This qualitative study examined diabetes self-care in three South Ethiopian hospital clinics. Its 21 participants comprised 12 adults with type 2 diabetes and nine healthcare providers. Patients described medication, diet, physical activity, foot care, and glucose monitoring as daily work negotiated against financial pressure, drug shortages, beliefs, food costs, limited support, and health-system constraints. Patients tended to foreground structural barriers, whereas providers placed more emphasis on knowledge and forgetting advice. Continued counselling, reliable medicines, and peer support emerged as priorities, but these are interview-derived design directions rather than tested interventions.
Key findings
- Patients described five core self-care activities—medication, physical activity, diet, foot care, and glucose monitoring—but defined well-being more broadly than glucose control: being able to work, fulfil family roles, participate socially, and retain hope.
- Economic and health-system constraints were prominent. Nine of 12 patients lacked insurance; stock-outs forced private purchases, testing and recommended foods were costly, and some participants reduced doses or delayed follow-up because of cost.
- Patient and provider accounts diverged. Providers often foregrounded knowledge gaps and forgotten advice, while patients said they knew what to do but could not act when medicines were unavailable or unaffordable. Education alone may therefore miss a central bottleneck.
- Participants proposed repeated counselling, reliable medicines, and peer support. Although some patients welcomed peer networks, one divergent case raised stigma concerns, so a single peer-support format should not be assumed suitable for everyone. No programme was evaluated in this study.
Why this matters globally
The study adds lived-experience evidence from a resource-constrained sub-Saharan African setting that diabetes self-management is not merely an individual responsibility. Medicine availability, food prices, transport, insurance, and provider time shape whether advice can be followed. These findings can guide primary-care and non-communicable-disease service design in comparable settings, but cannot estimate barrier prevalence, effect sizes, or how much counselling, peer support, or supply-chain reform would improve HbA1c or complications.
Thai researcher contribution
Four authors have Thai institutional affiliations and publisher-specified roles. Temesgen Anjulo Ageru and Cua Ngoc Le are affiliated with Walailak University’s Public Health Research Program and Excellent Center for Public Health Research; Apichai Wattanapisit is with Prince of Songkla University’s Faculty of Medicine; and Charuai Suwanbamrung is with Walailak University. Ageru’s roles include conceptualisation, data curation, formal analysis, investigation, methods, resources, software, funding, and writing; Le’s include conceptualisation, data curation, investigation, supervision, validation, and visualisation; Wattanapisit’s include conceptualisation, data curation, formal analysis, methods, supervision, validation, and drafting; and Suwanbamrung’s include conceptualisation, data curation, formal analysis, funding, methods, project administration, resources, supervision, validation, and drafting. Wattanapisit and Suwanbamrung are marked as corresponding authors. Funding came from the Walailak University Research Fund (CGS-RF-2024/16).
Limitations to consider
The sample comprised only 12 patients and nine providers from three hospital clinics, providing depth rather than population estimates. People who had disengaged from care may face more severe barriers than clinic attenders. Self-report invites social-desirability bias; family caregivers were not included; and excluding adults over 60 omits multimorbidity, dependence, and isolation among older people. Translation from Amharic to English may have lost nuance despite cross-checking. Provider accounts contextualise the system but cannot substitute for patient experience. No HbA1c, medication-use, or other clinical outcome data link narratives to disease control. Full transcripts are not public because of ethical restrictions.