What Hepatitis C Taught Us About the Clinical Record

In 2002, my colleagues and I reported that peginterferon plus ribavirin cured a little more than half of patients with chronic hepatitis C, and for most of the following decade that regimen remained the standard of care. Our trial enrolled patients who were healthy enough to complete 48 weeks of a challenging treatment regimen, so it answered the efficacy question cleanly but left the question of effectiveness largely to the clinic. The real-world experience was a different story, because many of the patients I cared for at UNC had more advanced liver disease or contraindications to therapy that would have excluded them from the trial; they were patients in need of treatment for whom data were lacking.

Developing the first infrastructure for real-world hepatology evidence

I co-founded HCV-TARGET at the end of 2011, as telaprevir and boceprevir were entering practice, because the phase 3 programs could not tell us how these drugs would perform in diverse groups of patients, particularly those at the extremes of disease severity and those underrepresented in the trials. We enrolled patients prospectively at academic and community centers across the United States, and later in Canada, Germany, and Israel. In a novel twist, we did away with site-completed case report forms and brought the medical records of consented patients to a central coordinating center. There, trained abstractors captured each event against common definitions, and independent monitors reviewed the entries for completeness and accuracy. That design reduced the burden on participating sites and, more importantly, removed the variability that comes from each site interpreting its own charts, so that reporting of adverse events and disease progression was standardized across the different practice settings participating in HCV-TARGET.

The initial results were sobering. Among more than 2,000 patients treated in routine practice with a first-generation protease inhibitor (telaprevir or boceprevir) combined with peginterferon and ribavirin, 38% had cirrhosis, two-thirds developed anemia, nearly four in ten stopped treatment early, and only 52% were cured, appreciably fewer than in the triple-therapy registration trials. When guidelines endorsed simeprevir (an NS3/4A protease inhibitor) with sofosbuvir (an NS5B polymerase inhibitor) as one of the first all-oral combinations for hepatitis C on the strength of a phase 2 study, HCV-TARGET enrolled more than 800 patients, of whom 84% achieved a sustained virologic response. Over the following years, the network grew to 60 sites and more than 14,000 patients, and these data informed AASLD and EASL guidance, supported label expansions, and answered numerous safety queries from regulatory agencies and pharmaceutical partners.

What I took from those years is that the credibility of real-world evidence rests on accurate clinical documentation and on outcomes defined and interpreted by the clinicians who treat the disease, rather than on the size of the cohort alone. That principle was relatively easy to honor in hepatitis C, where the course of treatment was short and the primary endpoint was an easily measured viral load, but, as we have learned, it becomes more challenging in other liver diseases.

The credibility of real-world evidence rests on accurate clinical documentation and on outcomes defined and interpreted by the clinicians who treat the disease, rather than on the size of the cohort alone.

Developing modern infrastructure for real-world hepatology evidence

Hepatology now faces a more complex challenge than we faced in 2011. Across nearly every cause of liver disease, therapies are advancing or receiving accelerated approval on the basis of surrogate measures, among them biochemical response in primary biliary cholangitis (PBC) and fibrosis improvement or MASH resolution on biopsy in MASH. Confirmatory evidence of long-term benefit, such as fewer hepatic decompensations, transplants, or deaths, may take years to accrue, and these outcomes are rarely observed in trials of practical size, a problem that PBC and now MASH have brought to the forefront.

Furthermore, the patients who will be treated in usual clinical practice do not fully resemble those enrolled in clinical trials. They are often older and have more comorbidities, and their disease is staged clinically and with non-invasive tests such as elastography and serum markers rather than by liver biopsy. Whether long-term benefit extends to them will be settled largely by longitudinal real-world data, collected and analyzed with the same rigor we applied to hepatitis C.

Where this work continues

Pedestal Health carries forward the work that began with HCV-TARGET. Each variable and outcome measure was defined a priori by world-class hepatologists who recognized the complexities of extracting data from the medical record, as Steve Swanson described last week. As the registry grew to hundreds of thousands of patients, new methodologies were required to handle data at a scale that dwarfs the early days of HCV-TARGET.

Today’s capabilities for data acquisition, analysis, and quality measurement are beyond my wildest dreams from when we began our humble registry in 2011.

I am delighted that this work will continue to benefit the patients my colleagues in hepatology and I see every day in our clinics.