Economic Evaluation of a Community Based Diagnostic Pathway to Screen Adults for Non-Alcoholic Fatty Liver Disease

Background and Aims: Current diagnostic algorithms for detecting non-alcoholic fatty liver disease (NAFLD) are based in secondary care and associated with considerable costs and late diagnosis. We have created an integrated community based pathway utilising transient elastography and a community hepatologist review, to stratify patients at risk of developing NAFLD. This study investigated the cost-effectiveness from an NHS England perspective of this innovative diagnostic pathway (IDP). Methods: Markov modelling of the natural history of NAFLD was combined with results from a prospective cross sectional feasibility study of IDP and compared to standard care. Patients were stratified into the health states, no/ mild liver disease, significant liver disease or compensated cirrhosis and could progress to decompensated cirrhosis, hepatocellular carcinoma, liver transplant and death. Different transition probabilities were used dependent on whether early stages of liver disease were diagnosed or not. Transition probabilities, utility and resource use data were preferentially chosen from up-to-date UK sources or published literature which reflected the population. An expert panel of hepatologists was consulted to generate indicative estimates when data could not be identified. Starting age was 68, cycle length 1 yr, time horizon lifetime and cost year 2014. An incremental cost-effectiveness ratio (ICER)was estimated and one-way sensitivity (OSA) and probabilistic sensitivity analyses (PSA)were performed to assess robustness of the results. Results: We found an ICER of 2,138 per extra quality adjusted life year (QALY). OSAs demonstrated ICERs were most sensitive to varying the rate of fibrosis progression and the effect of treatment on reducing this. PSA demonstrated a 37% probability that IDP dominates SC and an 85% probability of cost effectiveness at the UK willingness to pay threshold of 20,000/QALY. Conclusions: We have not considered other likely health benefits such as reduced cardiovascular events and improved diabetic control and are probably underestimating the true impact of the pathway. Further work is also needed to improve specification of transition probabilities. However, despite the presence of significant uncertainty around estimates, the implementation of our pathway for earlier identification and management of NAFLD appears cost effective compared to standard care.