The benefits, harms and costs of surveillance for hepatocellular carcinoma in people with cirrhosis: synthesis of observational and diagnostic test accuracy data and cost–utility analysis

RESEARCH QUESTION What are the benefits, harms and costs of different surveillance regimens (compared with each other and no surveillance) for hepatocellular carcinoma (HCC) in people with cirrhosis?

BACKGROUND An influential 2007 NIHR HTA assessed the cost effectiveness of HCC surveillance in people with cirrhosis. However, the liver disease environment has changed in many ways since then. Our project will provide an up-to-date assessment of whether and how surveillance should be offered to people with cirrhosis.

METHODS There will be 4 work-packages (WPs). WP1 is a systematic review and meta-analysis of cohort studies comparing HCCs found under surveillance with those diagnosed incidentally or symptomatically. Outcomes cover tumour characteristics, treatment received and survival. As all such evidence is subject to a high risk of bias, we will assess methods carefully and evaluate any attempts to adjust for known biases.

WP2 is a systematic review and synthesis of diagnostic accuracy data relating to tests for HCC in people with cirrhosis. We will look at imaging (ultrasound, CT, MRI), conventional biomarkers (AFP, DCP) and genomic analytes (microRNA, circulating tumour cells). We will use Bayesian statistical techniques to estimate how tests work at all possible thresholds and in comparison with each other, especially when it comes to picking up HCC at an early stage.

In WP3, we will develop a decision-model to estimate the lifetime costs, benefits and harms of different HCC surveillance regimens, incorporating estimates of test accuracy from WP2. It will simulate the natural history, diagnosis and treatment of HCC. As well as benefits, we will simulate harms associated with surveillance.

We will explore whether the balance of benefits, harms and costs varies according to patients’ characteristics (e.g. aetiology of cirrhosis, presence of other long-term conditions, eligibility for HCC treatment).

In WP4, we will develop a print-based patient decision-aid, using what we have learned in WPs1–3 to quantify the expected benefits and harms of surveillance. It will aim to support shared decision-making for people with cirrhosis about whether to start – or stop – surveillance.

PATIENT AND PUBLIC INVOLVEMENT Our PPI strategy comprises 3 elements: (1) recruiting 2 liver patients as co-investigators; (2) holding 2 online workshops with 10–15 participants; at the second, we will alpha-test the decision-aid developed in WP4; (3) partnering with British Liver Trust (BLT), who will sit on our advisory group and facilitate engagement with the liver disease community (c25,000 forum users; c16,000 social media followers) to shape our research and enhance dissemination.

DISSEMINATION BLT are going to help us share our findings with patients and professionals; they will be invited to badge and publicise the decision-aid from WP4. We will target NICE and national liver organisations to maximise the impact of our research for population-level decision-making. We will publish in scientific journals and present at conferences. We will make our statistical and economic code freely available for future researchers.

PROJECT PLAN The project will take 21 months. The project team includes clinicians with expertise in hepatology and radiology, statisticians, health economists, evidence reviewers and patient experts. We have recruited an advisory group of methods, clinical and patient experts who will provide oversight and advice throughout the project.