Abstract
Quality gaps exist in the hepatitis C virus (HCV) care process from diagnosis to cure. To better understand current gaps and to identify targets for quality improvement, we constructed an HCV care cascade in a patient-centered medical home (PCMH) with an emphasis on the specialty referral process. We performed a retrospective study of HCV-infected patients in a PCMH using electronic health record (EPIC) data. Patients with a first positive HCV RNA between 2012 and 2019 were included. With an adaptation to analyze linkage to specialty care, we created an HCV care cascade that included the following: (1) a positive HCV RNA, (2) referral to a specialty provider, (3) a scheduled specialty appointment, (4) attendance at a specialty visit, (5) prescription for HCV therapy, and (6) evidence of sustained virological response (SVR). Patient and referring clinician characteristics were analyzed at each step of the care pathway, and the proportion of patients completing each step was calculated. Of the 256 HCV RNA-positive patients, 229 (89.5%) received a specialty referral; 215 (84.0%) had an appointment scheduled; 178 (69.5%) attended the specialty appointment; 116 (45.3%) were prescribed antiviral therapy; and 87 (34.1%) had documented SVR during the study period. Of the 178 patients attending a specialty visit, 62 (34.8%) did not receive a prescription, and the barrier most often noted was the desire for further workup (40.3%). Gaps occur at all stages of the HCV care continuum, with drop-offs in care occurring both before and after linkage to specialty care.
Significance of this study
What is already known about this subject?
An estimated 2.4 million persons in the USA are chronically infected with hepatitis C virus (HCV), and many remain untreated despite the efficacy of direct-acting antiviral (DAA) therapy.
Gaps in the hepatitis C care cascade have historically appeared in HCV antibody testing, HCV RNA confirmation, and receipt of DAA therapy.
Little is known about the gaps in the specialty referral process linking patients to treating providers.
What are the new findings?
Racial differences and primary care provider (PCP) experience were associated with differences in specialty provider visit attendance.
A desire for further workup was identified most often in specialty care notes where DAA prescriptions were not provided at the attended specialty visit.
No patient or provider factors were associated with significant differences in achieving sustained virological response for those patients undergoing DAA therapy.
How might these results change the focus of research or clinical practice?
The gap between specialty visit scheduling and specialty visit attendance demands further exploration, with attention paid to transportation barriers and the communication of the scheduled specialty visit, as current practices may promote disparities in care.
Addressing the barriers related to further workup may fall on patient-centered medical homes to team with specialty providers to streamline additional HCV-related evaluation and to ensure its completion at the time of the specialty visit.
PCPs with access to specialty care will need to decide whether to optimize the specialty referral process for HCV care or to take on DAA therapy on their own.
Introduction
An estimated 2.4 million persons in the USA are chronically infected with hepatitis C virus (HCV) and new cases increased every year from 2013 to 2017.1 2 Direct-acting antiviral (DAA) therapy has drastically transformed care for patients chronically infected with HCV, offering the opportunity for cure.1 3 4 DAA therapy leads to sustained HCV eradication in over 90% of patients (sustained virological response (SVR)), often with fewer than 12 weeks of treatment and fewer side effects than traditional interferon-based therapy.5–7 Despite the efficacy of DAA, many patients chronically infected with HCV remain untreated.8
The patient’s journey from HCV diagnosis to treatment and cure is complex. The guideline-based management pathway, the HCV care cascade, includes antibody (Ab) screening and confirmatory RNA testing, referral and establishment with a treating provider, genotyping and hepatic staging as indicated, prescription of treatment, and follow-up.5 Each of these steps allows for disconnects in care and varies based on patient, provider, and system factors.8 9 Previously published studies identified interruptions at every step of the process, with gaps in HCV Ab testing, HCV RNA confirmation, and receipt of DAA therapy standing out as the most glaring opportunities for improvement.8–15 Healthcare systems and professional society guidelines have addressed some of these gaps by heightening clinician awareness for those patients in need of testing, automating (or ‘reflexing’) HCV RNA testing in those with positive HCV Ab results, and loosening DAA restrictions based on fibrosis stage or recent substance abuse.5 16–21 However, few studies have focused on the steps between diagnosis and treatment, and little is known about the gaps in the specialty referral process linking patients to treating providers.20 22–24
In academic medical centers and multispecialty healthcare systems, where gastroenterology (GI) and infectious disease (ID) specialists are available, the process from primary care HCV diagnosis to specialty treatment deserves further evaluation. As improvements in HCV care continue to progress, primary care providers (PCPs) and patient-centered medical homes (PCMHs) with system access to specialty care must decide whether to treat patients chronically infected in primary care practice or to refer to a specialist. Previous studies show DAA treatment by non-specialty providers is equally effective to that provided by specialists, but the necessary resources for a successful program remain a concern.25–27 Intensive support, both from infrastructure and pharmaceutical resources (ie, dedicated assistance in the medication prior to the authorization process), is likely necessary for primary care-based treatment efforts to succeed.28 While PCPs in remote locations may opt to treat out of necessity, those with access to specialty care may need to understand specialty referral practices prior to deciding to initiate primary care treatment.
Gaps in linking primary care patients with HCV to specialty care for prescription of appropriate DAA are opportunities for PCMHs to improve HCV care delivery. This study evaluates the completion of HCV care cascade milestones in RNA-confirmed patients in PCMH, with an emphasis on the specialty referral process, and evaluates the association of patient and provider characteristics with completion of each stage of the cascade.
Materials and methods
We performed a retrospective study of electronic health record (EHR, EPIC) data from an academic internal medicine PCMH to identify patients with RNA-confirmed HCV. We created an HCV care cascade modeled after peer-reviewed studies but adapted to specifically analyze linkage to specialty care, including referral to an HCV specialist, specialty appointment scheduling, and specialty appointment attendance.8–10 13 14 19 22 29 Patient and referring provider characteristics were analyzed at each step of the HCV pathway.
Study sample
All patients seen in the academic Internal Medicine PCMH at the Medical University of South Carolina (MUSC) between January 1, 2007, and January 31, 2019, were evaluated. The practice delivers care to a diverse (49.1% non-white), adult (mean age 58.6 years) population with chronic and complex medical problems and conducts approximately 38,00 patient visits yearly. The clinic is located in an urban setting and houses the faculty and resident physicians at a single site adjacent to the tertiary care hospital at MUSC. The standard of care at MUSC calls for referral of all patients identified with HCV to GI or ID for evaluation and treatment. The ID clinic is located within the same building as the PCMH, and the GI clinic is located across the medical campus (<1 mile). All referrals are placed via computerized physician order entry, and referral scheduling is done through a university-wide central scheduling department. Communication of scheduling occurs via telephone. All HCV results were obtained through a computer-based EHR data query and the HCV RNA results were quantitative and defined as ‘positive’ when HCV RNA was >12 IU/mL (the reference range in the MUSC lab). From January 2007 to January 2019, the clinic delivered care to 36,841 unique patients, of whom 10,061 (27.3%) had results for at least one HCV test. Of these patients, 9896 had an HCV Ab result, and 743 (7.5% of those with Ab testing) possessed a positive Ab test (qualitative test result). Six hundred thirty-nine (86.0%) of the HCV Ab-positive patients had RNA testing and 165 patients received an initial HCV RNA test (no identified HCV Ab precursor). Of the 804 total patients receiving HCV RNA testing, 560 (69.7%) had positive HCV RNA results. A sample of 256 (256/560) unique patients with a first known positive HCV RNA test in our institution between January 1, 2012, and January 31, 2019, was constructed with an emphasis on care outcomes following a positive HCV RNA result. The start date was chosen to take advantage of the comprehensive EHR (EPIC) implemented in 2012, enhancing the fidelity of referral and care coordination data. Improving HCV Ab testing and HCV RNA follow-up is a critical need, and efforts to address these issues in this setting are ongoing.15 30
Outcome measures
Completion of each step of the HCV care cascade served as an outcome of interest. HCV care cascade steps included (1) a positive HCV RNA, (2) referral to a treating specialty provider, (3) a scheduled appointment with a specialty provider, (4) patient attendance at a specialty provider visit, (5) prescription provided for HCV therapy, and (6) evidence of SVR (a negative HCV RNA test 3–6 months after the end of therapy). Outcomes were assessed via chart review using an abstraction tool and were conducted by two investigators (DB and AS). Chart review was conducted to identify the outcomes (DB) and to confirm the administrative HCV RNA data used to create the sample. The chart review for barriers to prescription therapy (AS) then confirmed the specialty referral attendance and the prescription outcomes from the previous reviewer. Data were collected during the period of January 1, 2012–July 31, 2019, to allow 6 months of follow-up from the last positive RNA test included in the study (January 31, 2019).
Data collection and other variables
Administrative and chart review data were collected for each included patient following their first known positive HCV RNA test in our system. Administrative level data came from Medical University Hospital Authority Enterprise and EPIC (EPIC Systems Corporation, Wisconsin) Clarity databases, and included structured demographic, laboratory, and comorbidity data. Age was determined at the time of the first known positive HCV RNA test and was analyzed as a continuous variable. Categorical variables included gender (male/female), race (black/white/other), marital status (married/single), payor (Medicare/Medicaid/private insurance/uninsured), and distance lived from the clinic (<50 miles/>50 miles). The distance reflected the number of miles between the center of the patient’s zip code and the medical campus. Race was coded as black/non-black in the univariate and regression analyses, owing to the small number of non-black, non-white patients identified in the administrated data (n=3). Comorbidities included alcohol abuse, drug abuse, HIV infection, and depression, and were identified by International Classification of Diseases, 9th and 1oth Revisions (ICD-9/10) codes (at any time during the study period) using an Elixhauser coding algorithm, and chosen based on clinical considerations historically necessary to address prior to HCV treatment (online supplemental table S1).29 31 Since some patients in our sample received a positive HCV RNA result prior to the approval of DAA therapy for HCV, a ‘pre-DAA’ categorical variable was created to reflect this timing.32 33 Patients with an HCV RNA in 2012 or 2013 were designated as having a pre-DAA diagnosis.
Supplementary data
Data from patient charts were abstracted, specifically to determine evidence of specialty referral orders to GI (hepatology included) or ID, scheduled specialty visits, attended encounters with specialists, the presence of a prescription for DAA therapy, and evidence of SVR. For patients with multiple referrals, evidence of scheduling or attendance at any future specialty appointment was used for the completion of care milestones. For those patients referred, information regarding the ordering provider (at the time of referral) was collected and categorized by the specialty of the clinician (specialist vs primary care), level of training (attending physician vs trainee (resident or fellow)), and relationship to the academic medical system (internal or external (for patients previously engaged with a non-PCMH PCP)).
Evidence of prescription for antiviral therapy (via chart review of all active and inactive medications) at any time following referral to specialty care was documented. For those patients attending a specialty visit but not receiving a prescription for antiviral therapy, specialist visit notes were reviewed to ascertain the reason a prescription was not initiated. Barriers were categorized as alcohol or drug abuse, compromised health status, administrative issues, a desire for further workup, or not clearly addressed. Categorization required specific mention in the specialty care note of current drug or alcohol use; severe illness, poor functional status, or severe mental health limitations (compromised health status); plans to begin paperwork, but not placing a prescription order (administrative); explicit statements desiring HCV genotype testing, imaging, repeat HCV RNA, or other laboratory data before proceeding to treatment (desire for further workup); or no reason for not prescribing provided (not clearly addressed).
Analysis
The proportion of HCV RNA-positive patients completing each step of the HCV care cascade was calculated. Data are displayed as proportions with the 256 patients entering the cascade as the denominator.
Univariate analysis of independent, categorical variables was performed using χ2 and Fisher’s exact tests. Univariate analysis of independent, continuous variables was performed using two sample t-tests. Multivariable logistic regression models (five) for the dependent binary variable of successful completion of each step of the care cascade were developed. The predictor variables included demographic and clinical variables as described previously. Characteristics of the referring provider were not included in the model for the first step of the cascade, with the outcome being whether the patient received a specialty referral. Variables were selected for the model based on clinical relevance and findings in previously published peer-reviewed studies.14 22 34 Model parameters were estimated using maximum likelihood estimation. Variables in the models were tested for multicollinearity. SAS V.9.4 was used for the statistical analysis.
Results
The 256 HCV RNA-positive patients had a mean age of 56 years at the time of diagnosis; 68.0% were male; 54.7% were black; 77.3% were single; 95.3% were insured (50% Medicare, 29% Medicaid, and 16% private), and 89.5% lived within 50 miles of the clinic (table 1). Burdens of comorbidity included 43.0% of patients with a history of alcohol abuse, 48.8% with a history of drug abuse, 40.2% with a history of depression, and 4.3% with HIV coinfection. The majority of the first known HCV RNA test results (85.9%) came after the year 2014.
Of the HCV RNA-positive patients, 229 (89.5%) had a specialty referral to a treating provider; 215 (84.0%) had a specialty appointment scheduled; and 178 (69.5%) attended a specialty appointment with GI or ID (figure 1). Of the 256 patients, 116 (45.3%) were prescribed antiviral therapy, and 87 (34.0%) had documented SVR during the study period.
In the univariate analysis, the mean age of patients with a scheduled specialty visit exceeded the mean age of those without (56.3 vs 50.4 years, p=0.04), but no other significant differences appeared in the proportion of patients referred and having a specialty visit scheduled by the patient and provider variables studied (table 2). A significantly higher proportion of patients referred by an attending physician completed the specialty visit as compared with trainee providers (93.9% vs 77.8%, p<0.01). Significantly lower proportions of patients with histories of drug (54.9% vs 72.9%, p=0.02) and alcohol abuse (56.6% vs 71.6%, p=0.04) who attended a specialty visit received a prescription for antiviral therapy, as did only 14.3% of uninsured patients (p=0.02). Higher proportions of black (79.0% vs 63.0% for all non-black patients, p=0.06) and local (<50 miles, 75.7% vs 38.5%, p<0.01) patients prescribed medication achieved SVR.
The five logistic regression models developed for each step of the care cascade revealed no significant associations between the predictor variables studied and referral placement or specialty visit scheduling (table 3). Black patients had significantly lower odds (OR 0.33, 95% CI 0.13 to 0.89) of attending a specialty visit compared with non-black patients when controlling for the other variables. Patients referred by attending physicians as compared with trainee physicians had significantly higher odds of attending their specialty visits (OR 3.92, 95% CI 1.14 to 13.5) when controlling for other factors. No significant associations appeared for the outcomes of receiving a prescription for DAA medication or for achieving SVR.
Of the 178 patients who attended a specialty visit, 62 (34.8%) did not receive a prescription for HCV treatment. The barrier most often noted as a deterrent to therapy was the desire for further workup (40.3%, table 4). At least 49 (19.1%) of the 256 patients died during the study period. All analyses were also performed with the deceased patients removed, with no significant changes to the results.
Discussion
Our study identified gaps in the HCV care cascade for PCMH patients with RNA-confirmed chronic HCV infection. Only 69% of the patients studied were successfully linked to a specialty provider for HCV treatment; 45% underwent DAA therapy; and 37% had documented SVR, with findings consistent with previous work.8 10 14 19 23
To better understand the role of the specialty referral process in the care delivery gaps, we used an adapted care cascade design to evaluate each step of the referral process, including (1) a care provider ordering a referral to specialty care, (2) the health system successfully scheduling a specialty appointment, and (3) the patient attending a specialty visit. Each component of the referral process contributed to interrupted care, with 89% of patients receiving a referral, 83% getting scheduled, and only 69% attending the specialty appointment. No significant differences in patient or provider characteristics appeared for the referral ordering or scheduling care steps. However, patients had lower odds of attending a scheduled specialty visit if they were identified as black (OR 0.33, 95% CI 0.13 to 0.89) or referred by trainees (non-attending physician, as attending physician OR 3.92, 95% CI 1.13 to 13.5). The racial disparity in specialty care linkage parallels that reported in HIV care, but other studies focused on racial disparities in linkage to HCV care have yielded mixed results.22 24 35 Historically, racial disparities in the receipt of DAA therapy existed for black Medicare patients, though this gap had closed by 2016.36 Despite the lower odds of attending a specialty visit for black patients in this study, no racial differences appeared in those patients receiving a DAA prescription or achieving SVR. This gap between specialty visit scheduling and specialty visit attendance demands further exploration, with attention paid to transportation barriers and the communication of the scheduled specialty visit, as current practices may promote disparities in care. Further work might also focus on the patient experience of linkage to HCV care, particularly relating to cultural biases and stigma, as these remain significant barriers to pursuing care in Black patients with HIV.37 38 The association between trainee referral and lower odds of specialty visit attendance reflects previous work revealing a lower likelihood of attaining chronic disease outcomes for patients cared for by resident physicians when compared with patients receiving care from attending physicians39 Trainees generally have less experience navigating healthcare systems, which may contribute to fewer patients successfully linked to specialty care.39 Continuity in patient care may contribute to this difference, but this has been difficult to measure. There may also be socioeconomic factors that were not captured. Efforts to standardize or improve specialty referral tracking and follow-up in a PCMH with attending and trainee providers may help to address this gap.
Another significant gap in our cascade occurs at the point of receiving a DAA prescription, a finding consistent with other published studies, with only 45% (and only 65% of those attending a specialty visit) of patients prescribed therapy.8–14 The most frequently noted reason for not prescribing was the desire for further workup (40%), followed by current alcohol or substance disorders (18%). This finding deviates from other previously published studies that cite drug use and significant mental health disorders as the main barriers to DAA.13 Hopefully, substance use-related gaps in DAA therapy close over time as treatment paradigms become more universal and less restricted, particularly to include patients using intravenous drugs.5 27 Addressing the barriers related to further workup may fall on PCMHs to team with specialty providers to streamline additional HCV-related evaluation and to ensure its completion at the time of the specialty visit. This gap may close over time as well, as current guidelines focus primarily on imaging and non-invasive liver fibrosis scoring systems such as FIB-4.29
In our institution, HCV treatment has historically been provided by specialists (GI or ID). As 31.5% of patients in our study were lost to follow-up prior to attending a subspecialty appointment, our findings raise the possibility that gaps in HCV care directly resulting from the referral process could be eliminated if portions of HCV workup and/or treatment were provided by PCPs. Further workup such as hepatic staging (the most commonly (40%) cited reason for not prescribing DAA therapy) may be more easily attained by the PCP, especially since non-invasive staging using lab-based scoring systems like FIB-4 is both simple and recommended by current HCV treatment guidelines.29 40 Such an approach would take advantage of the patient’s existing care engagement with the PCP and enable the patient to receive a prescription for DAAT at the first specialty visit. With sufficient resources, some PCPs can prescribe DAAT. While the PCP as prescriber approach has succeeded in some cases, this has not been a consistent finding.25 26 28 Before prescribing, PCMHs would need to consider the degree of practice engagement, the current availability of pharmacy support, and comfort with ongoing surveillance for HCV-related complications. On this last point, recent work suggests that patients with advanced fibrosis at the time of treatment remain at risk of hepatocellular carcinoma despite successful therapy and should undergo routine surveillance.41 Another possible model includes PCP as prescriber for DAAT, followed by referral to specialty care for HCV-related complications. PCMHs will need to weigh the resources at their disposal against the availability and accessibility of a specialty provider to determine the extent of HCV workup and care they are able to provide.
Limitations in this study include its focus on the HCV care cascade following a positive RNA result. As our study sample selection shows, improving HCV Ab testing and HCV RNA follow-up is a critical need, and efforts to address these issues in this setting are ongoing.15 30 Additionally, this study does not include liver fibrosis staging, a factor that may influence therapy decisions (although all patients should receive treatment regardless of liver stage). We only have access to liver biopsy results on 41 of these HCV RNA-positive patients, and even fewer liver elastography results in our PCMH dataset. We are currently evaluating FIB-4 scores in primary care, but have confronted challenges in FIB-4 interpretation due to acute illness, issues with missing data, and variable intervals. Non-invasive fibrosis risk assessment is a target for future work. Also, this study comes from a single site, with a population that is not representative of all settings (predominantly urban, publicly funded) and one that is already engaged in primary care. However, the diversity and complexity of the PCMH resembles many large, urban general internal medicine practices nationally, which may contribute to generalizability. Not all primary care practices have access to specialty providers in ID and GI, which may limit the impact of this work’s focus. However, many of the relationships explored in this work may provide more evidence in barriers to HCV care, and gaps in the specialty referral process. As in all fields, practice patterns also changed over this time period, as new guidelines and therapies emerged, and the contribution of these factors is difficult to quantify. Missing data in this EHR-based study is a concern, specifically limiting the ability to evaluate patients with laboratory findings outside of the institution. Patients with long-standing HCV were likely under-recognized (since the first known HCV RNA was used), as were those who may have achieved SVR but did not have confirmatory laboratory testing in the system (a particular concern in those patients residing far from the clinic). The study also relies on ICD-9/10 medical coding for the comorbidities of alcohol abuse, drug abuse, and depression, codes that have historically been imprecise.42 43 These conditions can resolve over time, and unfortunately, our data are unable to capture the dynamic nature of these diagnoses. The relationships between these variables and the results should be interpreted with caution. Over time, deaths and patients lost to follow-up during the study period may also have confounded the findings, although when analyses were conducted with deceased patients removed, results were unchanged.
In summary, this study suggests that each step in the HCV care cascade represents a target for improvement. Linkage to specialty care represents a large gap in the cascade, for which black patients and those referred by trainees may be disproportionately affected. More integrated care at each step, with knowledge of system-specific and population-specific barriers, may increase linkage to care with treating providers (both primary and specialty) and resolve chronic HCV infection.
Footnotes
Contributors All authors made significant contributions to the conception of the work. DB, AS, JZ, and JM were responsible for the acquisition of data, and all authors contributed to the analysis and interpretation of data for the work. DB and AS drafted the original manuscript, but all authors significantly contributed to critical revisions and intellectual content. All authors gave final approval of the version to be published, and all authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding Dr Schreiner is supported by a career development award (K23) from the National Institute of Health and National Institute of Diabetes and Digestive and Kidney Diseases (1K23DK118200 (principal investigator: AS).
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval This study was approved by the institutional review board at the Medical University of South Carolina.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. Data for this study are deidentified, securely stored in an encrypted server, and used as permitted by the institutional review board at the Medical University of South Carolina.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.