| Abstract: | BACKGROUND & AIMS: Globally, 25% of people have nonalcoholic fatty liver disease (NAFLD), and, currently, there are no approved pharmacologic treatments for NAFLD. With a slow disease progression, long-term impact of pharmacologic treatments can be assessed only by complementing emerging clinical trial evidence with data from other sources in disease progression modeling. Although this modeling is crucial for economic evaluation studies assessing the clinical and economic consequences of new treatments, the approach to modeling the natural history of NAFLD differs in contemporary research. This systematic literature review investigated modeling of the natural history of NAFLD. METHODS: A systematic literature review was conducted searching PubMed, Scopus, Cochrane, and the National Health Service Economic Evaluation Database to identify articles focusing on modeling of the natural history of NAFLD. Model structure and transition probabilities were extracted from included studies. RESULTS: Of the 28 articles identified, differences were seen in model structure and data input. Clear definitions of nonalcoholic steatohepatitis and NAFLD often were lacking; differences in the granularity of modeling fibrosis progression, the approach to disease regression, and modeling of advanced liver disease varied across studies. Observed transition probabilities for F0 to F1, F1 to F2, F2 to F3, and F3 to compensated cirrhosis varied between 0.059 to 0.095, 0.023 to 0.140, 0.018 to 0.070, and 0.040 to 0.118, respectively. CONCLUSIONS: The difference in disease progression modeling for seemingly similar models warrants further inquiry regarding how to model the natural course of NAFLD. Such differences may have a large impact when assessing the value of emerging pharmacologic treatments. |
