S3-E39 - Reviewing a Decision Model for Early NASH Diagnosis

Send us Fan MailOne serious challenge in the overall management of Fatty Liver disease involves creating cost-effective methods for "early" diagnosis. The term "early" is relative because, as Ian Rowe puts it, a "substantial proportion" of people admitted to hospital with various symptoms of decompensating cirrhosis or hepatic encephalopathy never received a diagnosis of advanced liver disease before they presented. Simply diagnosing these patients during advanced fibrosis (F2 or F3) can save lives, improve longevity and quality of life for these patients and save money for healthcare systems, all at the same time. Ian spends the first ten minutes of this episode describing a model he and Richard Parker developed to evaluate five diagnostic strategies:Targeted -- simple non-invasive liver screening Targeted + risk stratification -- targeted assessment plus FIB-4 or VCTE (FibroScan) for patients believed to be at high riskiLFT -- automated assessment of abnormal LFTs analyzed with other systemic measures in a medical record system "Comprehensive" -- a "kitchen sink" approach that runs iLFT analysis and conducts FIB-4 +/- VCTE on every patient"Fibrosis first" -- an approach that deploys viroserological and iron testing plus FIB-4 +/- VCTE for every patient "Fibrosis first" scored best in cost effectiveness (cost per correct diagnosis of treatable liver disease) and "decision curve analysis", which looks at true positive outcomes and false positive outcomes, correctly identifying 85% of treatable liver disease patients (vs. 90% for the comprehensive approach and less than 15% for the current targeted approach.)  Fibrosis first has the added benefit of not progressing treatment for people who may have steatosis without fibrosis. The rest of the conversation entails Ian, Louise Campbell and Roger Green sharing questions and observations. Highlights:Louise agrees that the current targeted pathway approaches do not work and  identifies lack of access to advanced testing in primary care as a central issue. She asks how many additional patients with Type 2 Diabetes, other metabolic diseases or CVD could be identified through this approach. Ian cannot answer because the question is outside the scope of the research. He points out that to assess different approaches accurately, researchers and policymakers need to settle on whether the goal is to identify steatohepatitis or something broader. Louise asks about the frequency and nature of referrals. Ian notes that one issue with iLFT is a high level of referrals for conditions that might not require treatment, the most common of which is "abnormal LFT without fibrosis."Roger notes that iLFT might help patients who complain they never heard about their NASH earlier. He also recalls Quentin Anstee and Stephen Harrison discussing recently that FIB-4 can serve as a prognostic measure not only for liver-related deaths, but also CV deaths and all-cause mortality. Louise suggests that patients will do better if they adopt a  behavior belief model instead of a sick patient approach and that the earlier in disease we intervene, the more success we will have. Roger and Ian note that Louise's focus is broader in scope than his model. Ian further notes that we do not know when non-fibrotic patients should be tested again, although he has some thoughts about this. Louise notes that one element that might make iLFT costly is that only 45% of patients picked up on iLFT are referred into the model. Ian notes that this insight might require changes in the model.As the conversation winds down, it shifts toward policy issues: should we focus testing on the workforce (Louise) and it the best point of intervention to limit advertising of unhealthy foods (Ian).