Study population, setting, and recruitment
This cross-sectional survey collected self-reported questionnaire-based data, from December 2018 to February 2019, in people with T2D in Thai Binh Province, Vietnam, who had been diagnosed with T2D after the age of 40 years. Thai Binh Province is a rural province in the northern part of Vietnam and covers an area of 1542 km2, with a population of 1,860,447 people (2019). The province is divided into seven districts and one major city. This survey is part of the larger interdisciplinary project VALID, ‘Living Together with Chronic Disease: Informal Support for Diabetes Management in Vietnam’ [24].
In Vietnam, the health care system is divided into four levels: national, provincial, district, and communal level. For diabetes, primary treatment is provided by district hospitals; those with diabetes-related complications receive care at province and national hospitals. From the health authorities in Thai Binh Province, we obtained information about people with T2D (by district and commune of residence), who had received diabetes treatment at a district hospital. Two districts, Quynh Phu District in the northern part of Thai Binh and and Vu Thu District in the southern part, were randomly selected. In each district, we selected four communes: Two communes with the highest number of people with T2D, and a neighboring commune of each of the two selected communes, which was convenient for data collection.
We recruited a total of 963 people, who resided in the eight selected communes and who were identified as having been diagnosed with or treated for T2D at a district hospitals. Among them, 37 (3.8%) refused to participate, and 78 (8.1%) did not stay at the address reported to the hospital or had moved away by the time of data collection. A total of 848 people participated in the cross-sectional survey. Among these, 42 individuals were excluded from the data analysis because they were diagnosed with T2D before the age of 40 or did not remember when they were diagnosed. A total of 806 participants were eligible for the analyses ‘Fig. 1’.
Data collection
To collect data, we trained 16 village health workers from the 8 selected communes at a 2-day workshop followed by 1-day field-based training. To be invited to the study and to schedule the interviews, participants with T2D were contacted by mobile phone or via a personal visit. At the beginning of the interview, the outline of the project was explained, and persons agreeing to participate were asked to sign a consent form. After signing the consent form, the interviewer conducted the interview face-to-face, using a structured questionnaire developed for this study (eight domains, 176 questions, including 1) the study participants and their household (33 questions); 2) health status and use of health care services (46 questions); 3) life style regarding smoking, alcohol and exercise (6 questions); 4) homework - self-management of diabetes (19 questions); 5) informal social support (51 questions), including 6 questions related to unmet need for informal support); 6) sexual well-being and emotions (10 questions, including 5 questions related to distress); 7) use of smart phones, internet and other types of media (7 questions); and 8) comments to questionnaire, consent regarding further contact to support persons (4 questions) ‘Supplementary‘. In the present study, only domain 1, 2 and 5 were used. To maintain confidentiality, a village health worker could not collect data in her/his own commune and no other people than the interviewer and the participant could take part in the interview. Thus, the questionnaire contained questions that could be considered sensitive e.g. receiving support from family members and sexual health. All the participants were interviewed in their home.
Outcome variable
To assess the experiences of DRD in the study population, we used the Problem Areas in Diabetes scale (PAID) [8]. We chose PAID as it focuses on emotional concerns related to coping with diabetes and diabetes complications, whereas other tools such as the Diabetes Distress Scale is more focused on motivational and behavioural problems associated with diabetes self-management and emotions related to physicians involved in diabetes treatment [25, 26]. We used the short PAID form (PAID5) as it has shown high sensitivity (94%) and specificity (89%) with the 20 item PAID and was chosen to minimize the interview burden on the participants. PAID5 was translated to Vietnamese, back-translated into English, and pilot-tested in a group of Vietnamese persons with diabetes to test the understanding, interpretation, and perception of the questions. The questions were also assessed in light of the project’s qualitative component where we found that the questions resonated with local framings of the problems. In other words, the topics that the questions address were covered in the ethnographic research on the project before being put to use quantitatively ([23]; Gammeltoft, Tine M: The Force of Love: Type II Diabetes in Vietnam as Tentatively Transformative Experience, forthcoming). PAID5 contains five questions: 1) Feeling scared when you think about living with diabetes; 2) Feeling depressed when you think about living with diabetes; 3) Worrying about the future and the possibility of serious complications; 4) Feeling that diabetes is taking up too much of your mental and physical energy every day; and 5) Coping with complications. Answers were scored at five levels: not a problem (0 points), minor problem (1 point), moderate problem (2 points), somewhat serious problem (3 points), and serious problem (4 points). The total score can range from zero to 20, and higher scores indicates more serious emotional problems [2]. To identify more severe experiences of DRD, a score of ≥33 on the 20-item PAID has been shown appropriate [27]. This score corresponds to a total score of ≥8 points on PAID5 [8], which we used to assess if people in Thai Binh had experienced DRD.
Exposure variables
Social support is understood as help provided by family, friends, neighbors, relatives, and other informal care givers, and includes different areas of assistance, such as informational, emotional, companionship, and practical help [16,17,18]. An unmet need for social support was defined as inadequate amount of help received from informal sources in relation to daily management of diabetes [19]. We chose 6 areas of social support, which people with diabetes often receive from informal caregivers [16, 18] or have reported as unmet needs of support [28]. The 6 areas were: (i) Transport and company when visiting health facilities; (ii) Reminders to take medication; (iii) Purchase and preparation of food; (iv) Reminders to engage in physical exercise; (v) Emotional support; and (vi) Financial support related to any of the above-mentioned items. Answering options were: 1) I receive too little support: 2) I receive adequate support; and 3) I receive too much support. For each item, unmet need for the support was coded as a binary variable with “I receive too little support” coded 1, whereas the other answering options were coded as no unmet need for support (0). We also created a variable for the total number of unmet needs, categorized as 1, 2, 3, ≥4 unmet needs.
Covariates
Demographic and socio-economic characteristics including gender (man, woman); age group (40–49 years, 50–59 years, 60–69 years, and ≥ 70 years); marital status (single, living with spouse, not living with spouse/divorced, or widowed); occupation (unemployed/stays at home, farmer, small trade business/worker/government employee/private company, and retired); size of household (one member, i.e. living alone, two members, 3–4 members, and ≥ 5 members); self-reported economic status of the household (poor, medium, and wealthy); self-reported physical health (poor, fair, and good); and duration of diabetes (< 5 years, 5–10 years, and > 10 years).
Statistical analysis
The descriptive characteristics of people with T2D who experienced DRD are presented as frequencies. Chi-square tests were performed to compare the frequency of DRD by demographic, socio-economic, and health-related characteristics. A statistical significance level of p < 0.05 was used. We used logistic regression to model DRD as an outcome of each unmet need for social support to manage diabetes, and as an outcome of total number of types of unmet needs for social supports. All models were adjusted for age, gender, and household economic status (model 1). As a next step, we further adjusted for self-reported physical health (model 2). Lastly, we also included adjusted for number of co-residing household members, and duration of being diagnosed with diabetes. Results are presented as odd ratios (OR) with 95% confidence interval (95%CI). Data analyses were performed using SPSS (IBM Statistical Package for Social Science software) version 22.