Cost-effectiveness of non-invasive screening for alcohol-related liver fibrosis.
Alcohol-related liver disease is often undetected until irreversible late-stage decompensated disease manifests. Consequently, there is an unmet need for effective and economically reasonable pathways to screen for advanced alcohol-related fibrosis. We used real-world data from a large biopsy-controlled study of excessive drinkers recruited from primary and secondary care, to evaluate the cost-effectiveness of four primary care initiated strategies: 1) routine liver function tests with follow-up ultrasonography for test-positives, 2) the enhanced liver fibrosis (ELF) test, with hospital liver stiffness measurement (LSM) for positives, 3) three-tier strategy using Forns Index to control before strategy two, 4) direct referral of all to LSM. We used linked decision trees and Markov models to evaluate outcomes short-term (cost-per-accurate-diagnosis) and long-term (quality-adjusted life-years, QALYs). For low-prevalence populations, ELF with LSM follow-up was most cost-effective, both short-term (accuracy 96%, 196$ per patient), and long-term incremental cost-effectiveness ratio (ICER) $5,387-8,430/QALY, depending on whether diagnostic testing had lasting or temporary effects on abstinence rates. Adding Forns Index decreased costs to $72 per patient and accuracy to 95%. The strategy resulted in less QALY's due to more false negatives, but an ICER of $3,012, making this strategy suited for areas with restricted access to ELF and TE, or lower willingness-to-pay. For high-prevalence populations, direct referral to LSM was highly cost-effective (accuracy 93%, $297 per patient), with ICERs between $490 and $1,037/QALY CONCLUSION: Non-invasive screening for advanced, alcohol-related fibrosis is a cost-effective intervention when different referral pathways are used according to the prevalence of advanced fibrosis. Patients in the primary healthcare sector should be tested with the ELF test followed by LSM if test-positive, whereas direct referral to LSM is highly cost-effective in high-prevalence cohorts.