Chronic viral hepatitis causes an asymptomatic disease that leads to cirrhosis and death in a large proportion of those infected. Often not associated with abnormalities of liver function, these diseases preclude identification by simple non-specific testing. The prevalence of viral hepatitis is at least ten times greater in immigrants than in the indigenous community (prevalence in the UK is ~0.5% but >5% in immigrants) and there is abundant data that the mortality and morbidity in ethnic minorities is growing rapidly. Early identification of people with chronic viral hepatitis allows effective therapy, shown by NICE to be highly cost effective. Given the rising mortality in ethnic minorities from chronic viral hepatitis, the Advisory Group on Hepatitis and The National Screening Committee have considered screening immigrants but there is no data on whether this is feasible and clinically and/or cost effective.
The UK has one of the lowest rates of therapy for viral hepatitis in Europe and this is undoubtedly contributing to the rising mortality from liver disease seen here. Studies in the UK have shown that access to therapy for patients known to have viral hepatitis is poor, with only a minority of diagnosed patients going on to receive treatment. Strategies that improve access to treatment are likely to have a major impact on treatment uptake but alternatives to hospital based treatment have not been studied.
The wider collaborative project aims to determine whether:
Screening immigrants for chronic viral hepatitis in primary care is feasible, acceptable to the communities at risk and cost effective.
Therapy for viral hepatitis in primary care is achievable, and both clinically and cost effective.
Community treatment improves engagement and increases the number of patients who access effective therapies that are NICE approved.
PenCHORD will carry out the cost effectiveness analysis aspect of this project, particularly the cost effectiveness of screening and community care. This study will build on previous work conducted for hepatitis C in former injecting drug users and prison populations. Models from this work utilised baseline values derived from published studies, information on health-related quality of life (QoL) data, alongside a randomised controlled trial. In addition, QoL values for patients with decompensated liver disease, hepatocellular carcinoma, and transplant related states were drawn from a large UK liver transplantation study. These utility values were chosen since they were obtained from UK patients with hepatitis C using a preference-based measure, in contrast to previous studies that relied on expert opinion. For the PenCHORD project, these QoL measures will form a starting point for the cost effectiveness analysis but the model values will be refined and a thorough literature review will be conducted to investigate any recent sources of utility values relevant to the analysis and these will be incorporated where appropriate.
The previously developed economic models, used to assess case finding for viral hepatitis in other groups will be adapted to immigrants. Differences in key parameters will be considered such as prevalence, engagement and varying response rates caused by the different distribution of viral genotypes. Factors specific to the ethnic minority population, as well as variations due to the differing forms of hepatitis will be explored. To establish values for utility, the relevant evidence base to determine valid estimates for QoL in relation to these additional factors will be considered.
The relative costs and benefits for the different scenarios of interest will be the key outputs (representing alternatives for opportunistic case finding, screening, and treatment). Benefits will be assessed in terms of the aggregated Quality Adjusted Life Years (QALY) over the life-time of the modelled patient cohort, using existing utility data sourced for the different patient states in the model. The incremental costs for compared scenarios will adopt an NHS and Personal and Social Services perspective. Extensive one-way and probabilistic sensitivity analysis will be used, both to explore the uncertainty surrounding the utility parameters but also importantly, to identify which utility parameters are instrumental in driving the outputs of the model. Any utility values to which the model is particularly sensitive can then be a focus of more thorough investigation to derive refined estimates and confidence limits.
The existing model incorporates a testing and diagnosis component (implemented as a decision tree). Treatment is incorporated as a Markov model, which represents the progression of HCV disease as transitions between discrete health states (mild, moderate, severe hepatitis; cirrhosis; decompensated cirrhosis; transplant and death). This model has been successfully used to assess the cost-effectiveness of case finding for former drug users, as well as other scenarios. A hypothetical cohort of patients undergoing case-finding and treatment was compared to a cohort in whom no case-finding was implemented but spontaneous presentation was allowed. The model showed the cost-effectiveness of case-finding at normal willingness-to-pay thresholds for a variety of scenarios. This model will be re-configured to study screening in immigrant populations by re-structuring and re-parameterisation. To support this, the work on community treatment will provide detailed cost and outcome data on a cohort of patients treated in primary care. This will be used to extend the existing model, ensuring it reflects the study population. An extensive cost analysis will be undertaken to ensure that all marginal costs for alternative scenarios are incorporated. It is envisaged that immigrant screening will incorporate service payments, described above and the cost effectiveness implications of this and other payments will be examined.
Outputs from the cost-effectiveness model will provide estimates of the relative cost-effectiveness for a range of scenarios for community versus hospital-based therapy. A range of outputs and graphical presentation methods will be used to ensure that the outputs of interest (including all statistics relating to patient events) are clearly reported. A key part of the study will be extensive one-way and probabilistic sensitivity analysis. This will analyse the statistical uncertainty surrounding each of the model input parameters (treatment costs, response rates etc). Sensitivity analysis provides a means to identify those parameters which have most impact on outcomes. This analysis will report the probability that the intervention of interest (community based therapy) is cost effective at cost per QALY thresholds typically used by NICE (£20-30,000 per QALY).
This model will fully examine which parameters are most important in driving decision uncertainty and, if appropriate, will use value of information analysis to assess the impact of uncertainty within the model and the value and importance of further research in the area.