Participants were recruited from a sample of MSP health facilities. Initially, we selected the only type C health centre, and a random selection of six type A/B centers from the 27 available. We aimed to recruit 160 participants from the type C and 40 per A/B center. A randomized list of additional health facilities was generated to serve as replacements until the required sample size was achieved. Facilities were contacted through official channels, and physicians provided patient lists. Initial telephone recruitment was hindered by contact errors, prompting a shift to a mixed approach, combining face-to-face recruitment and scheduled meetings. Consecutive sampling was then applied until the required sample size was achieved.

Of 939 possible participants who were treated in 12 MSP health centers, we finally included 332 participants, from whom we were able to extract clinical information (Figure 1). Of these, 263 (79%) had at least one record of glucose measurement.

The research team, composed of two women and two men, recruited participants between April 2022 and March 2023. The survey collected information about sociodemographic characteristics, access to health services, clinical data, behavioral risk factors, perceived SS and the Diabetes Health Profile-18 (DHP-18). Following informed consent, clinical information regarding treatment, complications, healthcare visits, health metrics (weight, BMI, blood pressure), and glucose status was retrieved from participants’ digital medical records. The follow-up period covered the 12 months following each participant’s survey date. Clinical data extraction was conducted between July and September 2024.

The main outcome variable, controlled glycaemia, was extracted from clinical records prioritizing measures hierarchically based on clinical evidence, using HbA1c, then fasting blood glucose, and finally random blood glucose. When multiple measurements were available, the measurement closest to the one-year follow-up first was prioritized. GC was defined according to ADA guidelines in controlled when HbA1c ≤ 7%, fasting capillary blood glucose between 80 and 130 mg/dL or random blood glucose <180 mg/dL, and in poorly controlled for any other case. As such, individuals who had no glycemic data in their clinical files were included the analysis, classified as “uncontrolled” together with those with poor glycemic results. Sensitive analyses indicated that this categorization did not affect the overall findings (Supplementary Table A). This approach allowed retention of all participants while acknowledging that missing data may introduce some misclassification bias.

We assessed participants’ perceived SS using the multidimensional scale of perceived social support (MSPSS) instrument (26). For this study, we employed the version adapted by Arechabala and Miranda (27), originally developed for Chilean older adults with hypertension and later applied to Chilean patients with T2DM (28). This adaptation uses a 4-point Likert-type scale ranging from 1 (almost never) to 4 (always or almost always), which provides an overall score ranging from 12 to 48, as well as scores for three subscales: family, friends, and significant other, each ranging from 4 to 16. Higher scores indicate a greater perception of SS. The instrument was reviewed to ensure comprehension and cultural relevance in the study context.

Household income was divided into two categories based on the basic salary in Ecuador at the moment of the questionnaire $0–$375 and over $375. The duration of T2DM was calculated as the difference from the self-reported diagnosis date to the interview date and was treated as a continuous variable. We considered oral antidiabetics and insulin treatment for T2DM, and we define polymedication as the use of three or more medications for different chronic indications.

We used descriptive statistics to summarize the sociodemographic and clinical characteristics of the study population and compared these characteristics in patients with controlled and uncontrolled glycemic levels. The chi-square test assessed differences in proportions for categorical variables, while the Mann–Whitney U test evaluated median differences for continuous variables. Lastly, we conducted a logistic regression analysis to examine the relationship between the level of SS (overall score and its three subscales) with GC. We calculated both crude and adjusted odds ratios (ORs) along with their corresponding 95% confidence intervals (CIs) for each subscale. We adjusted the ORs for sex, age, and duration of diabetes to account for potential confounding effects. All analyses were conducted using Stata version 15.0 (StataCorp, College Station, TX, USA).

We conducted in-person interviews with 16 individuals living with T2DM from the district 17D06, Quito in May and June 2023. We used purposive sampling to ensure heterogeneity in terms of gender, age, education level, employment status, ethnic group and socioeconomic status. Sample adequacy was guided by richness and diversity of accounts rather than data saturation (29). We reached out to potential participants by phone explaining the study’s objectives and inviting them to participate on a mutually convenient day and time. Interviews were conducted either at the participant’s residence or within the van used by our research team, based on the participant’s preference. We developed a semi-structured interview script to gather information about participants’ personal experiences managing T2DM, including life changes since diagnosis and resources needed for control (full script in 10.5281/zenodo.15719473). It also explored sources of SS such as family, friends, community, and healthcare providers, and the types of support received (informational, practical, emotional). MP-G conducted the interviews, and MH registered non-verbal information and took notes. The duration of the interviews averaged 36 min, ranging from 14 min to 1 h 10 min. All interviews were audiotaped and transcribed verbatim. Tone of voice, silences and interruptions were registered, contributing to a comprehensive dataset for analysis.

To ensure the trustworthiness of the qualitative component, we followed Guba and Lincoln’s criteria (30, 31). Credibility was strengthened through triangulation between researchers, engagement in the study setting, and consultation with local experts; in addition, member checking was carried out with the community as part of a complementary participatory research. Transferability was supported by detailed descriptions of the study context and participant characteristics. Dependability was ensured through supervision by two researchers not involved in data collection, and confirmability was reinforced by keeping reflexive notes, applying thematic analysis with triangulation during coding and writing, and including participants with diverse characteristics. The semi-structured interview guide was developed based on the study objectives, reviewed by local public health researchers to ensure content validity, and evaluated by a native speaker for linguistic and cultural appropriateness. The guide was iteratively adapted during the first interviews to improve clarity and comprehensiveness.

The verbatim transcripts were analyzed using the framework analysis (32) in ATLAS.ti 8. MP-G used a line-by-line coding approach to identify relevant pieces of information from the data and conducted inductive coding to capture participants’ perspectives on SS and diabetes management. The initial codes described needs, sources and types of support, as well as attitudes and sociodemographic factors that influenced the previous aspects. After coding the first four transcripts, MP-G refined these codes and grouped them into broader categories, including Types of Support, Sources of Support, Diabetes Management and Attitudes. At this point, AB and MP-G triangulated the data and agreed on the codes and categories used. The refined framework was applied to all transcripts, indexing the data into the agreed-upon categories. Using ATLAS.ti 8, we summarized and presented the data into a framework matrix, structuring the results around sources of support (e.g., family, friends, significant others) as the central theme and explored relationships between SS, diabetes management, and sociodemographic characteristics to develop explanations beyond quantitative findings. We conducted the analysis of the interviews in Spanish, and subsequently, the selected excerpts were translated into English. The original Spanish versions of the excerpts are available in Supplementary material 2.

Emotional support from family emerged as a cornerstone of coping with diabetes, particularly during moments of emotional distress following diagnosis.

Participants identified family support as a crucial motivator for treatment adherence and healthier behaviors, reinforcing self-discipline.

Partners provided both emotional and instrumental support, this dynamic was particularly evident in couples where both partners had diabetes, as they supported each other and found it easier to adapt to behavioral changes together. However, economic constraints sometimes limited their ability to fully manage the condition.

Two participants in a migration situation, both originally from Venezuela, highlighted the critical role of financial support from their families who had also migrated. One of them, living alone, expressed a desire to return to her family in Venezuela. However, this financial assistance was used as leverage to discourage her from going back to Venezuela, citing the lack of public healthcare resources necessary to manage her diabetes.

Many participants highlighted the adaptation of household dietary practices as crucial for diabetes management, with family acceptance of the changes providing practical and emotional support. This collective effort facilitated treatment adherence, improved household eating habits, and reinforced shared responsibility. Notably, these narratives were predominantly from women, reflecting their role as primary caregivers.

The lack of this understanding was a significant factor contributing to strained family relationships, as one woman explained:

Financial insecurity added to these challenges, often making it harder to follow the dietary recommendations needed for effective diabetes management.

Beyond dietary support, participants frequently highlighted the practical support provided by family members. This included help with insulin injections, companionship during physical exercise such as walking, and financial support for basic needs, including food, living expenses, and purchasing medications during shortcuts in the public system.

However, a few participants expressed feelings of being a burden, reflecting the emotional strain of relying on others for financial or practical support in managing their diabetes. This perception often led to self-silencing and reluctance to seek help, further complicating the challenges of living with a chronic condition.

In contrast, others found strength and a sense of purpose in feeling useful within their support networks. For one woman, this sense of mutual care was particularly meaningful after her divorce:

The role of family in informational support was less mentioned, but having a family member with diabetes provided valuable advice and also helped participants accept their diagnosis. Health professionals, however, remained the primary source of informational support.

Aligned with the findings from the MSPSS scores, interviewees generally mentioned weak friendship networks, particularly among men.

Participants described the challenges of managing social pressure during social gatherings, where expectations often conflicted with their dietary restrictions. For example, difficulty of refusing certain foods or drinks without offending others, as well as frustration with friends who failed to respect their condition.

Male participants often associated their friendships with alcohol consumption, which became incompatible with diabetes-related restrictions. As a result, these friendships were lost, with men being ignored by their friends because they no longer drink, or the person with diabetes avoiding their friends because of social pressure to keep drinking.

The link between friendships and alcohol consumption was noted across all education levels, underscoring its cultural significance in the country. However, some participants described exceptions, highlighting friendships based on physical activities, such as sports or taking walks.

In a context of insecurity in streets and parks, particularly due to robbery, the lack of having someone to accompany them restricted their opportunities for outdoor physical activity.

All participants taking part in a health-based support group emphasized its role in creating a sense of community (companionship) and offering practical tools for managing diabetes. These groups provide spaces for sharing experiences, having fun and relaxing, engaging in physical activities, and receiving informational support from health professionals and advice from their peers.

Time constraints, often driven by caregiving responsibilities for family members, particularly grandchildren, emerged as a significant barrier to participation in health-based support groups.

Participants also identified community-based groups and social organisations, including sports clubs, urban gardening initiatives, traditional dance associations like ‘Los Negros de la Magdalena’, and religious groups such as ‘Cucuruchos.’¹ These networks fostered a sense of belonging and facilitated close community ties, often serving as sources of social engagement, commitment, and recreation.

Work-related relationships, mentioned mostly by men, often dissolved after leaving or losing a job. However, one man explains how his friendships were sustained through labor unions, which provided a more lasting source of connection beyond the workplace.

Less frequently, neighbors appeared as part of the extended network of support for some participants, providing companionship, sharing meals, and walking together. For example, one participant described her neighbor as a sister, highlighting the role of informal community ties in providing emotional and instrumental help.

Interactions with healthcare professionals varied widely. Some participants reported positive experiences, showing the importance of informational support from doctors in improving their self-management. However, others highlighted challenges, including frequent changes in assigned doctors and perceived lack of interest from healthcare professionals.

Overall, these narratives show the multifaceted nature of SS in diabetes management, showing how various formal and informal sources intersect to provide emotional, instrumental, and informational aid. Such support is vital for individuals managing diabetes and other chronic diseases, significantly influencing their ability to cope and maintain their wellbeing:

Based on the participants’ accounts, we developed a conceptual framework to illustrate the main sources of social support in diabetes management and the multiple pathways through which this support may take form (Figure 3). While family remain a central pillar, other significant sources include neighbors, community groups, and patient clubs. In this low-income context, practical support emerged as particularly relevant: assisting with medication adherence, reinforcing health-promoting behaviors (social control), and facilitating their adoption. The figure also reflects counterproductive aspects of social support seen in the interviews, such as reluctance to seek help due to fear of being a burden, or social pressure during events. Overall, the figure summarizes how different types and sources of support interact along the pathway toward improved diabetes management and, ultimately, better GC.