Social Support for Self-Care: Patient Strategies for Managing Diabetes and Hypertension in Rural Uganda

Background: Low-income countries suffer a growing burden of non-communicable diseases (NCDs). Self-care practices are crucial for successfully managing NCDs to prevent complications. However, little is known about how patients practice self-care in resource-limited settings. Objective: We sought to understand self-care efforts and their facilitators among patients with diabetes and hypertension in rural Uganda. Methods: Between April and June 2019, we conducted a cross-sectional qualitative study among adult patients from outpatient NCD clinics at three health facilities in Uganda. We conducted in-depth interviews exploring self-care practices for hypertension and/or diabetes and used content analysis to identify emergent themes. Results: Nineteen patients participated. Patients said they preferred conventional medicines as their first resort, but often used traditional medicines to mitigate the impact of inconsistent access to prescribed medicines or as a supplement to those medicines. Patients adopted a wide range of vernacular practices to supplement treatment or replace unavailable diagnostic tests, such as tasting urine to gauge blood-sugar level. Finally, patients sought and received both instrumental and emotional support for self-care activities from networks of family and peers. Patients saw their children as their most reliable source of support facilitating self-care, especially as a source of money for medicines, transport and home necessities. Conclusion: Patients valued conventional medicines but engaged in varied self-care practices. They depended upon networks of social support from family and peers to facilitate self-care. Interventions to improve self-care may be more effective if they improve access to prescribed medicines and engage or enhance patients’ social support networks.

clinics. There, a nurse or clinician checks weight, blood pressure, and blood sugar; delivers health education; prescribes medications; and schedules monthly follow-up appointments. Medications are dispensed at a clinic-based pharmacy.

STUDY POPULATION
We included adult patients who 1) attended outpatient NCD clinics at one of the three selected facilities and 2) were diagnosed with diabetes and/or hypertension at least three months prior.

SAMPLING
We purposively selected participants for maximum variation by condition and gender, selecting at least two patients with hypertension, two with diabetes and two with both conditions from each facility. We selected at least one male and one female for each condition category. We initially selected 18 participants. During interviews, we learned about the phenomenon of informal patient group leaders. We therefore selected the nineteenth participant to include at least one patient leader. Research assistants approached patients in the waiting area to inform them about the study prior to receiving care. After the patient completed their visit, they were invited to interview.

DATA COLLECTION
Three (one male, two female) trained research assistants with backgrounds in social science interviewed participants using a semi-structured interview guide between April and June 2019. The guide was developed and reviewed by the research team and pretested among four patients with diabetes and hypertension. After pretesting, we revised study instruments to improve clarity and alignment with the overarching research question.
The interview guide (Appendix 1) included five key domains: perspectives on their condition(s), experiences receiving care, challenges to medication adherence, symptom recognition and responses, and practices intended to prevent adverse conditions or be well. All interviews were conducted in-person in the local language, Luganda, and lasted 30-45 minutes. Participants were compensated 15,000 Ugandan shillings (UGX), equivalent to $4 United States Dollar (USD) at the time of the study, for their participation. Audio recordings were transcribed verbatim in Luganda and translated to English.

ANALYSIS
We used a conventional approach to thematic content analysis to inductively derive categories [19]. A team of four researchers, including AKT and MAH, independently read and coded three transcripts to develop initial codes. Coders discussed and reached consensus about the initial code structure, then applied it to further transcripts while remaining open to new codes. Another meeting was held to finalize the structure. The resulting codebook was applied to all transcripts using Atlas.ti 8.5. Emerging findings were refined through works-in-progress sessions as part of a training grant (NIHD43TW009607-07) and review with co-investigators.

HUMAN SUBJECTS
We received approval from the Makerere University School of Medicine Ethics Review Committee, Yale Human Subjects Committee, and Uganda National Council of Science and Technology (HS 2698). Health facilities provided administrative clearance. All participants provided written informed consent.

RESULTS
We enrolled all 19 patients invited to participate ( Table 1). Their mean age was 55 years (standard deviation ±12). Most were farmers (n = 10, 53%) and had attained at least primary-level formal education (n = 17, 89%). Six (32%) participants were living with hypertension, six (32%) with diabetes, and seven (36%) with both conditions. Three overarching themes emerged. First, patients preferred prescribed medicines as their first resort. However, they used traditional medicines to mitigate the impact of inconsistent access to prescribed medicines or as a supplement to those medicines. Second, patients adopted a wide range of vernacular practices to supplement treatment or self-monitor in the absence of basic equipment or materials. Finally, patients relied on social support from family members and patient peers to mitigate the impact of uncertain access to prescribed medicines.

PATIENTS PREFERRED CONVENTIONAL MEDICINES BUT USED TRADITIONAL MEDICINES
Patients reported that clinician-prescribed conventional medicines were their key resource for selfcare. However, their ability to adhere to prescribed medicines was limited by inconsistent access. One explained: When patients bought medicines at private pharmacies, they typically secured fewer doses than prescribed.
Unable to access medicines at public facilities or afford those at private pharmacies, many patients sought other remedies. Some described using traditional remedies, including local herbs, bitter vegetables, and saltwater, in place of unavailable or costly prescribed medicines. Patients perceived that these remedies would prevent complications, reduce their blood pressure, stabilize blood sugar, and/or relieve pain in the absence of conventional medicines (Table 2A). Many saw herbal remedies as a stopgap for times when they were unable to obtain prescribed medicines: When patients were able to obtain all prescribed doses, some supplemented prescriptions with traditional medicines. They used traditional medicines to attain a desired health status, such as weight loss, relief from pain, or because they hoped to cure their condition (

PATIENTS ADOPTED MANY VERNACULAR SELF-CARE PRACTICES
Besides using medicines, patients restricted diet, sought ways to control stress, and devised self-monitoring strategies in order to maintain their health. They adopted a variety of self-care practices in response to specific symptoms, such as walking barefooted for relief when their feet were numb, soaking swollen feet in salt water, and massaging with warm water to relieve itching (  When patients with diabetes felt weak or experienced symptoms they linked to low sugar level, many said they consumed sweet things to restore blood sugar:

When I get a foam in the mouth to an extent that my tongue cannot move, I know that my sugar level is low. So, I lick some sugar or dissolve it in water and I drink. It is my first medicine. Even other diabetic patients have told me that I should take sugar the moment I notice that my sugar level has reduced. (IDI-12, DM, Female, 78)
Because patients with diabetes were anxious to ascertain their sugar level but could not obtain glucometers, they improvised self-monitoring practices. For example, some reported tasting urine to gauge sugar level: Patients saw these practices as adjuncts to conventional treatments and diagnostic tests to which they lacked access, rather than as preferred practices for maintaining health.

PATIENTS SOUGHT SOCIAL SUPPORT FOR SELF-CARE
Patients attempted to mitigate uncertain access to conventional medicines by drawing on other resources in their social environment. They typically saw their children as their most important source of emotional and instrumental support to sustain self-care. In particular, children were a primary source of instrumental support, providing money for medicines, transport, and necessities at home:

The children buy the medicines and bring them. They can bring for me some medicines from Kampala where they stay. Sometimes children send money and I buy medicines.
(IDI-2, HTN, Female, 52) Patients also looked to spouses for social support. Some husbands looked to their wives for emotional support and preparation of meals. Wives tended to look to husbands solely for financial support. These patients counted on their children for financial support for treatment when support from the spouse dwindled: Patients considered their children a final but frequently-called-upon resort for financial resources to sustain supplies of medicine.
Finally, patients reported receiving peer support through unofficial patient groups. Some -mostly diabetes patients-described organizing themselves into groups for the purpose of mutual support. In addition to emotional support, patients received instrumental support in the form of cash assistance, transport, or medicines from other group members. One patient described how his patient group organized and pooled resources for mutual support:

DISCUSSION
In this interview-based study of patients' self-care for diabetes and hypertension in rural Uganda, we found that patients used both conventional and traditional medicines, but preferred conventional medicines as their first resort. They also adopted vernacular practices to supplement conventionally recommended self-care practices. Partly in response to inconsistent access to conventional medicines, patients frequently sought support from family and peers.
Crucially, we found that patients relied on networks of social support to mitigate uncertain access to medicines and meet costs associated with care. Studies of chronic HIV care in Uganda demonstrate that families, friends and other significant persons can be important sources of similar support for adherence to anti-retroviral therapy [20,21]. Similarly, a study in Nigeria identified families as the most accessible source of financial support for patients with hypertension and diabetes and found that this marginally impacted health outcomes [22]. Limited economic resources may increase the importance of such social networks for accessing treatment [8]. We add to this literature the notable finding that patients relied most on their children for support, rather than peers or spouses-even when they characterized their children as financially unstable or living outside the household. We identify a critical need for interventions to facilitate additional sources of instrumental and emotional support for patients.
We also found that many patients with diabetes participated in loosely organized patient groups. Patients pooled their own limited resources to provide test strips, transport support, emergency funds for medicines, and even food relief. While groups spontaneously arose in clinic waiting rooms, they were not affiliated with the health facilities. This is consistent with an earlier study in Uganda, which found that informal diabetes support groups helped patients manage frequent stock-outs of medicines [23]. Our study affirms the centrality of group support as a psychosocial and financial resource for managing health among rural patients. This finding is particularly important because it suggests a viable local model for a cost-effective, patient-centered intervention to mitigate resource variability. Patient-driven groups can provide additional sources of peer support for all patients. Such groups may be cost-effective for supporting and extending existing conventional chronic care delivery models, particularly among low-income and vulnerable populations [24]. For instance, Uganda's differentiated service delivery platform, which offers facility-based groups for patients living with HIV, reduced total costs for antiretroviral drugs, laboratory commodities and health workforce needs [25].
Strengthening and scaling these groups for patients living with non-communicable diseases could address gaps in social support, decrease pressure on families, increase access to essential medicines and supplies, facilitate effective self-care, and improve adherence. Further research should develop and evaluate interventions to expand and strengthen the patient peer support model for rural settings, including mHealth tools to augment peer support. Building on the Centers for Disease Control's Chronic Disease Self-Management Program and existing differentiated care models established for HIV in LMIC, patient group interventions could include components such as interactive education, symptom management activities, and mutual support for the costs of care [25,26].
We also found that patients valued prescribed medicines and overwhelmingly preferred them for disease management. Indeed, patients may even overestimate the capacity of these medicines. However, patients also reported that their access to prescribed medicines was poor or inconsistent. Insufficient medicine is the commonest reason for poor adherence to medication in Uganda [7,27], where public facilities are the primary source of free medicines for patients with NCDs. Recent work has demonstrated that the majority of doses prescribed for diabetes and cardiovascular disease at these facilities are not dispensed due to real or anticipated shortage [28]. Therefore, patients must obtain their prescriptions elsewhere, if at all [27,29]. New strategies are urgently needed to improve access to medicines; this, in turn, could improve adherence and reduce reliance on non-recommended care practices.
Because they had no test supplies, patients in our study devised their own self-monitoring strategies. Chronic care in Uganda is persistently affected by poor availability of essential materials and equipment [30][31][32]. Similar gaps in NCD service provision have been reported in other developing countries, like Malawi [33], Ethiopia [34], and Bangladesh [35]. There is urgent need for strategies to improve access not only to medicines but also to essential materials and equipment. Strategies to improve self-care in these settings must attend to the limited resources available for self-monitoring at home. Patient groups may mitigate the impact of poor access to essential materials by enabling patients to pool resources, including test supplies [23]. Additionally, patient group activities could potentially influence patient behavior to reduce dependency on traditional medicines. Finally, we found that patients used traditional medicines in part to mitigate the negative impact of inconsistent access to prescribed medicines. The use of traditional medicines to compensate for rationed prescription medication may imply that patients perceive traditional treatments to play a similar role as prescribed medicines, as has been reported in other studies in Uganda and India [23,29,36]. Traditional medicines are common in Uganda and promoted by healers, peers, family members and media [7,23,29]. Notably, patients in our study did not report ceasing conventional medicine in favor of traditional medicine, as others have found in south-western Uganda [29]. Rather, patients in our study believed their prescribed medicines to be central to managing their condition. Additionally, patients in this study relied on biomedical monitoring to judge the effectiveness of the traditional medicines they consumed. This work should be regarded within the context of some limitations. Most notably, we recruited participants from health facilities. Participants may face fewer barriers to accessing care or have different health-seeking behaviors than the general population. Findings may not reflect experiences or practices of people not engaged in clinical care. Future research should examine patterns detected through this exploratory work in larger, representative samples. For instance, further investigation could explore how selfcare practices relate to culture, especially in settings where traditional medicines are central to self-care for NCDs. In addition, future research should explore providers' perspectives on the use of traditional medicines for self-care.
This study also has strengths. Our sample includes a balance of men and women with hypertension, diabetes, or both conditions across three different types of health facilities, enhancing its external validity. Second, our analytic strategy involved a multi-disciplinary coding team, consensus meetings about code structure, and reviews by all investigators, which facilitated both emergence of divergent interpretations and cross-coder reliability. Finally, this work generates novel directions for interventions to promote self-care for NCDs in low-income, rural settings.

CONCLUSION
Conventional prescribed medicines were a key resource for self-care among patients living with diabetes and hypertension in rural Uganda. However, patients faced uncertain and inconsistent