Health System Factors Serving as Facilitators and Barriers to Rheumatic Heart Disease Care in Sudan

Background: Rheumatic heart disease (RHD) remains a leading cause of morbidity and mortality in sub-Saharan Africa despite widely available preventive therapies. In this study, we sought to characterize facilitators and barriers to RHD treatment in Sudan. Methods: We conducted a mixed-methods study, collecting survey data from 398 patients who had enrolled in a national RHD registry between July and November 2017. The surveys included information on demographics, healthcare access, and patient perspectives on treatment barriers and facilitators. Factors associated with increased likelihood of RHD treatment adherence were assessed using multivariate logistic regression. These data were enhanced by focus group discussions with 20 participants, further exploring health system factors impacting RHD care. Results: Our analysis revealed that female gender (Odds ratio (OR) = 1.94; 95% CI 1.14-3.29), increased household income (monthly income above 2000 Sudanese pounds; OR = 2.02; 95% CI 1.26-3.25), and a perceived appropriate level of healthcare stang by doctors and nurses (OR = 2.07; 95% CI 1.16-3.71) were all factors associated with increased odds of treatment adherence. Further, qualitative data revealed signicant barriers to RHD treatment arising from health services factors at the systemic level: The primary barriers to RHD treatment included patient lack of understanding of disease, lack of access due to inadequate healthcare stang, lack of faith in local healthcare systems, and poor ancillary services. Facilitators of RHD treatment included stronger interpersonal support. Conclusions:

with RHD are advised to take prophylaxis in the form of intramuscular benzathine penicillin G (BPG) monthly for either ten years or until the patient turns 21 years of age (whichever is longer), with some guidelines suggesting lifelong prophylaxis. [8][9][10] Secondary prophylaxis has proven to be effective in preventing progression of disease in RHD patients. 9,11,12 Despite the e cacy of BPG, the REMEDY study, a multicenter RHD survey across LMIC settings, revealed that over 20% of enrolled patients did not regularly receive a monthly dose. 3 Previous studies in other countries have suggested potential reasons for low treatment adherence in RHD patients, including urban versus rural setting, education level, pain associated with injections, and availability of transportation funds. 13 Other qualitative analyses eliciting attitudes towards treatment and barriers to secondary prophylaxis found that key impediments to adherence included lack of resources (transportation, medications, clinic availability), injection pain, and poor communication between patients and providers. 14 In Sudan, low adherence rates to secondary prophylaxis in populations with RHD are believed to be linked to high rates of preventable heart valve injury. However, there is inadequate knowledge of the true adherence rates and subsequent correlation with valvular disease due to a lack of adequate tracking in national public hospitals and registries. There is also a lack of pre-existing literature describing causes of low adherence rates in Sudan.
A more comprehensive understanding of treatment adherence in the RHD patient population in Sudan will enable public health interventions to effectively target the root causes of non-adherence and help providers limit further complications from RHD. Thus, we conducted a mixed methods analysis evaluating barriers and facilitators to RHD treatment adherence, utilizing a comprehensive patient survey, along with a series of in-depth focus group discussions.

Study Population
We completed a mixed methods study (qualitative study with quantitative methodology) at the Alzaeim Alazhari University in Khartoum, Sudan. We collected survey data from 397 patients aged 12 to 90 years who were enrolled in a regional RHD registry between July and November 2017. The population included all patients within the university hospital catchment areas of Al-Shaab Hospital and Ahmed Gasim Hospital. Survey participants were selected via consecutive sampling using admitted patients or individuals attending routine clinic visits. Surveys were administered during the hospital admission or after clinic appointments by Sudanese medical students or physician trainees in English or Sudanese Arabic depending on participant preference.
To enrich our analysis by identifying key themes not captured by our surveys, we conducted four focus group interviews with 20 patients and patient family members aged 20-66. Focus group participants were selected via convenience sampling that occurred during clinic visits. All participants gave written consent to be included in the analysis. The study was conducted in accordance with the National Research Ethics Review Committee of Alzaeim Alazhari University, No. 4-5-2017, and Institutional Review Board of Stanford University, Protocol #40884.

Study Design
The surveys included data on demographics (age, household income, education level), healthcare access (distance from facility, insurance status), and opinions on treatment barriers (Supplementary Material 1).
The primary outcome variable was medication adherence, de ned as survey responses that indicated monthly BPG prophylaxis based on the prior six months. Since six injections in the preceding six months would indicate monthly BPG prophylaxis (standard treatment), participants were classi ed as nonadherent if they reported fewer than six BPG injections and adherent if they reported six or more injections. Patients and the public were involved in the study design, by providing the list of most common barriers to care for RHD patients, participating in the pilot survey, and ensuring that the research was culturally appropriate.
Next, focus groups were administered by two Sudanese psychologists, who were trained in qualitative research methods and performed as semi-structured interviews of 3-4 patients and/or family members in Arabic using a prepared focus group discussion guide created by the investigators (Supplementary Material 2). Themes of both documents include speci c programmatic and more general barriers and facilitators of receiving RHD treatment, speci cally focusing on BPG treatment.

Data Analysis
Demographic variables and survey responses were described as counts, medians, and proportions (%). To assess which demographic and healthcare system factors were associated with an increased likelihood of adherence, we rst constructed univariate binomial logistic regressions with each variable. To control for major demographic confounding factors, we constructed a multivariate logistic regression model using forward stepwise variable selection with a signi cance cutoff of p-value < 0.05. For both the univariate and multivariate logistic regression models, a p-value of < 0.05 was used to determine statistical signi cance regarding hypothesis testing. Robustness of the multivariate model was assessed using the Hosmer-Lemeshow goodness-of-t test and area under the receiver operating characteristic curve (ROC). All quantitative statistical analyses were completed using Stata-SE, version 16.1 (College Station, TX).
Focus group interview data were translated from Arabic into English by the research team and coded until thematic saturation was achieved using the Dedoose qualitative analysis package (Los Angeles, CA) and Microsoft Word (Redmond, WA). Two independent readers from the analysis team (JE, AC) reviewed the transcripts and subsequently compared coding for the purposes of internal validity. Differences in coding were resolved with discussion. From this coding process, key themes and concepts were identi ed and classi ed.

Demographic Factors
The demographic distribution of the survey (Table 1) revealed that the participants were mostly female (74.8%) and their ages ranged from 12 to 90 (median 40). Most of the respondents were homemakers or unemployed (72.5%) and had limited formal education, with the majority reporting their highest level of education as primary school or no formal schooling (66.7%). Nearly all participants (94.7%) reported a household monthly income less than 4000 Sudanese pounds (~ 88 US dollars) with a median household size of 6 people. 28.7% of participants reported that they had a history of carditis, though 51.4% of participants claimed to have had heart valve surgery. Only 32% of participants were found to be optimally adherent to BPG prophylaxis.

Qualitative Analysis
Twenty individuals aged 20-66 (median age of 41) were represented in four focus groups, with eleven patients and nine patient family members. 55% were female, 55% reported a monthly income level less than 2000 SDG, and 75% of participants lived in a rural setting (Table 3). Of the eleven patients, four (36.4%) had undergone valve surgery. Analysis of the focus group discussions revealed recurring major and minor themes that were categorized as facilitators or barriers to receiving treatment in one of three domains: individual, interpersonal, or systemic ( Fig. 1). She is very active and goes out with us without being annoyed from her disease and -thanks to Allah-she has done the operation trusting Allah that she will be ok and recover. She is the one who is encouraging me to do the operation and get relieved from this stress. Praise to God she was travelling to the operation the next day and we had the wedding of her cousin in our same house and everybody was coming to say farewell to her. She was normal and the wedding proceeded normally and she did not suffer, God protected her." Family support Emotional support during treatment The rst time they said they widened a valve, I wondered are they going to widen both valves? I left and came a year later, they said they would do a widening operation but after 9-10 days I got even worse, so I just paid for nothing. After this operation I laid in bed for 4 years and couldn't move. I'm from Al-Duwaim and in the operation there was hemorrhage that they sealed shut and they didn't tell my kin about it. Eventually they made us pay for the operation. There were so many shortages and mistakes in the past that I could have avoided." "Also getting a de nite date for surgeries is important. Especially for people who live in far areas. Meeting a doctor several times and not getting a de nite date is frustrating. It would be very much comforting to be assigned a date at an early time." "Her illness became when we went to the doctor and he didn't give her good medicine. The medicine itself causes diseases." Poor quality of ancillary services (nursing) in local hospitals "In nursing there are some clever people and they are in the right place but there are those who are in the wrong place and these maltreat people. The responsible people in the ministry of health when they appoint people should send them for training for two to three years not two or three months and come back to the hospital and they still have lessons and other things [training] left." Long waittimes for treatment Delay in diagnosis "There is a lot of pressure on the emergency department, they see more than a 100 [patients] and may put patients in beds and the oor. I met a fair colored female doctor when I came from the morning, I was sitting on a chair and we lay down on the stairs till the sunset [before being seen]." Delay to see specialist "At the hospital… you have to wait. You get to the ER where they sedate you and after that you meet the specialist on Monday or someday and so you nd that you waste time." Delay to get surgery " The operation takes a lot of time [before occurring]. For one patient, it was one month, but for [me] it was one year." Perception of poor quality care by younger/less experienced providers No subtheme present "With all due respect, doctors are different, some of them are useless and don't know anything. They are carrying stethoscopes and examine you and write a prescription which do not make you better and even make you worse. In these hospitals there should be doctors with good certi cates, who are well trained and experienced not those who just graduated 2 or 3 months and entered a hospital. These won't treat you, they will make you more sick." Perception of higher quality care in other settings No subtheme present "Now I have a [insurance] card and I go for treatment in Asia at a private hospital. They are good and cure me." "All these people are sick, 95% of Sudanese are sick but don't know where to go and get treated. All people now go to Egypt, they take a bus for 600 SDG going, 600 SDG going back and take 10 or 20 thousand SDG and go and get treated well and return back." Individual facilitators to receiving RHD treatment included adequate knowledge of RHD by the patient, perception of improved symptoms with treatment, and the positive in uence of faith on attitude towards treatment. Interpersonal facilitators of treatment were a motivation to undergo treatment as a means of supporting family members, multifactorial family support with treatment, and a positive in uence from other's treatment success. One participant recalled consistent support from their mother, stating "All of this was a Individual barriers to receiving RHD treatment included misconception of the disease process or treatment, as well as pain from BPG injection. The various misconceptions regarding RHD included one participant stating that the cause of RHD was "…drinking too much coffee and cigarettes. All these things cause complications and affect the heart because the heart is the main building which pushes the body. Also, too much stress affects the heartbeat, which affects the heart arteries" (Table 4B) Systemic barriers to treatment adherence were numerous, including poor quality of disease/treatment education by providers, signi cant costs associated with treatment across several domains, perception of lower quality care by younger/less experienced physicians, a lack of trust in local healthcare systems, a lack of transportation to appropriate healthcare facilities, long wait times associated with treatment, and a perception of higher quality care provision in other settings. Highlighting systemic issues with transportation, one participated noted, "Transportation is a major problem between the hospitals and from the house to the hospital. Autumn is 3-4 months [long] and because of the rain you can't leave your house and get medical care; there is also the issue of the bus fee."

Discussion
Our study offers a detailed survey of the barriers and facilitators to receiving adequate RHD care in the urban setting of Khartoum, Sudan. Although prior analyses have offered estimates of the epidemiologic characteristics of RHD, few have focused on the speci c LMIC healthcare system barriers to treatment contributing to the disparate prevalence compared to that of high-income countries. This study's strength lies in its mixed-methods design that bolsters the ndings of our quantitative surveys with themes independently identi ed in our focus groups. Applying these methods to the relatively poorly studied region of Sudan allows us to identify targeted interventions for context-speci c issues.
Re ecting on the patient-level barriers identi ed in the survey, a lack of understanding of RHD was identi ed as the primary barrier to receiving adequate care that was most strongly associated with poor BPG adherence. This nding was reinforced by the major themes of poor disease education quality and poor communication between patients and providers in the focus groups. Though this appears to be a robust factor, poor education on RHD can be addressed in a multitude of cost-effective ways. Patient education campaigns, public service announcements (PSAs), and healthcare provider training are all lowcost interventions that can improve patient understanding of disease and serve as both primary and secondary prevention of RHD. In particular, PSAs have been successful at increasing patient knowledge in LMICs in the Caribbean, though sociopolitical differences must be accounted for when adapting that strategy to SSA. 12,15 Female gender and a higher household income were also found to confer a higher likelihood of adherence with RHD treatment in univariate analysis whereas suburban household setting was associated with a decreased likelihood of adherence, which provides insight into possible appropriate targeting of educational interventions. In Zambia, a public-private partnership exempli ed the e cacy of targeted educational interventions based on the results of mixed-methods research, which serves as a model for designing interventions in Sudan from this project's ndings. 16 The systemic barriers to treatment adherence identi ed in our study (high treatment costs and limited access to appropriate healthcare) overlapped with prior studies in Uganda. 14 Our analysis is unique in eliciting patient attitudes toward health system factors, and we identi ed that perceived inadequate healthcare sta ng was a robust barrier to optimal BPG adherence given its signi cance in univariate and multivariate analysis. In qualitative analysis, this was bolstered by our nding that ancillary health center services such as facility cleanliness, nutritional services, and security also impacted patient trust in the RHD care system. This further outlines the long-term healthcare infrastructure shortfalls existing in Khartoum as it relates to care for chronic illness. Partnership with the Sudanese government can most feasibly be achieved through following the needs assessment tool for developing effective RHD programs, as outlined by Zühlke and colleagues. 17 Based on that approach, this study's results combined with additional stakeholder interviews in Sudan could be used to design community-based interventions.
Another approach to achieving policy change could involve guidance from cost-effectiveness models, which may inform policymakers of the prudent investment in primary and secondary prevention costs as compared to the workforce and surgical repair costs associated with severe RHD, as was proposed by researchers in Kenya. 18 Though transportation costs were not found to be a statistically signi cant factor associated with lower treatment adherence in our survey, transportation issues were highlighted as major themes throughout our focus group discussions. Given the large distances between many Sudanese towns and its capital, Khartoum, it is understandable that transportation remains a major barrier to receiving care, as identi ed by 17% of the cohort. Transportation issues could be alleviated through the use of mobile health clinics, such as those utilized to mobilize maternal health care in Sudan in 2015. 19 Another approach would include the decentralization of RHD care at specialized district level centers, as was done in Uganda. [20][21][22] Similarly, though injection pain was only cited by 5% of participants as the primary barrier to treatment adherence, it emerged as a major theme in focus group discussions. Because of the higher e cacy of BPG injections over oral penicillin equivalents for secondary prevention, however, this is unlikely to be addressed outside of providing analgesia to those patients. 23,24 Regarding primary prevention, further development of a previously studied GAS vaccine could provide primary prevention of GAS infections and its complications. Unfortunately, promising vaccine candidates have displayed cross-reactivity with human tissue and the high number of GAS subtypes makes vaccine development challenging at present. [25][26][27] In addition to tackling the barriers to receiving adequate RHD care, emphasis should be placed on the identi ed facilitators of treatment adherence. Family support was revealed as a facilitator of treatment adherence in focus groups, with many participants noting that family members helped with treatment costs, transportation and lodging associated with referral hospital visits, while providing emotional support. This contrasts somewhat with the survey data, which suggested that 89% of respondents selfreported strong family support, even though the adherence rate in that cohort was 32%.
The positive experiences of RHD patients at referral hospitals can also serve as a template for local medical facilities providing care for RHD patients. Though many of the inadequacies of local facilities were not limited to RHD care, these centers can use this information to improve the perception of care quality by those in their communities, for example through the establishment of RHD Centers of Excellence, as has been done elsewhere in sub-Saharan Africa. 28 Moving forward, a comprehensive approach to RHD control, such as the SUR I CAAN program adopted in Sudan, offers a practical model for addressing this multifactorial issue resource-limited settings. 29

Limitations
There are several limitations to our current study. As this was a cross-sectional study design, associations observed in our study cannot be considered causal and are susceptible to unmeasured confounding variables. Further, the study represented a convenience sample of patients during the enrollment time period, presenting an opportunity for selection bias. This is partially mitigated by the large number of participants in the study. Lastly, though this study's ndings would ideally be extrapolated to develop interventions for other resource-limited settings, the unique geopolitical situation of Sudan may limit the generalizability of this study.