Background and Aim Surveillance of non-dysplastic Barrett's esophagus (BE) is recommended to identify progression to dysplasia; however, the most cost-effective strategy remains unclear. Mutation or aberrant expression of p53 has been associated with the development of dysplasia in BE. We sought to determine if surveillance intervals for BE could be stratified based on p53 expression. Methods A Markov model was developed for non-dysplastic BE. Patients with non-dysplastic BE (NDBE) underwent p53 immunohistochemistry (IHC) and those with abnormal p53 expression underwent surveillance endoscopy at 1 year, whilst patients with normal p53 expression underwent surveillance in 3 years. Patients with dysplasia underwent endoscopic therapy and surveillance. Results On base-case analysis, the strategy of stratifying surveillance based on abnormal p53 IHC was cost-effective relative to conventional surveillance and a natural history model, with an ICER of $8,258 for p53 IHC based surveillance. Both conventional and p53 stratified surveillance strategies dominated the natural history model. On probabilistic sensitivity analysis (PSA), the p53 IHC strategy ($28,652, 16.78 QALYs) was more cost-effective than conventional surveillance ($25,679, 16.17 QALYs) with a net monetary benefit (NMB) of $306,873 compared to conventional surveillance ($297,642) with ICER < $50000 in 96% of the iterations. The p53 stratification strategy was associated with a 14.5% reduction in the overall endoscopy burden and a 59% increase in dysplasia. Conclusion: A surveillance strategy for BE based on abnormal p53 IHC is cost-effective relative to a conventional surveillance strategy and is likely to be associated with higher diagnostic rates of dysplasia.