Background: Chronic liver disease is an escalating problem both in the United Kingdom and worldwide. In the UK mortality rates have risen sharply over the previous 50 years predominantly due to alcohol, however the increasing prevalence of non-alcohol related fatty liver disease both in the UK and elsewhere is also of concern. Liver disease develops silently hence early detection of fibrosis is essential to prevent progression. Primary care presents an opportunity to identify at risk populations, however assessment largely comprises of indirect markers of fibrosis which have little prognostic value. We hypothesised that setting up nurse-led primary care based liver clinics using additional non-invasive testing would increase the number of new diagnoses of liver disease compared to usual care.
Methods: This was a prospective, cluster randomised feasibility trial based in urban primary care in Southampton, United Kingdom. 10 GP practices were randomised to either intervention (liver health nurse) or control (care as usual). Pre recruitment audits were carried out in each practice to ascertain baseline prevalence of liver disease. Participants were subsequently recruited in intervention practices from July 2014-March 2016 via one of 3 pathways: GP referral, nurse led case finding based on risk factors or random AUDIT questionnaire mailouts. Liver assessment included the Southampton Traffic Light test (serum fibrosis markers HA and P3NP) and transient elastography (FibroScan). Cases were ascribed as 'no fibrosis', 'liver warning', 'progressive fibrosis' or 'probable cirrhosis'. Post recruitment audits were repeated and incident liver diagnoses captured from July 2014-September 2016. Each new diagnosis was reviewed in a virtual clinic by a consultant hepatologist.
Findings: 910 participants were seen in the nurse led clinic-44 (4.8%) probable cirrhosis, 141 (15.5%) progressive fibrosis, 220 (24.2%) liver warning and 505 (55.5%) no evidence of liver fibrosis. 450 (49.5%) cases were due to NAFLD with 356 (39.1%) from alcohol. In the 405 with a liver disease diagnosis, 136 (33.6%) were referred by GP, 218 (53.8%) from nurse led case finding and 51 (12.6%) from the AUDIT mailout. 544 incident cases were identified in the intervention arm compared to 221 in the control arm in the period July 2014-September 2016 (adjusted odds ratio 2.4, 95% CI 2.1 to 2.8).
Conclusions: The incorporation of a liver health nurse into GP practices was simple to arrange and yielded a much higher number of new diagnoses of liver disease compared to usual care. Nearly half of all participants recruited had a degree of liver disease. Nurse led case finding and GP referrals were most effective compared to AUDIT questionnaire mailouts in an urban population in identifying unknown disease. Utilising study and previous data allowed quick and effective virtual review by a hepatologist. Identifying those who are at risk of liver disease from harmful alcohol use remains a challenge and needs to be addressed in future work.