Introduction At a time when mortality for many diseases is falling, deaths from liver disease have increased fivefold over the last 30 years and have almost doubled over the last 10 years. In order to reduce liver deaths we need to pick up liver disease at a much earlier stage in the community so we need easier more accurate diagnostic methodologies. We used two serum markers and standard tests to develop a simple traffic light diagnostic modality.

Aim To develop a simple one-step traffic light diagnostic modality that could be used to triage liver disease of all aetiologies in the community.

Method Subjects were out-patients or in-patients at Southampton University Hospitals Trust between 2003 and 2009. Serum fibrosis markers hyaluronic acid and collagen P3NP were combined with FBC, INR and LFTs in a test cohort (n=119) with a firm diagnosis of the stage of fibrosis. On the basis of the test data, a simple clinical traffic light algorithm was created using HA, P3NP, INR, Albumin, and platelets. An evaluation cohort of 591 subjects, of whom 278 had independent staging of fibrosis, was used to AUROC validate the model and compare a modified binary logistic model—in the abstract data sets are combined.

Conclusion The traffic light system detected 98% of subjects with cirrhosis and 88% of subjects with progressive fibrosis, in the latter mortality over 5 years was not increased. We envisage this system being used to inform subjects about the need to moderate high risk behaviour, and triage subjects with severe disease to secondary care. Repeat testing at 5-year intervals in increased risk groups should have the potential to pick up most cases of significant liver disease before a fatal presentation to hospital. The traffic light system has been applied to the detection of liver disease in a community sample of 10 000 in the ALDES study funded by NIHR.