Improving Hepatitis C Identification: Technology Alone Is Not the Answer

Introduction

The introduction of the hepatitis C virus (HCV) birth cohort–based testing guidelines by the Centers for Disease Control and Prevention (CDC; Smith et al., 2012) and the U.S. Preventive Services Task Force (Moyer, 2013) represented a significant paradigm shift in the clinical identification of HCV. The rationale for these new recommendations is the disproportionate number of individuals born between 1945 and 1965 who have chronic HCV infection; indeed, approximately 75% of patients with HCV are members of this birth cohort (also known as the “baby boomer” generation; Smith et al., 2012). Identification of HCV infection using a risk-based strategy alone has been demonstrated to be ineffective. In a review of 110,223 patients with HCV, Mahajan, Liu, Klevens, and Holmberg (2013) reported that only 27% would have been identified using risk-based criteria, whereas 68% met birth cohort–based criteria.

These updated testing guidelines represent a significant paradigm shift in the clinical identification of HCV, as well as a critical public health mandate. Currently, advanced liver disease due to HCV is the main driver of liver transplants in the United States (Ward, 2013a, 2013b). In addition, mortality related to HCV has increased steadily; by 2007, deaths due to HCV outnumbered those due to human immunodeficiency virus, and in 2011, the mortality rates for HCV surpassed a combination of 60 other reportable infectious diseases (Ly et al., 2012; Ly, Hughes, Jiles, & Holmberg, 2016). Among the HCV-related deaths that occurred between 1999 and 2007, 73.4% were patients within the baby boomer generation (Davis, Alter, El-Serang, Poynard, & Jennings, 2010). Recently, in a retrospective study of more than 1 million Medicare beneficiaries, Sayiner et al. (2016) demonstrated that the presence of HCV is independently associated with increased mortality and resource utilization among baby boomers.

Highly effective direct-acting antiviral therapy regimens have been approved for the treatment of chronic HCV and are recommended by clinical practice guidelines from the American Association for the Study of Liver Diseases in collaboration with the Infectious Diseases Society of America (American Association for the Study of Liver Diseases/Infectious Diseases Society of America/IAS-USA, 2016). As a result of these new treatments, it is now possible for virtually every patient with HCV infection to achieve a sustained viral response, which is a functional cure for HCV, and leading national and international health policy organizations have published goals for HCV disease control and eradication of the virus (National Academies of Sciences, Engineering, and Medicine, 2016; World Health Organization, 2016). However, studies suggest that as many as 50% of persons infected with chronic HCV are unaware of their disease (Denniston, Klevens, McQuillan, & Jiles, 2012; Udompap, Mannalithara, Heo, Kim, & Kim, 2016). Modeling data show that without improvement in the identification of patients with chronic HCV, as well as their engagement in care, the effectiveness of current HCV therapy may be reduced by as much as 75%, and the proportion of chronically infected patients who achieve sustained viral response may not be substantially better than that observed in the interferon era (Linas et al., 2014).

The identification of patients with HCV falls within the purview of already overtaxed primary care clinicians, who face numerous HCV-related barriers (Smith et al., 2012; U.S. Department of Health & Human Services, 2011). Experts have stated that a primary care physician would require 18 hours per day to provide all recommended preventive and chronic care services to a typical patient panel (Bodenheimer, 2008). Specific to the timely identification of HCV, numerous potential barriers in this setting include challenges related to patient willingness for testing, barriers related to clinician knowledge and acceptance of the need for testing, and systems-related obstacles (U.S. Department of Health & Human Services, 2011; van der Meer et al., 2014). The lack of awareness about viral hepatitis has resulted in inadequate health care resource allocation for HCV testing and treatment (Ward, Lok, Thomas, El-Serag, & Kim, 2012). In a review of testing patterns among 8,981 patients across three clinical sites, Southern et al. (2014) demonstrated substantial provider nonadherence with testing protocols for HCV; overall protocol adherence was 36.1% (range 0% to 92.4%). Interestingly, adherence to testing guidelines decreased over the 15 weeks of this study (59.1% to 13.7%), and despite educational programs focused on HCV testing, the authors described large variation in testing rates among physicians working in the same system, suggesting a need to address physician attitudinal barriers.

One strategy for improving care in alignment with contemporary evidence and clinical practice recommendations is the integration of reminders into the electronic health record (EHR; Jamtvedt, Young, Kristoffersen, O’Brien, & Oxman, 2006). However, data illustrate that large volumes of EHR reminders are adding to the time burden for primary care clinicians (Murphy et al., 2016). Murphy et al. (2016) report that primary care clinicians spend an average of 66.8 minutes per day processing notifications, a number that is significantly higher than the burden on specialists.

Studies demonstrate that local factors such as clinician motivation, experience, and leadership support can have a substantial effect on the success of an initiative designed to change clinical practice and that single, one-off activities often do not result in improvement (Kaplan et al., 2010; McCormack et al., 2002). Instead, interventions must be iterative, multifaceted, and responsive to both local issues and unexpected obstacles. Therefore, we designed a quality improvement (QI) program with multiple plan-do-study-act cycles to evaluate (and modify, as necessary) the effect of a series of interventions on HCV testing rates in alignment with birth cohort recommendations and the linkage of patients with HCV-positive tests to specialty care. The plan-do-study-act cycle promotes repetitive measurement over time to understand natural variation in dynamic practice settings and assess the effect of interventions on the outcome of interest (Damschroder et al., 2009; Powell, Rushmer, & Davies, 2009). This initiative was designed to employ multiple cycles of audit and feedback in collaboration with education and EHR reminders. The goal was to improve clinician knowledge and competence related to whom, when, and how to test for HCV; knowledge about appropriate engagement of specialty care and provision of comprehensive care for patients with HCV; and clinical performance related to HCV testing and referrals in alignment with birth cohort–based guidelines.

Materials and Method

The Virginia Mason health care system places substantial emphasis on continual improvement of care quality and has a mission to become the Quality Leader in health care. As HCV testing guidelines evolved, clinicians and continuing education leaders determined that a QI initiative aimed at assessing and improving HCV testing practices aligned with the mission and goals of the organization. Our study evaluated patterns of HCV testing in alignment with birth cohort recommendations and referral to specialty care within seven primary care sites in the Virginia Mason health care system, located in the greater Seattle area of Washington State, between August 1, 2014, and September 14, 2015. Clinicians in three primary care sites participated in a series of educational interventions (described in more detail below). Clinicians in four control sites were not offered similar educational activities or consistent review of HCV testing rates during our study (Figure 1). The Virginia Mason Institutional Review Board reviewed and approved this study.

OPEN IN VIEWER

Figure 1

Initiative Design

Each participating clinic designated a physician to be the project champion at that site. This champion was responsible for leading internal discussions about the initiative at his or her location, disseminating clinician surveys (n = 3) and educational materials (n = 7), sharing reports detailing audit results following each data review (n = 7), and communicating ongoing educational needs, questions, and comments to the project coordinators. We used electronic surveys (distributed and completed before, during, and after the initiative) to evaluate clinician knowledge, attitudes, and barriers related to testing for HCV in the primary clinic; specifically, the survey questions assessed clinician knowledge of current testing guidelines, clinician self-reported barriers to guideline adherence, resources necessary for improvement efforts, and clinician knowledge of current HCV treatment efficacy. We used a Med-IQ proprietary survey tool to administer the surveys. Survey respondents were offered a nominal incentive ($50) for participation, and clinicians had the opportunity to decline the stipend and answer anonymously. We developed a Continuing Medical Education/Continuing Education–certified digital publication to support clinician education in alignment with program objectives; this activity focused on the rationale and implementation of birth-cohort–based testing in the primary care setting and early steps for patients with newly identified HCV infection. Clinician participants also received six shorter publications (one to two pages each), which were designed to be practical educational tools that focused on key topics related to identifying and managing HCV in the primary care setting. Topics covered in these pieces included the rationale for birth cohort–based testing, the diagnostic process, practicalities of answering patient questions, available therapies, and patient–clinician communication about HCV risk, testing, and treatment in the primary care setting. Physician champions at each location were eligible for Maintenance of Certification IV credit for their efforts.

To determine the HCV testing practices of participating clinicians, we reviewed a total of 73,685 patient charts in the Virginia Mason Cerner EHR system for the primary objective of this initiative. Eligible patients were those born between January 1, 1945, and December 31, 1965, who had a visit to one of the seven sites during the abstraction time frames. For the baseline data, we evaluated testing rates among all eligible patients with visits between August 1, 2014, and October 31, 2014. Then, from March to September 2015, we evaluated changes in testing rates and linkage-to-care patterns for the prior 30 days of visits and shared those data with clinician participants in the intervention arm of this initiative. Testing rates were calculated as tested number/total eligible for each time frame. We used two-sample tests of proportions to evaluate the level of statistical significance of the differences from baseline to the follow-up periods in the percentages of eligible individuals who tested positive for HCV as well as the percentages of HCV-infected individuals who were subsequently linked to care within the three intervention groups. For evaluating the differences between the intervention and control groups at each time point, we used Z tests for two population proportions. Statistical tests were conducted using Stata Version 13.1 (StataCorp, 2013).

Results

Intervention Group

Among the 51 eligible clinicians in the 3 intervention clinics, 30 answered the baseline knowledge, barriers, and attitudes survey (58.8% response rate). The average time in clinical practice was 16 years, and clinicians reported seeing an average of 2.7 patients with HCV each month (Table 1). Clinicians reported a high level of baseline confidence regarding their knowledge of guideline recommendations for HCV testing and reported that their usual practice patterns related to HCV testing were in alignment with current practice guidelines. However, actual clinical performance as shown by HCV antibody tests recorded in the EHR demonstrated low baseline testing rates among patients born between 1945 and 1965 (Figure 2). In the survey following distribution of the baseline data, providers were asked to rate how closely their clinical practice data aligned with their perceptions of their clinical practice on a 7-point Likert-type scale, with 7 indicating complete alignment between reported testing rates and perception; 61.6% of responding physicians (n = 13) choose a score of 4 or lower. In the same survey, when asked to describe barriers to HCV testing in baby boomers, physicians noted health system–related barriers (time and cost of screening) and patient-related barriers (fear of stigma, lack of appreciation of risk, and belief that they had been vaccinated for HCV).

OPEN IN VIEWER

Table 1

Demographics for Participating Clinicians

Demographic	Participating Clinician Responses (n = 30)
Average no. of years in practice	16.1
Specialty
Family medicine	13.3%
Internal medicine	86.7%
Geography
Rural	3.3%
Suburban	33.3%
Urban	63.3%
% Patients who are baby boomers
<25%	3.3%
25%-50%	76.7%
51%-75%	16.7%
>75%	3.3%
Average no. of HCV-positive patients clinicians report caring for in 1 month	2.7

NOTE: HCV = hepatitis C virus.

OPEN IN VIEWER

Figure 2

Clinician Perceptions of HCV Testing Patterns Compared With Actual Practice

Overall, within the intervention group, significant changes in HCV testing among baby boomer patients were seen at each data review relative to baseline testing rates (Figure 3). Improvements were evident in each of the three intervention clinics. At baseline, testing rates were 7.6% (390/5,102) for Site 1, 5.0% (255/5,098) for Site 2, and 5.3% (129/2,427) for Site 3; at the conclusion of the intervention period, rates were 20.2% (318/1,574), 13.6% (203/1,495), and 19.6% (142/726), respectively. At baseline, 100% (33/33) of patients with positive HCV antibody test results had records of referral to specialty care, and no significant changes were seen during the course of the intervention (data not shown). Among the 13 clinicians who responded to the postintervention survey, most (91.7%) reported changing their clinical practice as a result of participating in the QI initiative.

OPEN IN VIEWER

Figure 3

Changes in HCV Testing Within Baby Boomers Over Intervention Period

Discussion

Many authors have demonstrated the success of using initiatives to facilitate testing processes that more closely align with current guidelines. Between 2012 and 2014, under the direction of the U.S. Senate, the CDC funded 25 programs in the United States to improve HCV testing and linkage to care for patients with a positive HCV antibody test result (Ward, 2016); 14 of these programs focused on birth cohort–based testing. In these birth cohort–focused programs, HCV testing was conducted in 24,966 baby boomers, 11.6% of whom tested positive for HCV (Patel, Vellozzi, & Smith, 2016). Importantly, among patients with positive HCV antibody test results, only 31.6% reported common HCV risk factors (e.g., past or present injection drug use, HIV-positive status), supporting the need for birth cohort–based testing (Patel et al., 2016). Similarly, in a 27-month QI initiative designed to improve testing rates in alignment with birth cohort–based recommendations, Gemelas et al. (2016) showed greater than 10-fold improvements (5% in 2012 to 76% in 2014). Importantly, in a postinitiative follow-up survey, 88% of survey respondents (n = 20) positively rated the effect of screening on HCV-related communication with their patients (Gemelas et al., 2016). However, a recent review of data from the 2013 and 2015 National Health Interview Survey reports only a small increase in the percentage of baby boomers who have been tested for HCV (12.3% in 2013 to 13.8% in 2015; p = .013; Jemal & Fedewa, 2017). From the National Health Interview Survey data, Jemal and Fedewa (2017) extrapolate that among the 76.2 million baby boomers, only 10.5 million have been tested for HCV as recommended. These data demonstrate a need for continued emphasis on birth cohort–based testing.

The incorporation of reminders for routine testing may help clinicians better align their practice with evidence-based guidelines. In fact, in the study reported by Gemelas et al. (2016), 100% of clinicians responding to the postinitiative survey indicated that the implementation of an EHR reminder was helpful for HCV screening.

However, considering the already burdensome volume of reminders that primary care clinicians are exposed to on a daily basis, adding reminders and/or relying on reminders to change clinical practice is unlikely to be successful (Murphy et al., 2016). Results from our initiative support the value of integrated education and continuous audit/feedback cycles for changing clinical processes, in addition to technology-based reminders. Finally, in our study, the discordance between clinician perception of current testing and actual testing practices as evaluated through chart review was striking. To our knowledge, ours is the first study to report such data. It is our perspective that this gap is a strong indication of the need for data-sharing processes within initiatives designed to make clinical practice changes.