The cost-effectiveness of screening for Helicobacter pylori to reduce mortality and morbidity from gastric cancer and peptic ulcer disease: a discrete-event simulation model.

Background Helicobacter pylori is a Gram-negative bacterium that causes chronic inflammation of the stomach (gastritis). Such infection is a major cause of peptic ulcer disease and gastric cancer. H. pylori infection fulfils many of the criteria for population screening. It can be detected by a simple test, and eradicated effectively. This speeds peptic ulcer healing and prevents recurrence, but the effect that eradicating H. pylori has on the risk of gastric cancer is unknown. Screening may not be cost-effective, given the decline in H. pylori prevalence and in gastric cancer, the scope for opportunistic eradication of H. pylori and the possibility of antibiotic resistance. Objective The objective of the study was to develop a discrete-event simulation model to evaluate the cost-effectiveness of population screening for H. pylori in England and Wales to prevent both gastric cancer and peptic ulcer disease. Method A model was developed using the ‘patient-oriented simulation technique’. In this model, without screening most H. pylori-positive individuals remained asymptomatic, but a proportion developed dyspepsia and presented to primary care where testing and eradication therapy could be given. H. pylori-positive individuals were assigned increased risks of developing peptic ulcer disease and gastric cancer. In the screening scenarios, the population was invited to attend screening; those found to be H. pylori-positive were offered eradication therapy. The risk of developing peptic ulcer was considered to reduce immediately to that of H. pylori-negative individuals. The effect of eradication on gastric cancer risk was modelled by assuming a time lag before such risk reverted to H. pylori-negative levels. UK data were used where possible. The decline in gastric cancer incidence was taken into account by age cohort modelling. Costs were NHS costs based at year 2000 prices. Discounting used base rates of 6% for both costs and benefits. The model was run for an 80-year period for the base case, to allow the impact of screening on gastric cancer risk to accrue. Sensitivity analyses were performed, to explore different scenarios and where there was uncertainty about the estimated values used in the model. Both incident and prevalent screening rounds were modelled, using the population of England and Wales. There were four age-related scenarios: screen all people aged 20–49 as a prevalent round and then 20-year-olds for the next 20 years as an incident round; screen those aged 30–49 as a prevalent round and then 30-year-olds for 30 years as an incident round; screen those aged 40–49 as a prevalent round and then 40-year-olds for 40 years as an incident round, and finally screen all 50-year-olds for 50 years as an incident round. Results Population screening would involve screening approximately 25 million individuals if uptake was 70%, with over 5 million people being treated. The number of deaths prevented falls with increasing age at screening, but so does the present value of costs because there would be less prevalent screening and costs are deferred. In the base case the cost-effectiveness of H. pylori screening improves with age and is under £10,000 per life-year saved (LYS) for all age groups, though over an 80-year follow-up. Lowering the discount rate for benefits significantly improves the cost/LYS to under £2000 in all groups. Increasing the time lag for reversion of gastric cancer risk to 20 years or increasing the level of opportunistic eradication reduces the relative advantage for screening. Screening at age 40 might be the most pragmatic policy, balancing cost-effectiveness and the feasibility of screening. The cost/LYS for the base run at age 40 is £5866 falling to £1027 if the benefit is discounted at 1.5%. Screening by serology is more cost-effective than using the urea breath test. Using a less efficacious but cheaper eradication regimen is as cost-effective but with fewer deaths prevented. The cost-effectiveness is sensitive to the H. pylori prevalence, lag time, relative risk, cohort estimate and compliance. Moreover, cost/LYS rises to over £20,000 if there is a high level of opportunistic eradication of H. pylori in patients presenting with dyspepsia and a reduced efficacy of eradication on gastric cancer risk. The benefits of screening take time to accrue and are very sensitive to the discount rate. At 6% rates the cost-effectiveness does not fall below £20,000 for 30 years. Conclusion Population screening for H. pylori is likely to be cost-effective with a cost/LYS of under £10,000 for the base assumptions, which compares favourably with other screening programmes. However the benefits take time to accrue and this cost/LYS is over an 80-year follow-up. Once-only screening at age 40 with a prevalent round for people aged 40–49 appears to be the most pragmatic policy. A major uncertainty is the effect of eradication of H. pylori on gastric cancer risk. The cost-effectiveness of H. pylori screening would be reduced if there were extensive H. pylori opportunistic testing of all dyspeptic individuals presenting to primary care.