Social support has been identified as the strongest predictor of adherence to GDM treatment among Scottish women20. However, social events21,22, expectations21,23, and external pressures, such as familial control and stigma21,24,25, negatively influenced women’s adherence. Additionally, women with GDM perceived high socioeconomic and financial status as facilitators of adherence to dietary advice, whereas low status posed a significant barrier24,26,27. Furthermore, nutritional interventions in GDM have been described as culturally insensitive and challenging to follow in late pregnancy, limiting adherence to prescribed diets27,28,29,30.

Difficulties in arranging appointments and limited consultation time also hindered women’s adherence to HCPs’ advice28,29. Women valued HCPs’ attentiveness and ability to inform them but disliked judgmental attitudes, impatience31 and stigma32. Some women preferred to adopt their own strategies, particularly around diet and physical activity, in an effort to gain autonomy, although they also followed the HCP recommendations23. Managing GDM was often seen as time-consuming24,28,33, adding responsibilities and pressure29,33, and complicating family planning33. While self-control was seen as empowering by women who recognised the need to change their health behaviours, many struggled with dietary changes33 due to misconceptions24, cultural beliefs or food preferences24,27,28,29.

Knowledge about GDM is instrumental in achieving behaviour change. Understanding GDM causes and consequences led to faster acceptance of the diagnosis33 and increased risk perception, thereby facilitating women’s adherence to health behavioural changes20,22,23,26,27,28. Conversely, a lack of risk perception about GDM inhibited adherence24,26. Nonetheless, opinions on health education vary: some women found the information provided by HCPs insufficient21,34, while others considered it adequate or excessive28. Moreover, HCPs often struggled to provide adequate support to disadvantaged and migrant women due to communication issues and cultural differences regarding eating habits28. Therefore, digital tools for intervening in health education and behaviour can offer a ubiquitous, accessible, and supportive form of promoting pregnant women’s learning and adoption of self-care practices for managing GDM35.

Mobile apps for GDM management

Most digital-based behavioural interventions targeting GDM use mobile apps (mHealth)35. This field has grown significantly since the first clinical trial in 2015, with a notable increase in randomised controlled trials36. Existing mobile apps’ key features include blood glucose measurements recording and storage, reminders, glycaemia classification, glucose prediction, diet recipes, GDM education, and specialist feedback35. Several apps have been tested in behaviour change interventions in GDM, including Pregnant+37,38,39, Habits-GDM40, Hola Bebé, Adiós Diabetes41, SweetMama42,43,44 and eMOM GDM45,46,47.

Medical guidelines, clinicians’ perspectives and behavioural models have guided the design of mobile apps, often overlooking the integration of self-care theory in their development, despite the importance of self-care (including self-management) in GDM23,48. While prior mHealth interventions positively affected clinical outcomes, e.g. maternal glycaemia and self-efficacy, they have been less successful in driving sustained health behaviour changes or enhancing health literacy40,41,42,43,49,50. We contend that integrating self-care models into the design of mHealth interventions will make them more sensitive to women’s practices, better supporting long-term health behaviour changes. However, the development of effective self-care interventions requires a patient-centred approach that prioritizes women’s needs and experiences in the first place19.

Tailoring materials to individual preferences, culture, health behaviours, and prior knowledge has the potential to improve intervention adherence and efficacy36,51. Some apps have been translated into multiple languages to meet population needs38,41. Still, they have yet to fully considered other cultural factors, e.g. daily consumption of culture/religious-specific foods, which also influence adherence to healthy behaviours. Moreover, HCPs suggest tailoring educational materials to health literacy, structuring educational content rigorously, and having materials screened by HCPs to enhance education and promote adequate behaviour changes52. Nevertheless, current solutions do not provide tailored content to individual characteristics, which may also affect women’s adherence to mHealth interventions and their effectiveness.

Pregnant women should be involved from early development stages to ensure their self-care and tailoring needs are considered, enhancing their engagement with the app and improve intervention effectiveness. Also, HCPs’ endorsement is crucial for developing user-centred designs that facilitate responsive, practical education, and clinical integrated solutions.

Development and refinement of Toolkit to support self-care in GDM

Existing mHealth solutions for managing GDM have sought to support women in making health behaviour changes. Still, they often overlooked women’s needs and characteristics and failed to frame self-care theories into their design. To address these gaps, this study aimed to explore patients’ (women with GDM) and HCPs’ experiences and perspectives in managing GDM, considering key theories of self-care53,54 and clinical nutrition care55, which frame relevant healthcare practices to consider in the development of self-care and nutrition interventions. Furthermore, it also aimed to develop and refine a user-centred, evidence-based digital Toolkit (the NUTRIA Toolkit) for supporting pregnant women with GDM (i.e. end-users) in self-care. The NUTRIA Toolkit aims to provide evidence-based educational content and behavioural tools by integrating the most recent scientific guidelines with practitioners’ expertise and patients’ values56, developed through a user-centred approach in line with users’ needs and preferences57. Additionally, this study intended to identify pregnant women’s characteristics, which may lead to the need of differential interventions, and tailor the Toolkit to their distinct characteristics to promote healthy and long-lasting behaviour changes while using a mobile app.

Methods

Study design

The development and refinement of the NUTRIA Toolkit followed a systematic mixed methods approach58. Data derived from observations, interviews, and co-creation sessions was analysed using qualitative methods, while survey data was analysed using both qualitative and quantitative methods. These analyses enabled the characterisation of women’s and HCPs’ experiences and perspectives in managing GDM, which informed the development and refinement of the user-centred, evidence-based Toolkit. To this end, data was triangulated through three research cycles (Fig. 1). Each cycle resulted in an iteration of the Toolkit requirements and/or content, with adjustments made after the second and third cycles. This approach is characteristic of design research59 and public health intervention research frameworks60. Figure 1 shows a detailed flowchart of the process, depicting how data collection steps and data analysis informed the development of the NUTRIA Toolkit.

Fig. 1
figure 1

Systematic mixed methods approach for Toolkit development and refinement. The nutrition appointment observations informed the definition of topics to explore during interviews. The qualitative analysis of the observations and interviews with women identified determinants of health behavioural changes and GDM-related self-care practices (informed by self-care frameworks). Determinants were grouped into five challenges women face during a GDM pregnancy and were associated with self-care practices, prompting the definition of Toolkit Modules requirements v1. Qualitative analysis of the observations and interviews with dietitians allowed the characterisation of the clinical nutrition care process in GDM (informed by the Model and Process for Nutrition and Dietetic Practice), particularly defining nutrition counselling variations, depending on clinical parameters and contextual factors, and education topics. This information, together with scientific literature and guidelines, supported the development of Toolkit Modules content v1. Note that BCTs were also selected from the BCTs Taxonomy, considering the BCW model, to integrate the Toolkit, which is composed of educational and behavioural content. These activities ceased the first cycle of design. Toolkit Modules content structuring, format, and tailoring were explored in two co-creation sessions with two dietitians, resulting in Toolkit Modules requirements and content refinement, including BCTs reselection, and completing the second cycle of refinement (v2). Endlessly, a survey presenting Toolkit content (v2) descriptions was applied to a group of HCPs to assess their agreement with the content using Likert rating scales, culminating in mostly high-agreement content selection to compose the Toolkit Modules content v3. GDM: gestational diabetes mellitus; v1/2/3: version 1/2/3; HCP: healthcare professional; BCT: behaviour change techniques; BCW: behaviour change wheel.

Full size image

Data collection

The study involved HCPs and pregnant women with GDM, with the latter being the NUTRIA Toolkit/mobile app end-users. Recruitment criteria for women included a past or current diagnosis of GDM, while HCPs were selected based on their current or previous involvement in the care of pregnant women with GDM. For each data collection stage, participants were recruited through purposive and snowball sampling61.

Step 1: Observations of nutrition appointments

MLN, BF and FN performed observations of GDM nutrition appointments in two Portuguese hospitals, one in the North (H1) and one in the South (H2) of Portugal. A total of 66 nutrition appointments were observed (Nobs H1 = 29; Nobs H2 = 37). Observations were conducted on the designated weekday for GDM appointments and depended on the availability of the dietitians. On observation days, the researchers were able to observe all the appointments of the day, i.e. eight to twelve a day. There were four dietitians involved, two from each clinical site. Meaning saturation was considered, and the stop criteria were defined based on it62. The researchers conducted direct observations of dietitian appointments, where they took notes and refrained from interfering in the appointment (passive participation63). At the beginning of each appointment, the dietitian introduced the research study goal and methods to the patient; after discussing the details, the dietitian requested the patient’s consent for the researcher to observe the session and take notes.

During these observations, researchers documented the major themes discussed, relevant information exchanged during the nutrition counselling, and the dietitian-patient interactions. MLN also performed observations of a nurse’s teachings on blood glucose monitoring and management in H2. Personal identification data of patients and HCPs, such as name, citizenship or clinical process identification numbers, were not recorded. Expanded notes were made shortly after each observation to support the analysis.

The notes from the nutrition appointment and nurse teachings observations were reviewed, and the researchers identified several topics for exploration in subsequent interviews with women and dietitians. Thus, these observations facilitated the researchers’ familiarisation with the subject and informed the development of the interview scripts.

Step 2: Interviews with women and dietitians

Interviews with women with GDM were held at the hospital (H1) following their nutrition appointments by FN and BF, and interviews with past patients were conducted online. Interviews with past patients were led by two students under the supervision of BF and FN. Interviews with dietitians were performed either in-person or online by BF and MLN. Some interviews were conducted online for participants’ convenience. Meaning saturation also determined the number of interviews conducted62. In total, we conducted 17 interviews with women and 11 interviews with dietitians. The interviews were semi-structured, following a general guide (available in the Supplementary Materials), but allowing for the exploration of additional issues raised by participants64. Before each interview, participants provided written informed consent.

During the interviews, women with past or ongoing GDM were questioned about their self-care routines and daily life with GDM, the healthcare they received, and their learning processes during pregnancy. The interviews with the dietitians focused on their professional experiences with patients with GDM, the organisation of the healthcare units where they have been working or worked while following pregnant women with GDM, the structure of the initial and follow-up nutrition appointments, the learning processes of pregnant women, and variations in the nutritional counselling based on pregnant women’s characteristics. The researchers also collected the interviewees’ demographic data to characterise the participants. All the interviews were audio recorded and manually transcribed verbatim by the researchers to support in-depth qualitative analysis.

Further analysis of observations notes and interview transcripts raised uncertainties about the structuring, format and personalisation of the Toolkit content, which prompted the planning of co-creation sessions to discuss them with the dietitians and refine the Toolkit content.

Step 3: Co-creation sessions with dietitians

MLN, BF and FN conducted two co-creation sessions with two dietitians from H1, who were currently following women with GDM and were not previously interviewed, at hospital sites. These sessions were conducted to explore Toolkit preliminary developments based on Steps 1 and 2. The first session focused on structuring and formatting the Toolkit module content, while the second session addressed its personalisation to different characteristics of pregnant women with GDM. Several exercises (available in the Supplementary Materials) specifically designed for these sessions were used to prompt discussion and collect the dietitians’ impressions about the materials while reflecting on their clinical practice and experience. These exercises enabled us to explore topics raised by dietitians and women with GDM during the interviews, as we asked the two dietitians to identify associations between topics to help inform the organisation and adaptation of content in the Toolkit, enhancing information understanding. The sessions were audio recorded, and the resulting materials were photographed for qualitative analysis.

Data from Steps 1, 2, and 3 allowed us to develop and refine the Toolkit content, but they lacked a quantitative analysis and the involvement of a broader range of HCPs. Therefore, through an online survey, we involved dietitians, physicians, and nurses in the following step of the Toolkit content validation.

Step 4: Toolkit content validation survey with HCPs

An online survey was administered to 17 HCPs who were currently following or have previously followed pregnant women with GDM, to validate the Toolkit content using soSci Survey [Leiner, D. J. (2024). SoSci Survey (Version 3.5.01) [Computer software]. Available at https://www.soscisurvey.de]. Nine participants were independent from the previous steps. The survey was developed by MLN and tested by the remaining researchers. It aimed to assess the experts’ agreement with several descriptions of the Toolkit content using a 5-point Likert scale from 0 - completely disagree - to 4 - completely agree - and took approximately forty-five minutes to complete. The survey was divided into four sections: (i) a description of the survey’s purpose and the HCPs’ tasks; (ii) collection of demographic data; (iii) an overview of the NUTRIA Toolkit, with a general definition and introduction; and (iv) items related to Toolkit content descriptions per Toolkit module (Nmodule 1 = 4; Nmodule 2 = 5; Nmodule 3 = 34; Nmodule 4 = 6, totalling 49 items) for HCPs’ to rate according to their agreement. The HCPs could also leave comments or suggestions about each content description for analysis.

Data analysis

Theory

In this study, theoretical frameworks of self-care53,54 were used to identify self-care practices and related concepts from collected data. We considered the complementary perspectives of self-care, i.e. the individual and the health system perspectives54. GDM reflects a condition where the care is equally shared between self-care and professional care, requiring the involvement of both patient and HCP in self-testing/monitoring, self-management and self-awareness. Self-care practices can be divided into self-maintenance, self-monitoring, self-management practices and symptom management, which is shared with HCPs. In contrast, self-management support and disease management are exclusively professional care practices53.

The Model and Process for Nutrition and Dietetic Practice from the British Dietitian Association55,65 supported the identification of dietitian’s clinical practices from data. According to the Model and Process, these practices are divided into six stages: (1) assessment, (2) nutrition and dietetic diagnosis, (3) strategy, (4) implementation, (5) monitor and review, and (6) evaluation. The stages in which the Toolkit could facilitate the Nutrition and Dietetic Practice in GDM were identified.

The Behaviour Change Wheel (BCW) model66 and Behaviour Change Techniques (BCTs) Taxonomy67 were used to select appropriate BCTs to integrate into the Toolkit development. The BCW model considers behaviour as a part of an interacting system, involving understanding the behaviour, identifying interventions functions and selecting corresponding BCTs using BCTs Taxonomy. The Taxonomy includes 93 consensually agreed, distinct BCTs that provide a structured method for specifying behavioural interventions67. These BCTs are distributed in the BCW model throughout intervention functions66.

Qualitative analysis

Data from Steps 1, 2 and 3 were analysed using thematic analysis, a method for identifying, analysing, and reporting patterns (i.e., themes) within data68,69.

At first, data from Steps 1 and 2 were analysed to characterise women’s and dietitians’ perspectives and experiences in GDM management, i.e. self-care and clinical nutrition care, respectively. The initial coding phases were performed using the Scrivener writing software. At this stage, MLN selected and named chunks of data related to healthcare practices, which FN concurrently reviewed in detail. The analysis was deductive, i.e. informed by key theories of self-care53,54 and the Model and Process for Nutrition and Dietetic Practice55,65, as well as it went beyond the semantic content of the data, i.e. researchers identified underlying ideas, assumptions, conceptualisations or ideologies that shaped and informed the gathered data68,69. In this phase, MLN proceeded to memo-writing to elaborate themes, specify their properties and define relationships between them70. As the analysis evolved and MLN generated themes from data, affinity mapping71 discussion sessions were conducted with all authors to iterate upon themes and double-check interpretations with the initial codes. The main themes identified include the determinants of adherence to health behaviour changes, self-care practices, and nutritional care process stages in GDM. This first analysis ceased the first research cycle after reaching meaning saturation62 and culminated with the initial Toolkit requirements and content (Fig. 1).

Step 3 proceeded the first research cycle, being informed by the analysis resulting from Steps 1 and 2. This enabled the exploration of topics relevant to refining the Toolkit in the second research cycle. The researcher analysed the materials from the co-creation sessions (Step 3) and performed a complementary thematic analysis68,69 by using the audio recordings of both co-creation sessions. The analysis was followed by affinity mapping71 discussion sessions involving all authors to further refine the Toolkit.

MLN also conducted a qualitative analysis of HCPs’ suggestions and comments from the survey. The suggestions and comments were directed to Toolkit descriptions and explored in terms of how they might influence the selection of the final Toolkit content by triangulating data with the results from the survey’s quantitative analysis.

Quantitative analysis

Data was analysed by MLN using Python 3.10. Descriptive statistical analysis of demographics data collected during Steps 2 and 4 was performed in terms of absolute frequency for categorical variables data and mean and standard deviation (SD) for continuous variables data. Regarding survey responses, descriptive statistics, including median, mode, maximum and minimum, and median absolute deviation (MAD) of ratings were determined.

Based on the survey’s responses, we could also select highly agreed content. The criteria for classifying an item with high-agreement was no HCP rated it as “completely disagree” (i.e., 0 points) or no more than one HCP rated it as “partially disagree” (i.e., 1 point). Additionally, items classified with medium-agreement had one HCP rating them as “completely disagree”, two HCPs rating them as “partially disagree”, or one HCP rating them as “completely disagree”, and another as “partially disagree”. Items not matching any of these criteria were classified as low-agreement.

Results

Participants characteristics

In nutrition appointment observations at H1, most consultations were conducted in Portuguese, with patients primarily from Portugal and Brazil, alongside a smaller group of non-Portuguese speakers, predominantly Chinese. At H2, more than half of the appointments were conducted in English, often requiring a translator, as many women did not speak Portuguese or English. Women were usually from India, Nepal and Bangladesh, while the remaining were Portuguese speakers from Portugal, Brazil or African countries, e.g., Mozambique, Angola.

Table 1 presents the characteristics of the interviewed women and dietitians, while Table 2 depicts the characteristics of the surveyed HCPs.

Table 1 Characteristics of the interviewed women and dietitians.
Full size table
Table 2 Characteristics of the surveyed HCPs.
Full size table

Women’s perspectives and experiences

The researchers selected 115 codes from Steps 1 and 2 (observations and interviews with women) and divided them into two major topics: determinants of adherence to health behaviour changes and GDM-related self-care practices.

Determinants of adherence to health behaviour changes in GDM perceived by women

After identifying the determinants of adherence to health behaviour changes, they were grouped into several key challenges: (1) Understanding the impact of health behaviour changes on the pregnant woman’s body and foetus - this challenge included codes related to the woman’s understanding of the physiological impact (e.g., health condition, symptom) of health behaviour changes. It focused on the direct physical effects of health behaviour changes, excluding the impact of the changes on logistics or social aspects; (2) Measuring blood glucose values - this challenge comprised codes related to the techniques of measuring blood glucose, reference ranges, and the importance of timing and context in monitoring; (3) Identifying dietary features to manage blood glucose levels - this challenge included features of meal timing, focusing on food fractioning and fasting periods throughout the day, and meal content, related to the quantity and quality of food in each meal; (4) Perceiving the multifactorial way to manage blood glucose levels – this challenge included codes mentioning factors other than diet that are relevant for managing blood glucose levels, such as stress, physical activity, and medication; and (5) Feeling supported by others in managing GDM – this challenge included codes expressing the importance of HCPs’ counselling and assistance from other individuals in managing GDM. These challenges encapsulate the critical areas that seem to influence women’s adherence to health behaviour changes necessary for managing GDM effectively.

Self-care practices in GDM experienced by women

By applying self-care frameworks53,54, the researchers were able to distinguish codes related to (1) self-awareness, (2) self-monitoring and (3) self-management practices: (1) Self-awareness included codes related to self-education and self-regulation actions, such as consulting a HCP, searching for information on the internet, analysing blood glucose levels (objectively measured), and assessing how the body responds to diet (subjectively perceived); (2) Self-monitoring comprised codes related to actively measure blood glucose values attributable to GDM and its therapy, such as regular self-monitoring of blood glucose levels; (3) Self-management included codes about taking actions autonomously or as recommended by HCPs to manage blood glucose levels, such as changing eating and/or physical activity habits, and taking medication as needed to manage GDM effectively. These self-care practices translate the multifaceted approach women with GDM take to manage their condition effectively.

Dietitians’ perspectives and experiences

From Steps 1 and 2 (observations and interviews with the dietitians), 319 codes were selected, along with 15 codes from co-creation session 2, which were grouped into relevant topics. The materials resulting from co-creation session 2 were also analysed.

By using the Model and Process for Nutrition and Dietetic Practice55,65, the stages in which the Toolkit could facilitate the Nutrition and Dietetic Practice in GDM were identified: assessment (stage 1), implementation (stage 4) and monitor and review (stage 5).

Assessment Assessment topics included assessing dietary intake, clinical parameters, and contextual factors. During interviews, we asked the dietitians to identify women’s characteristics that were used to tailor the nutritional intervention prescribed. The characteristics identified were as follows: clinical parameters, such as the mother and foetal anthropometrics, symptoms and glycaemia values; and contextual factors, including socioeconomic status, culture and religion, eating patterns and habits, food literacy and motivation. Additionally, based on outputs from the co-creation session 2, we narrowed the assessed variables by excluding foetal anthropometrics and pregestational body mass index (mother anthropometrics), retaining only gestational weight gain (mother anthropometrics).

Implementation Implementation included codes related to counselling, elaborating dietary plans or making dietary recommendations, promoting behaviour change, and building women’s knowledge or patient education. Promoting behaviour change was the least mentioned topic, while developing a dietary plan and education were the most common topics. As we intended to develop educational tools, we identified several key dietetic and nutritional education topics discussed in GDM nutrition appointment: (i) information on what constitutes a healthy diet during pregnancy and in GDM; (ii) information on adequate water intake; (iii) providing food alternatives (food equivalents or substitutions); (iv) instructions on homemade measures; (v) instructions for preparing food; (vi) tips for reading and interpreting food labels; and (vii) demystifying information about diet during pregnancy and GDM.

Monitor and review Monitor and review included codes related to reassessing interventions, indicating that adaptations were made based on changes in clinical parameters and contextual factors during pregnancy and making associations between clinical and dietary variables or between glycaemia values and other variables. We also identified variation of clinical parameters throughout pregnancy by trimester from co-creation session 2.

The Toolkit

The challenges and self-care practices identified guided the development of Toolkit requirements v1 (Fig. 1), which was consistent with our aim of supporting women with GDM in their self-care and health behaviour changes. To establish these requirements, we first mapped the actions women took to address each challenge (Fig. 2). Toolkit requirements v1 were organised into four main modules comprising 6, 7, 18 and 4 requirements, respectively.

Considering the available scientific evidence and key guidelines1,4,5,6, we incorporated the results from the observations and interviews with dietitians (Steps 1 and 2) to develop the Toolkit content v1 aligned with the Toolkit requirements v1 (Fig. 1). In addition, we associated the Toolkit content with BCTs using the BCTs Taxonomy67. For the Toolkit, we selected BCTs related to the intervention functions of education, enablement, training and modelling.

Co-creation sessions 1 and 2 (Step 3) allowed the refinement of the Toolkit (Fig. 1). The outputs from co-creation session 2 complemented dietitians’ perspectives and experiences, as previously described, allowing us to consider clinical parameters and contextual factors to tailor the content of the Toolkit. From co-creation session 1, 33 codes were selected and organised into the Toolkit modules, and the resulting materials were analysed. The outputs of both co-creation sessions, particularly of co-creation session 1, were instrumental in refining the Toolkit requirements and its content structuring and format, resulting in Toolkit requirements v2 and Toolkit content v2 (Fig. 1). This second version of the Toolkit had 1, 1, 8 and 2 additional requirements in each module, respectively.

Fig. 2
figure 2

Mapping of actions into challenges from interviews with women with previous or ongoing GDM. GDM: gestational diabetes mellitus; BGL: blood glucose levels; BGV: blood glucose value.

Full size image

Examples of data instances derived from Steps 1,2 and 3 and their connection to Toolkit requirements are detailed in Table 3, illustrating findings triangulation over the process (Fig. 1).

Table 3 Examples of data instances from different data sources (Steps 1, 2 and 3) and the resulting Toolkit requirement.
Full size table

Content

The NUTRIA Toolkit content is described and categorised based on its type (static/interactive) and its association (or not) with BCTs in Table S1 (Supplementary Material). There were 46 descriptions for Toolkit content v2, organised into four modules: (1) Gestational Diabetes (Nmodule 1 = 3); (2) Glycaemia Self-measurement (Nmodule 2 = 3); (3) Nutrition in Gestational Diabetes (Nmodule 3 = 32); (4) Beyond Diet and Nutrition (Nmodule 4 = 6 descriptions). Most of the content was assigned to module 3, which focuses on nutrition and dietetics in GDM. Screenshots of the NUTRIA Toolkit content on the mobile app are displayed in Fig. 3.

Most content was associated with BCTs, being designated as tools aiming at building women’s knowledge and promoting behaviour change. According to the BCTs Taxonomy and the BCW model66,67, the Toolkit includes tools to support: (i) Education, by informing women about antecedents [Ed/4.2], health consequences [Ed/5.1] and salience consequences [Ed/5.2]; (ii) Enablement, by goal setting (behaviour) [En/1.1], problem-solving [En/1.2], goal setting (outcome) [En/1.3], behaviour substitution [En/8.2] and comparative imaging of future outcomes [En/9.3]; (iii) Training by instructing on how to perform a behaviour [T/4.1]; and Modelling by demonstrating of a behaviour [M/6.1].

The content that is not related to BCTs provides context to the remaining content. Examples include 1.1 “Explaining woman what GDM is” (static) or 3.2 “Using the Portuguese Mediterranean Wheel with information about food group-specific adequate intake using an interactive health nutrition model” (interactive).

Validation

The Toolkit content v2 was validated by 17 HCPs, who rated their agreement with a total of 49 items in a survey (Step 4). Most of the seventeen respondents completed all 49 items in the survey, with the exception of one respondent who did not classify the final six items, which pertained to the content descriptions of Module 4.

Analysing the descriptive statistics, the maximum values for all Toolkit content descriptions were 5 points, with median and mode values predominantly also at 5 points, excluding median and mode values from three items. However, these demonstrated median and mode values of 4 or 3 points, not reflecting a disagreement with the content. MAD values were mostly 0, indicating low variability in respondents’ ratings, except for three items with a MAD of 1. Nonetheless, one item had a MAD of 1, which can indicate disagreement when subtracted from its median of 3 points. Additionally, some disagreement ratings were observed, as reflected in minimum values of 1 and 2 points.

We identified 15 out of 49 items of the NUTRIA Toolkit content v2 classified as low- and medium-agreement between the HCPs, and 34 out of 49 items with high-agreement. Descriptions with low-agreement were disregarded during the Toolkit validation stage, while descriptions with medium-agreement were assessed by complementing the quantitative analysis with comments/suggestions provided by the HCPs. For example, a physician highlighted the relevance of oral nutritional supplementation in pregnancy and that women should be informed about it, even if they had low socioeconomic status: “(About the option of making available only food alternatives to micronutrient supplements for women with low literacy and no information about micronutrients’ health impact) I think it’s crucial to clarify that all pregnant women, regardless of their level of literacy, should take folic acid, iron and iodine supplements. I also think it’s important to demystify multivitamin supplements [physician]”. This comment, together with well-established guidelines, led to our decision to provide information about relevant micronutrients to women with low literacy levels.

The selected BCTs were mainly centred around the education and enablement intervention functions, followed by the training intervention function. Furthermore, 38 static and interactive artefacts were developed to meet descriptions of NUTRIA Toolkit content v3, such as the examples in Fig. 3.

Fig. 3
figure 3

Examples of screenshots from the NUTRIA Toolkit on the NUTRIA mobile app prototype: (a) Menu screen that allows accessing the Toolkit modules and displays the user’s progress in exploring the Toolkit content; (b) Module 1 screen related to consequences of uncontrolled GDM, showcasing information in one of the dropdown boxes about the consequences for the mother in short- and long-term (behavioural content); (c) Module 2 screens on descriptions of equipment required for blood glucose measurement with footnote information about equipment’s name and function (educational content), a tutorial in video format with audio and text format about measurement technique (behavioural content), information about why and when to measure glycaemia (educational content), and demonstration of real-world scenarios (behavioural content); (d) Module 3 screens associated with Submodule 1 on Healthy diet displaying the interactive model of the Portuguese Mediterranean Wheel (educational content), and Submodule 3 on Eating with gestational diabetes providing food combinations to manage BGL with illustrations of glycaemia results depending on food combinations (behavioural content); (e) Module 4 screen related to physical activity recommendations during pregnancy (behavioural content). (b–e) Examples derived from requirements described in Table 3. GDM: gestational diabetes mellitus; BGL: blood glucose level.

Full size image

Discussion

This study reports on a systematic, mixed methods approach to develop and refine a user-centred, evidence-based digital Toolkit for supporting self-care in GDM. Behavioural interventions in GDM enable pregnant women to be agents of self-care19 and adequate nutritional care reflects positive health outcomes for both the mother and foetus14,15. Yet, it requires a multidisciplinary approach, where consistency between HCPs concerning behavioural change advice is of foremost importance, namely concerning dietary advice16,17,18. The developed Toolkit intends to serve as a valuable intermediary in education and behaviour change support in this regard, complementing the counselling provided by HCPs and promoting self-care.

Firstly, the main challenges experienced by pregnant women with GDM were defined based on groups of determinants identified from observations and interviews with women, which complemented the existing evidence. The analysis of observations and interviews with women was guided by theoretical frameworks of self-care53,54 and allowed the identification of self-care practices in the domains of self-awareness, self-monitoring and self-management. The Toolkit requirements v1 were established based on the main challenges faced by pregnant women with GDM and their self-care practices, mapped as actions into the challenges. Then, Toolkit content v1 was developed to meet the requirements raised by patients, but it was also based on current nutrition and dietetic practice, incorporating dietitians’ insights, scientific literature, and key guidelines. Secondly, the Toolkit content v1 structuring, format and personalisation to women’s characteristics were explored in co-creation sessions, resulting in the refinement of the Toolkit modules requirements v2 and content v2. Thirdly, we assessed HCPs’ agreement with the Toolkit content descriptions for each module, resulting in its final version (v3). The NUTRIA Toolkit content v3 consisted of 4 main modules with 38 static and interactive artefacts, including tools to support women’s education, enablement, and training towards GDM management.

The challenges to health behaviour changes felt and the self-care practices experienced by pregnant women with GDM directly informed the four main modules of the Toolkit and its interaction with base functionalities in the mobile app. Concerning the first challenge, “understanding the impact of health behaviour changes on pregnant woman’s body and foetus”, it aligns with previous research indicating that women with GDM who understand the causes and consequences of GDM show faster acceptance of their diagnosis and a heightened perception of risk, facilitating their adherence to health behaviour changes. This first challenge was mainly related to self-awareness activities, such as using the internet to search for information or consulting a HCP and self-monitoring, namely measuring BGL. The requirements associated with this challenge focused on the need to inform pregnant women about the consequences of GDM, explain what a healthy gestational weight gain is and the necessary health behaviour changes for a lower-risk pregnancy with GDM, enable meal recording and reporting of hunger and out-of-range blood glucose values, and support gestational weight gain monitoring. The resulting module, “gestational diabetes”, introduces women to GDM and explains that its consequences are primarily due to glycaemia values being out of the reference range. Additionally, the module emphasises the importance of adequate food intake and gestational weight gain for foetal health.

The second challenge, “measuring BGV”, seems to be the key determinant for successful GDM management and for women to gain autonomy through self-management strategies, e.g., diet and physical activity changes. Autonomy and self-control were also highly valued by women in previous research23,33. However, prior work23 and our findings have shown that some women feel ashamed and embarrassed about measuring blood glucose in public. This challenge was associated with self-monitoring and self-awareness activities, such as consulting a HCP and analysing BGL. The requirements covered the necessity of notifying and supporting women in blood glucose measurement and registration in graphics, explaining the importance and technique of measuring through a video tutorial, providing reference ranges in table format, enabling diary entries and mealtime adjustments, and supporting the recording of medication timing. The corresponding module, “glycaemia self-measurement”, provides a step-by-step video tutorial on measuring blood glucose and includes information on factors that can influence measurement values, such as hand washing and health behaviours, as well as reference values.

The third challenge, “identifying dietary features to manage BGL”, emerged as a relevant determinant of adherence to dietary changes, consistent with previous studies24,27,28,29. In existing literature, some women highlighted the need for training and health education to better self-manage GDM21,34. Women also declare that nutritional interventions can be culturally insensitive and challenging to follow in late pregnancy27,28,29,30. Self-care practices of self-awareness and self-management in diet and nutrition were assigned to this challenge. These challenge-related requirements included notifying mealtimes, supporting meal tracking, communicating with the dietitian, offering food planning, selection and preparation tools – e.g., food safety information, suggesting safe and healthy packed meals, offering food choice substitutions and label interpretation, listing culinary techniques, explain food equivalents and fractioning, demystify diet-related myths – with content in various formats including graphics, video and figures. The module “nutrition in gestational diabetes” is the most informative in the Toolkit, focusing on dietary recommendations divided into three submodules: “healthy diet”, “nutrition during pregnancy”, and “eating with gestational diabetes”. It begins by emphasising the importance of a healthy diet tailored to pregnancy needs, which forms the foundation of an adequate diet plan for GDM.

The fourth challenge, “perceiving the multifactorial way to manage BGL” – encompassing diet, physical activity, stress, and metabolism – underscores the complex interplay between these factors for effectively managing GDM, as previously mentioned in the literature23,33. This challenge included activities of self-awareness and self-management, such as physical activity and medication management. The requirements focused on the need to enable various personalised notifications, explain multifactorial blood glucose management using graphics, offer low-intensity physical exercises through images or videos, provide meditation sessions to reduce stress in audio format, share tips on sleep hygiene, and highlight the management of diet and physical activity. The resulting module, “beyond diet and nutrition”, covers multiple factors affecting blood glucose management, illustrated through graphics that show glycaemia variations due to different factors. The module also provides brief information on physical activity, stress, sleep, and metabolic changes during pregnancy, along with low-impact exercise sessions with figures, audio-guided meditation sessions, and sleep hygiene recommendations.

The last challenge, “feeling supported by others in managing GDM”, also arose as a pivotal determinant in our analysis. This finding is strongly supported by existing literature emphasising the crucial role of social20,21,22,23,24,25 and HCPs’ support in GDM management31. Moreover, different migrant backgrounds pose challenges in communicating with HCPs and women during nutrition appointments28. Self-care practices of self-awareness by exchanging experiences with other patients with diabetes were identified from our analysis. This challenge resulted in requirements for more interactive functionalities rather than static content, including the need to communicate with women and send them educational materials, supporting their GDM management journey. The Toolkit interacts with key functionalities – diet plan, blood glucose registry, and diary – by sending notifications based on self-monitoring variables, pregnancy stage, and educational progression. The NUTRIA Toolkit customises educational content based on socioeconomic status, culture, religion, eating patterns, habits, and literacy, while motivational content is tailored based on the woman’s readiness to change.

Strengths and limitations

This Toolkit is one of the first evidence-based approaches for supporting self-care in GDM that was systematically developed using a mixed methods approach, integrating insights from multiple stakeholders and scientific literature, together with key guidelines. It was guided by key theoretical self-care frameworks53,54 and the Model and Process for Nutrition and Dietetic Practice55, which informed deductive qualitative analysis to characterise self-care and clinical nutrition care in GDM in a systematic way, with the potential to contribute to improved behavioural health outcomes. Its development was also guided by the BCTs Taxonomy67 to promote health behaviour change, similar to other digital behavioural interventions72. Similarly to other commercial apps, e.g., Pregnant+, this approach integrated HCPs’ perspectives and experiences to develop the Toolkit content. Nevertheless, it also prioritised the challenges and self-care needs perceived by women with GDM, considering the necessity for patient-centred health interventions19. Additionally, we were able to tailor the content based on HCPs’ experience to women’s clinical parameters and contextual factors, addressing the need for personalised solutions36,52. Involving pregnant women with GDM from the outset and providing tailored content has the potential to increase women’s engagement with the Toolkit, promoting their adherence to future Toolkit-based interventions. Furthermore, the Toolkit can aid in overcoming insufficient assistance by HCPs, by providing digital educational materials, and cultural barriers by customising the content to the diverse cultural backgrounds of patients28,29,31. Thus, the approach taken robustly supported the development of this user-centred, evidence-based Toolkit for promoting self-care in GDM. It is expected that a Toolkit-based intervention will effectively impact health behaviours and improve health literacy together with clinical outcomes, such as gestational weight gain and maternal glycaemia control in women with GDM.

The development and refinement of this user-centred, evidence-based Toolkit was not without limitations. The oral consent process for observing appointments was implemented to minimize disruptions to the consultations. However, its verbal nature may have limited pregnant women’s reflections about their decision to participate in the study. Still, the study’s main limitations are related to the purposive and snowball sampling recruitment process, which may have resulted in a selection bias. First, the characteristics of interviewed women may have been narrower, limiting the potential applicability of the Toolkit to other pregnant women with GDM with different lifestyles (urban/rural) and diverse migrant backgrounds27,28,29,30. So, extending the study to various cultural and socioeconomic contexts in the future could result in a more inclusive Toolkit. Second, the number of co-creation sessions with dietitians was limited to one clinical site. Thus, including more dietitians from different clinical sites could provide more information about specificities in clinical nutrition care. Third, given the project’s primary focus on nutritional care, the input from dietitians was deemed sufficient for the development of the Toolkit. However, interviews or co-creation sessions with HCPs other than dietitians may provide different rationales and insights and should be considered in future studies. Fourth, we did not perform a co-creation session with patients, as the Toolkit requirements were already derived from patient interviews. While this approach provided valuable insights, the lack of a co-creation session with patients may have limited further iterations to the Toolkit content. Future research should aim to conduct larger co-creation sessions that include diverse patient populations, particularly those from varied cultural and socioeconomic backgrounds, as they may have different needs and preferences. Nevertheless, feasibility testing with a broader group of women is planned to comprehensively evaluate the content, ensuring that a future Toolkit-based intervention is relevant and effective in real-world settings. Lastly, the reduced number of surveyed respondents may have limited the applicability of the results to other contexts, hindering also the use of more in-depth statistical analysis methods. Collecting responses from more participants, from different regions, healthcare settings, and even different countries will help evaluate the applicability of the insights to other settings. Nonetheless, the mixed methods approach strengthens the study findings as multiple data sources enrich the analyses and interpretation of the findings.

Conclusion

A systematic, mixed methods approach to developing digital solutions based on evidence, theoretical insights, and user needs may significantly contribute to advancing behaviour change science and improving solutions to support self-care in GDM. Furthermore, testing the feasibility and acceptability of the NUTRIA Toolkit with pregnant women with GDM and conducting a randomised controlled trial to evaluate the effectiveness of a Toolkit-based intervention can provide a unique opportunity to examine the effect of the Toolkit features and their interaction in GDM self-care.