1. Background
Hepatitis C virus (HCV) is associated with substantial morbidity and mortality worldwide, which is mainly attributable to the sequelae of infection, including liver cirrhosis, liver failure, and hepatocellular carcinoma (HCC) [
1,
2,
3,
4]. In 2019, 58 million people globally were living with chronic HCV, and approximately 290,000 people died from the virus [
5]. Between 55% and 85% of acute infections progress to chronic HCV if not treated early [
6]. Notably, while deaths from AIDS, malaria, and tuberculosis have declined [
7], HCV-related mortality remains high, based on 2021 World Health Organization (WHO) estimates [
5]. HCV is also associated with a high economic burden, both with high drug costs and the costs associated with disease progression [
4,
7,
8,
9,
10]. Liver cirrhosis was among the top ten causes of death in low-income countries in 2019 [
7].
The advent of a short course, well tolerated, and highly effective HCV treatment has revolutionized the management of the disease. However, while timely diagnosis and delivery of antiviral therapy can cure and prevent progression to later stages of disease, access to HCV diagnosis and treatment has been limited [
9,
11,
12,
13]. In 2019, only 21% of people living with HCV infection globally knew their status [
6], and approximately 62% among those diagnosed had access to treatment. Access to HCV diagnosis and treatment is particularly challenging in sub-Saharan Africa (SSA) [
14], despite the region accounting for 20% of global infections [
15].
One reason why newer treatments for HCV have had limited use in SSA countries is their high cost and the costs associated with viral testing. However, in recent years, manufacturers have shown a willingness to lower these prices in resource-limited settings [
16]. For example, in 2017, the government of Rwanda negotiated a price reduction for direct-acting antivirals (DAAs) that reduced treatment costs in the country from USD 1200 in 2015 to USD 60 in 2018 [
17,
18,
19].
In 2016, the World Health Assembly set an ambitious goal of eliminating HCV by 2030. In 2017, the World Health Organization published five strategies to guide countries willing to eliminate HCV as a public health concern. These strategies include continuous collection and analysis of HCV-related data to inform policymakers, increasing screening and treatment coverage, sustainable and scalable financing, and developing new affordable diagnostics, vaccines, and treatments. In line with this, the government of Rwanda initiated a mass screening and treatment program across the country and launched a 5-year HCV elimination plan [
20,
21]. While similar programs have been considered elsewhere, there is a lack of evidence regarding the impact of such programs in terms of the progression of patients through the care cascade and outcomes, particularly in SSA. The HCV cascade of care is an approach used to evaluate patient retention throughout the different stages of care that are needed to achieve HCV treatment success [
22].
Evidence related to the HCV cascade remains limited to high-income countries where medicine accessibility, laboratory equipment, and HCV awareness are high. A recent global systematic review that evaluated estimates on the cascade of care for HCV included mostly high-income countries, and no country in SSA had estimates of their HCV cascade [
23]. Furthermore, while the number of studies reporting HCV epidemiology has been increasing, there remains limited evidence on the total number of people who initiate treatment and their treatment outcomes in SSA [
3,
24]. SSA is among the regions with the highest burden of HCV and other co-infections, such as HIV; therefore, it is crucial to review the HCV cascade in these settings.
Prior studies have estimated the HCV population prevalence in Rwanda to be between 4% and 8% [
19,
23]. One study that assessed the cascade of care in two rural districts of Rwanda found that 83.4% of patients tested for HCV had detectable viral loads. Almost everyone considered for treatment initiated treatment, while 93.7% of patients achieved SVR12 [
25]. Although this study provided information on patients’ journeys in HCV care, the findings cannot be generalized, since the study was conducted in only two of the poorest districts of Rwanda and used a small sample size. Additionally, over 40% of the patients who initiated treatment did not have SVR12 test results. Therefore, although DAA treatments are highly effective in curing HCV, treatment success represents only a small proportion of the study population. Lastly, the study did not assess factors contributing to outcomes and dropouts at each stage of care. Therefore, we studied the progression of patients through the cascade of HCV care in Rwanda’s general population using data from all public hospitals providing HCV care. We also assessed factors contributing to outcomes at each stage of the care cascade and dropouts.
2. Materials and Methods
2.1. Study Context
Over the past 20 years, Rwanda has shown progress in improving access to healthcare services across the health sector [
26]. However, this same progress has not been seen in HCV care: in 2015, less than 1% of the estimated 55,000 HCV patients received drug treatment [
14]. Before 2015, patients who obtained drug treatment could only do so through four hospitals that had hepatitis C specialists, and it was delivered to only a small group of comparatively wealthier patients [
27].
To take advantage of the negotiated price reduction, the Rwanda Ministry of Health (MoH) launched the first voluntary mass screening and treatment campaign for HCV. The program first targeted people living with HIV and was subsequently expanded to the prison population and then to the general population. The general population screening was performed in age cohorts, starting with those who were 45 years and above and then everyone who was 15 years and older at the time of the campaign. Testing and treatment were available at one or more public health facilities (health centers or hospitals) in each of Rwanda’s 30 districts.
2.2. Study Design
We conducted a retrospective cohort study using secondary data to describe the HCV treatment cascade of care and assessed factors associated with outcomes and dropouts at different stages of care.
2.3. Study Sites and Population
We studied all patients screened at 46 of 47 public hospitals that provided HCV testing and treatment between April 2017 and October 2019. The excluded hospital is a neuropsychiatric hospital that did not provide HCV services. We also excluded patients who had indeterminate screening tests, refugees, those who died during follow-up, and those who received care from private health facilities. Eighty-five percent of Rwanda’s population seeks care at public health facilities [
28], and during the mass screening, patients were only referred to public hospitals.
2.4. Data Sources
We used an electronic database compiled throughout the screening activities by the Rwanda Biomedical Centre (RBC). This database contained patients’ sociodemographic characteristics and other risk factors collected at the screening stage. This database has previously been used in analyses of HCV, HIV, and HBV in Rwanda [
24,
28,
29]. We linked this database to additional information on treatment initiation and treatment outcomes extracted from patient charts. Each HCV patient had a record at the hospital that contained demographic characteristics, clinical variables (e.g., treatment and testing results, medical history, and behavior characteristics), and hospital visit dates. We used a unique identification number and demographic characteristics to link data extracted from patients to the electronic database. Data extraction was conducted by two trained nurses in each hospital who were overseen by an HCV nurse mentor.
2.5. Cascade of Care
The cascade of care for patients in Rwanda consisted of three major stages:
2.5.1. Screening
Anti-HCV screening took place in public venues, including stadiums and playgrounds, or near other public places, such as markets, and was free of charge. Each district had at least one screening center. Trained nurses collected demographic information and blood samples from attendees at the screening venues [
29,
30]. The venous blood (5 mL collected in an ethylenediamine tetraacetic acid tube) samples and laboratory request forms were transported to the nearest of 13 facilities that provided enzyme-linked immune sorbent assay (ELISA) services. Murex-ELISA (version 4.0; DiaSorin S.p.A., Italy) was used across all testing sites. The samples were stored at room temperature (20–25 °C) for less than 12 h without refrigeration and then transported to laboratory facilities for ELISA testing. The sample transportation used a prespecified route to ensure efficient distribution across testing sites. At the testing sites, samples were tested for the presence of HCV antibody (anti-HCV), hepatitis B (HBsAg), and HIV. Starting in April 2019, capillary blood was drawn to test HCV antibodies using SD Bioline™ rapid diagnostic tests (RDTs) at the screening venues. Finger pricks were conducted for the present patients and immediately tested at the screening sites. These tests had near 100% sensitivity and a specificity of 99.4% [
31]. Although RDTs were pre-approved by WHO for HCV screening, the Rwandan Ministry of Health, through the national referral laboratory, also conducted an internal validation to evaluate their performance before official use in Rwanda. The findings showed 97% sensitivity and 99% specificity (unpublished internal report). Patients who had a positive antibody test were provided with information on RNA viral load testing and were referred to a specialist for HCV RNA testing at their nearest hospital, where a blood sample for HCV RNA testing was obtained. From 2019, after a decentralization and task shifting intervention, patients were treated at the nearest health center by trained nurses and doctors. Those who screened as anti-HCV positive had to pay the fees associated with HCV RNA testing and additional assessment tests using health insurance or pay out-of-pocket. The cost of screening in Rwanda is approximately USD 1, while HCV RNA costs USD 9, which is equivalent to 10,000 Rwandan francs.
2.5.2. HCV RNA Testing
Following referral and sample collection, trained laboratory technicians conducted HCV RNA testing at 9 HCV viral load testing sites [
21]. Some blood samples were collected at hospitals that did not provide viral load testing, so they were transferred on the same day to testing sites. Each sample had patient and hospital identification information. The hospital at which the sample was taken called patients to inform them of their results. Patients who were confirmed to be positive received counseling and were provided information on the treatment process by an HCV specialist and referred for treatment.
Figure 1 provides more details on the screening and testing process.
2.5.3. Treatment
All patients who had a positive HCV RNA test (detectable 15–20 copies/mL) received treatment free of out-of-pocket charges [
29,
30]. During the study period, the HCV treatment available in Rwanda included ledipasvir/sofosbuvir (Harvoni
®) and sofosbuvir/daclatasvir (DCV). DAAs are known to have a higher cure rate (90%) and fewer side effects than other therapeutic options [
23]. After treatment completion, all patients were asked to return to the hospital for viral load testing. All patients whose HCV results were undetectable (below 15–20 copies/mL) within 12 weeks after the date of treatment completion were considered cured.
2.6. Outcome Variables
Table 1 lists the outcomes that we used to describe the cascade of care for patients with HCV. Along with the number and proportion of patients who completed each stage of care, we also estimated the proportion of patients who dropped out at two stages of the cascade: (1) patients with no confirmed diagnosis and (2) patients who did not return for their final check-up 12–24 weeks following treatment completion. We defined treatment failure as having an HCV viral load that was above 20 copies/mL between 12 and 24 weeks post-treatment.
2.7. Other Variables
We included covariates identified in previous studies [
14,
22,
32], including the following demographic and socio-economic factors: age, sex, Social Economic Status, health insurance, and marital status. In Rwanda, households are classified into four different Social Economic Status (SES) (known as ubudehe) categories based on their household income and assets, with category 1 being the lowest wealth group and category 4 representing the highest wealth group [
33]. This variable includes a category called unknown that represents new families that have not yet been classified into any SES category.
We also collected clinical and behavioral features, including previous HCV treatment, hepatitis B vaccination status, previous transfusion and surgical history, family history of HCV, comorbidities (hepatitis B, HIV, diabetes, liver diseases, cancer, and renal failure status), and traditional operation practices. For treatment outcomes, we included variables related to the type of drugs a patient received, previous HCV treatment history, baseline viral load, and cirrhosis status (stage of HCV). Lastly, we included hospital-level variables, such as number of staff and hospital type (referral, teaching, provincial, or district).
2.8. Data Analysis
We started by generating descriptive statistics to estimate the proportion of patients at each stage of care and treatment success, followed by bivariate analysis using chi-square tests and t-tests to assess the association between covariates and outcomes. Following this, we used a hierarchical logistic regression model to assess factors associated with HCV positivity, dropout at different stages of care, and treatment failure while controlling clustering in the data. Continuous variables were centered or standardized for modeling and interpretation purposes. Some covariates had missing values, and we included a missing category for each variable with missing values to enable us to retain all observations in the analysis [
34].
Our model-building process included three steps. First, we estimated the interclass correlation coefficient using an empty model. Second, we included all patient-level and hospital-level variables that were significant in the bivariate analysis and those that were important based on prior literature. Covariates were included in the model using a forward stepwise selection process using the AIC and the likelihood ratio test. Finally, we tested all interaction terms between variables for significance at α = 0.05 and retained those that were significant. All analyses were conducted using STATA 16.0.