Improved Hepatitis C Cure Cascade Outcomes Among Urban Baby Boomers in the Direct-Acting Antiviral Era

Abstract

Objectives:

We compared outcomes of the hepatitis C virus (HCV) cure cascade (ie, the path a patient follows from diagnosis to cure), including antiviral treatment outcomes, from 2 HCV screening programs. Our objective was to assess whether treatment uptake and HCV cure rates improved in the cohort screened after the release of all-oral HCV direct-acting antiviral therapies.

Methods:

We retrospectively compared outcomes of the HCV cure cascade from a cohort of newly diagnosed patients screened during 2012-2014 (period 1) with outcomes from a cohort of newly diagnosed patients screened during 2015-2016 (period 2) at Grady Health System in Atlanta, Georgia. Cure cascade outcomes included HCV antibody (anti-HCV) and RNA testing, linkage to care, antiviral treatment, and sustained virologic response.

Results:

During period 1, 412 of 5274 (7.8%) persons screened were anti-HCV positive, and 264 (69.3%) of those tested were RNA positive. During period 2, 462 of 7137 (6.5%) persons screened were anti-HCV positive, and 240 (59.3%) of those tested were RNA positive (P = .003). The percentage of newly diagnosed patients who were treated during period 2 (64.0%) was 3 times that of newly diagnosed patients treated during period 1 (21.2%; P < .001). Both cohorts had similarly high levels of linkage to care (95.8% during period 1, 95.4% during period 2) and cure (92.6% during period 1, 95.5% during period 2).

Conclusions:

Over time, the prevalence of anti-HCV and HCV RNA positivity declined substantially, and linkage-to-care and cure rates remained high. Treatment uptake increased significantly after the introduction of all-oral direct-acting antiviral therapy. These findings suggest that combining large-scale screening initiatives with treatment programs can speed progress toward HCV elimination.

Hepatitis C virus (HCV) infection is a major public health threat in the United States; an estimated 3.4-6 million persons had chronic HCV infection in 2010.¹ The number of deaths due to hepatitis C rose in the past decade, driven by a rise in end-stage liver disease and hepatocellular carcinoma.² HCV infection is also costly. Costs to the US health care system are estimated to top $9 billion per year by 2024.³

On a positive note, treatment for HCV infection has transformed in recent years. The introduction of all-oral direct-acting antiviral (DAA) therapies has led to a substantial improvement in the tolerability and efficacy of HCV treatment compared with older regimens. In 2014, second-wave DAA therapies became available to treat multiple genotypes.⁴ DAA therapy is effective and well tolerated, including in underserved populations that are historically difficult to reach for screening and linkage to care.^4-6

However, effective antiviral therapy alone cannot achieve HCV elimination by 2030, a goal set by the World Health Organization.⁷ Elimination relies heavily on targeted screening and treatment strategies, given that only 50% of persons infected with HCV have been diagnosed.^8,9 In 2012, the Centers for Disease Control and Prevention (CDC) updated its HCV screening recommendations to include one-time screening for persons born during 1945-1965 (ie, baby boomers).⁹ This recommendation was supported by the US Preventive Services Task Force in 2013.¹⁰ Since 2013, screening programs targeting baby boomers have shown success.¹¹

We previously described 2 HCV screening and linkage-to-care programs that ran successively, one in 2012 and the other in 2015.^12,13 Both demonstrated the feasibility of high-yield birth-cohort HCV screening and linkage to care in an urban, underserved, and predominantly African American population. However, treatment uptake and cure rates, the downstream results of these programs, were not compared. The objective of this study was to compare 2 cohorts of newly screened hepatitis C antibody (anti-HCV)–positive patients at the Grady Health System (GHS): 1 cohort screened before the introduction of all-oral DAA therapies and the other screened after DAAs were introduced. We compared these 2 groups for progression along the HCV cure cascade (ie, the path a patient follows from diagnosis to cure), including rates of treatment uptake and HCV cure.

Methods

We conducted a retrospective study of patients screened for HCV infection at GHS to compare outcomes of the HCV cure cascade during 2 screening periods: 2012-2014 (period 1) and 2015-2016 (period 2).

Grady Health System

GHS is a safety-net health system in Atlanta, Georgia, that serves an inner-city population and has approximately 620 000 visits per year. Clinical sites include the Primary Care Center located at the main hospital and satellite clinics located within 10 miles of the main hospital. Patients move fluidly between these centers and the main hospital for specialty care. Most patients are African American and have an annual household income <$20 000. Approximately 30% of patients are uninsured.

Grady Liver Clinic

The Grady Liver Clinic (GLC), located in the main hospital, was founded by general internists in 2002 to provide comprehensive HCV care and treatment for underserved patients. GLC provides care targeting each milestone along the HCV cure cascade¹⁴ (Figure 1).

Figure 1.

Sequential milestones of the hepatitis C virus (HCV) cure cascade for anti-HCV–positive patients, Grady Health System, Atlanta, Georgia, 2012-2016. Abbreviations: anti-HCV, hepatitis C antibody; SVR12, sustained virologic response 12 weeks after completing therapy.

GLC staff members introduced an HCV screening and linkage-to-care program in the Primary Care Center in October 2012. Screening activities ran through September 2014, and the linkage-to-care program continued through March 2015. The program involved training internal medicine residents to implement screening in their primary care clinics via automated electronic health record reminders and one-on-one education.¹⁵ The program focused on screening baby boomers (ie, birth cohort) and excluded tests performed on non–birth-cohort patients. Most patients who tested anti-HCV positive were linked to care at GLC, and a minority of patients were linked to GHS primary care clinics or infectious disease clinics. We included in the analysis only patients who received care at GLC.

A successive HCV screening and linkage-to-care program was implemented in October 2015 and was analyzed through September 2016. During period 2, the program was expanded to include 5 GHS satellite clinics. These clinics accounted for an additional 90 000 patient visits per year in addition to the 60 000 patient visits per year at the Grady Primary Care Center. Physicians at the community clinics received education about HCV screening and an electronic health record best practice advisory. During period 2, all patients newly screened—including non–baby boomers—were counted. We included in the analysis only patients who received care at GLC.

Data Management

For the period 1 cohort of patients, a data manager extracted data monthly from the electronic health record into a database.^12,15 Twenty months after completion of the screening program, we reviewed the electronic health records of patients who were newly identified as anti-HCV positive during the screening program. We abstracted data from electronic health records for the type of initial linkage to care (infectious disease, primary care, or GLC), number of follow-up visits, number referred for antiviral treatment, reasons for nontreatment, and treatment response.

For the period 2 cohort of screened patients, a data manager extracted data monthly from the electronic health record into a database. We generated a summary report of anti-HCV–positive patients who were newly identified from October 2015 through September 2016 from that database on April 1, 2017. The report included data on milestones in the cure cascade, including anti-HCV result, HCV RNA result, and linkage-to-care location and date. The report did not include information on HCV antiviral treatment. To obtain treatment data, the report was cross-referenced with GLC treatment records for the study period. A research assistant flagged for review any records that were missing treatment data that were needed to determine milestones of the cure cascade.

Analysis

The only demographic data available for all screened patients (anti-HCV positive and negative) during both study periods was year of birth. Additional data, including data on race/ethnicity and sex, were available for anti-HCV–positive persons. For this analysis, we classified data on race/ethnicity into 3 categories: black/African American, white, and other.

For period 2, we included patients in each stage of analysis on the basis of inclusion criteria (Figure 2). We excluded patients from analysis if they were diagnosed with HCV infection before the study (n = 4). Of persons newly identified as anti-HCV positive (n = 462), a research assistant conducted medical record review only for persons with detectable HCV RNA (n = 240), as was done for patients during period 1. Medical record review identified linkage to care (defined as any appointment during which HCV was addressed as a problem) and linkage to specialty care (defined as a completed appointment at GLC, an infectious disease clinic, or a gastroenterology clinic). To directly compare outcomes from period 2 with outcomes from period 1, we conducted a medical record review for treatment and cure status only for patients who were seen at GLC (n = 165). Treatment milestones were (1) initiated HCV treatment and (2) achieved SVR12 (sustained virologic response), defined as undetectable HCV RNA 12 weeks after completing therapy.

Figure 2.

Inclusion and exclusion criteria of patients who achieved hepatitis C cure cascade milestones at each stage of analysis, Grady Health System, Atlanta, Georgia, 2015-2016. Abbreviations: anti-HCV, hepatitis C antibody; GLC, Grady Liver Clinic; SVR12, sustained virologic response 12 weeks after completing therapy.

Statistical Analysis

We compared categorical values between period 1 and period 2 cohorts by using OpenEpi version 3.01.¹⁶ We used the Pearson χ² test of homogeneity, with P < .05 considered significant. We did not compare SVR12 success rates because sample sizes were insufficient for comparison. The Emory University Institutional Review Board reviewed the screening and linkage-to-care portions of the study and determined the study to be non–human subjects research. The Emory University Institutional Review Board and the Grady Research Oversight Committee approved the analysis of treatment outcomes.

Results

Demographic Characteristics

A total of 5274 patients were screened during period 1, with a mean of 2137 patients screened each year (Table 1). Of 5274 patients screened, 412 (7.8%) were anti-HCV positive. Of the 412 anti-HCV–positive patients, 245 (59.5%) were male, 386 (93.7%) were black/African American, and 22 (5.3%) were white.

Table 1.

Characteristics of patients screened for anti-HCV and for patients who tested anti-HCV positive during 2012-2014 (period 1) and 2015-2016 (period 2), Grady Health System, Atlanta, Georgia^a

Characteristic	Period 1 (2012-2014)	Period 2 (2015-2016)	P Valueb
Total no. of patients screened	5274	7137	—
Mean no. of patients screened per year	2137	7137	—
Among patients screened for anti-HCV, patients who were anti-HCV positive, no. (%)	412 (7.8)	462 (6.5)	.004
Among patients who were anti-HCV positive, no. (%)
Year of birth
Before 1945	0	7 (1.5)	.01
1945-1965	412 (100.0)	397 (85.9)	<.001
After 1965	0	58 (12.6)	<.001
Sex
Male	245 (59.5)	281 (60.8)	.70
Female	167 (40.5)	181 (39.2)	.70
Race/ethnicity
Black/African American	386 (93.7)	387 (83.8)	<.001
White	22 (5.3)	49 (10.6)	.01
Other	4 (1.0)	26 (5.6)	<.001

Abbreviation: anti-HCV, hepatitis C virus antibody.

^a Period 1 included the Grady Primary Care Center and occurred before the availability of all-oral direct-acting antiviral (DAA) therapies. Period 2 included the Grady Primary Care Center and neighborhood clinics and occurred after the introduction of all-oral DAA therapies.

^b Fisher Exact χ² test of homogeneity, with P < .05 considered significant.

^c See Table 2 for further analysis.

A total of 7137 patients were screened during period 2, of whom 462 (6.5%) were newly diagnosed anti-HCV positive (Table 1). Of the 462 anti-HCV–positive patients, 397 (85.9%) were baby boomers, 7 (1.5%) were born before 1945, and 58 (12.6%) were born after 1965. Most anti-HCV–positive patients were male (n = 281; 60.8%) and black/African American (n = 387; 83.8%).

Cure Cascade Outcomes

Of 412 patients in the period 1 anti-HCV–positive cohort, 381 (92.5%) received HCV RNA testing, 264 (69.3%) of whom had detectable HCV RNA (Table 2). Of those, 95.8% (253 of 264) were linked to any HCV care and 75.0% (198 of 264) were linked to specialty HCV care at GLC. Of patients linked to GLC, 21.2% (42 of 198) initiated antiviral therapy for HCV infection; 27 of 42 (64.3%) patients returned for SVR12 testing, 25 (92.6%) of whom achieved SVR12.

Table 2.

Number and percentage of patients who reached hepatitis C cure milestones during 2012-2014 (period 1) and 2015-2016 (period 2), Grady Health System, Atlanta, Georgia

Hepatitis C Cure Milestones	Eligible or Tested, No. (%)a		Percentage-Point Difference	P Valueb
	2012-2014 (n = 5274)	2015-2016 (n = 7137)
Anti-HCV positive	412 (7.8)	462 (6.5)	–1.3	.004
HCV RNA tested	381 of 412 (92.5)	405 of 462 (87.7)	–4.8	.02
HCV RNA positive	264 of 381 (69.3)	240 of 405 (59.3)	–10.0	.003
Linked to any HCV carec	253 of 264 (95.8)	229 of 240 (95.4)	–0.4	.83
Linked to Grady Liver Clinicd	198 of 264 (75.0)	165 of 240 (68.8)	–6.2	<.001
Initiated treatmente	42 of 198 (21.2)	105 of 164 (64.0)	42.8	<.001
Achieved SVR12f	25 of 27 (92.6)	21 of 22 (95.5)	2.9	NCg

Abbreviations: anti-HCV, hepatitis C antibody; HCV, hepatitis C virus; NC, not calculated; SVR12, sustained virologic response 12 weeks after completing treatment.

^a See Figure 2 for inclusion criteria for each milestone.

^b Using the Pearson χ² test of homogeneity, with P < .05 considered significant.

^c Attendance at a specialty or primary care appointment where HCV was discussed, as documented in clinician's note.

^d Attendance at a Grady Liver Clinic or other specialty clinic where HCV was the primary problem.

^e Denominator is the number of patients eligible for treatment (life expectancy >1 year) among those linked to the Grady Liver Clinic.

^f Denominator is the number of patients who returned for SVR12 testing.

^g Numbers were insufficient for analysis.

Of 462 anti-HCV–positive patients in the period 2 cohort, 405 (87.7%) received HCV RNA testing, 240 (59.3%) of whom had detectable HCV RNA (Table 2). Of those, 95.4% (229 of 240) were linked to any HCV care and 79.5% (183 of 229) were linked to HCV specialty care—165 at GLC, 15 at an infectious disease clinic, and 3 at a gastroenterology clinic. Of the 164 patients linked to care at GLC who were eligible for treatment (life expectancy >1 year), 105 (64.0%) initiated antiviral therapy for HCV infection. Of the 86 patients who had reached this treatment milestone at the time of analysis, 22 (51.2%) were tested for SVR12 and 21 (95.5%) had achieved SVR12.

The percentage of patients who tested anti-HCV positive (P = .004) and who were tested for HCV RNA (P = .01) was significantly higher during period 1 than during period 2. The percentage of patients with chronic HCV infection (detectable HCV RNA) was significantly higher among patients tested during period 1 than during period 2 (P = .003). Finally, the percentage of patients who initiated treatment was significantly higher during period 2 than during period 1 (P < .001); patients during period 2 were 3 times as likely to have initiated treatment as patients during period 1.

Discussion

To our knowledge, this analysis is the first to examine HCV linkage-to-care and treatment outcomes before and after the introduction of all-oral DAA therapies. As expected, the rate of screening increased from period 1 to period 2 with the inclusion of 5 additional community clinics during period 2. We hypothesized that the greatest difference between cohorts during period 1 and period 2 would be in the initiation of antiviral treatment because of dramatic improvements in HCV treatment from 2012 to 2016. Indeed, the rate of HCV antiviral therapy initiation tripled from period 1 (21.2%) to period 2 (63.6%). Several factors may account for this increased treatment uptake:

1. Fewer contraindications to DAA therapy compared with interferon-based therapy increased the pool of patients eligible for treatment and led to more referrals.

2. Regimens that are shorter and easier to tolerate prompted greater treatment uptake among patients.

3. Relaxed fibrosis-based treatment restrictions by payors after 2014 allowed more patients to access DAA therapies.

4. We expanded our capacity at GLC with greater staffing dedicated to treatment.

The average prevalence of anti-HCV–positive patients in both cohorts (7.1%) was more than double the reported national prevalence among baby boomers (3.3%).¹⁷ However, these rates were similar to those reported in other urban primary care sites, where reported prevalence is approximately 10%.¹⁸ Unexpectedly, the percentage of anti-HCV–positive patients was significantly higher in the period 1 cohort (7.8%) than in the period 2 cohort (6.5%). This difference may be related to implementation of CDC’s new baby-boomer screening recommendation by clinicians. The period 1 program was implemented in 2012, shortly after CDC’s recommendation was released. We hypothesize that during period 1, clinicians accustomed to risk factor–based screening may have been biased (out of habit) toward preferentially screening patients with traditional risk factors who also were born during 1945-1965. If a clinician was faced with screening 2 baby boomers, one with elevated liver enzymes and the other without elevated liver enzymes, that clinician might have been more likely to screen the patient with elevated liver enzymes. Therefore, the higher anti-HCV–positive prevalence during period 1 likely represents a hybrid between a pure risk factor–based prevalence (expected to be higher) and a pure baby-boomer prevalence (expected to be lower). The lower prevalence of anti-HCV positivity during period 2 is likely closer to the true population prevalence within this primarily black/African American population. As screening expands, observed prevalence tends to trend toward the true mean. Indeed, the prevalence of anti-HCV positivity in the period 2 cohort was similar to that previously reported among non-Hispanic black patients in federally qualified health centers in Philadelphia from 2012-2014 (6.0%).¹⁸ This suspected pattern of screening could also explain the significantly higher rate of chronic HCV infection in the period 1 cohort than in the period 2 cohort. Again, if physicians were subconsciously selecting higher-risk patients for screening within the birth cohort, patients with true chronic infections are more likely to be identified than patients with naturally cleared infection.

The percentage of patients tested for HCV RNA was significantly lower in the period 2 cohort (87.6%) than in the period 1 cohort (92.5%), although it was comparable to rates of confirmatory testing reported at other sites with similar screening programs.^18,19 The decrease was likely partly the result of the screening protocol’s expansion to the community clinics, which increased the number of patients screened. Although a higher absolute number of patients screened positive for anti-HCV during period 1 than during period 2, the number of patient navigation staff members, who were part of the linkage-to-care team and tasked with reminding patients to return for RNA testing, did not increase proportionately. Thus, it was more difficult for staff members to keep up with the increased load of reminder telephone calls. This increase in staff member workload may explain why linkage to care at GLC was slightly lower during period 2 than during period 1, although the difference was not significant. During both periods, reflex RNA testing for anti-HCV–positive samples was not available at GHS; thus, all anti-HCV–positive tests required a second laboratory visit for RNA testing, which was a labor-intensive process. Also, period 1 patients were followed for 2.5 years, whereas period 2 patients were followed for only 1 year. Completing milestones of the cure cascade, including HCV RNA testing, in a population with social and financial challenges can take several months. The ability to follow period 1 patients for a longer time increased the likelihood that they would eventually complete HCV RNA testing. One solution to mitigate barriers to HCV RNA testing is to implement reflex RNA testing for all anti-HCV–positive samples, as is current policy within Veterans Health Administration facilities.²⁰ GHS did ultimately implement reflex RNA testing in 2018.

Substantial drop-offs still occur along the HCV cure cascade among GLC patients. Although we have little control over the biological rates of virologic response, from a quality improvement perspective, we should be aiming for a 100% testing rate of virologic response during and after treatment. Our SVR12 rates among patients tested were high compared with rates in clinical trials.^21-23 Although our measured rates of SVR12 were >90%, we found that only half of the patients actually returned for SVR12 testing. This rate of loss to follow-up is higher than the rate reported at similar sites, indicating a need for increased patient outreach to return for confirmation of cure. One strategy that was found to be effective in a safety-net health system was to schedule patients for an office visit, as opposed to a laboratory-only visit, for an SVR12 check.²⁴ We are currently addressing this issue by adding an in-clinic patient navigator.

Limitations

This study had several limitations. First, this study was retrospective and required medical record reviews, which are subject to reviewer error and may have introduced selection and/or misclassification bias. Second, the periods for evaluating the screening programs were unequal, with 2.5 years for period 1 and 1 year for period 2. Given the length of time required for some patients to reach cure cascade milestones, this longer period may have resulted in improved outcomes for period 1 patients. Third, the demographic characteristics of the 2 cohorts, although similar, were significantly different. Period 1 had a higher percentage of black/African American patients than period 2, and black/African American patients are known to have lower response rates to interferon-based (non-DAA) therapy; as such, SVR12 rates may have been affected. Also, screening was expanded beyond the baby-boomer cohort to include a substantial proportion of younger patients during period 2 than during period 1. Younger patients may have had a different risk profile for HCV acquisition (eg, active injection drug use), which may have led to more spontaneous clearance of HCV infection. This could have contributed to the lower proportion of patients with detectable HCV RNA in period 2. Finally, we did not collect data on health insurance for the period 2 cohort, and the difference in health insurance status between patients in the 2 cohorts may have affected progression along the cure cascade. Further investigation into the relationships between race/ethnicity and/or health insurance status and treatment referrals and SVR rates may be warranted.

Conclusions

Although our large-scale screening initiatives coupled with a robust treatment program have been successful, barriers along the cure cascade remain and need to be addressed before elimination of HCV can be achieved. Treatment uptake appears to have increased dramatically in the DAA era. By addressing barriers to the cure cascade, including RNA testing rates, linkage to care, and treatment uptake, we can set our sights on elimination in the immediate future.