Empirical evidence of the continuing improvement in cost efficiency of an endoscopic surveillance programme for gastric cancer in Singapore from 2004 to 2010

BACKGROUND Endoscopic surveillance has been proven effective in prolonging the survival of gastric cancer (GC) patients. However, there is limited evidence on the cost efficiency of delivering this intervention, especially on a national level in spite of cost efficiency being a major determinant of the actual cost-effectiveness of a cancer prevention programme. The Singapore Gastric Cancer Epidemiology Clinical and Genetic Programme (GCEP) is a demonstration project offering scheduled endoscopy to the Chinese population aged 50 years or older in Singapore. By assessing the cost efficiency of the GCEP, this study aimed to provide empirical evidence on the cost structure and mechanisms underlying cost generation in conducting GC surveillance, thus informing resource allocation and programme budgeting for the Singapore government. METHODS From a societal perspective, we reported on the direct cost (resource consumption) of conducting endoscopic surveillance through the GCEP network. We retrospectively collected individual-level data of 216 subjects recruited at the National University Hospital, Singapore from 01/04/2004 to 31/10/2010. The Overall Cost, Clinical Cost, GCEP Cost and Personal Cost incurred in serving one subject was computed and discounted as 2004 US dollar (US$) per capita for every year. The Generalized Estimation Equation (GEE) was used to model the data. RESULTS All cost indices continuously declined over the 6.5-year costing period. For the total sample, Overall Cost, Clinical Cost, GCEP Cost and Personal Cost declined by 42.3%, 54.1%, 30% and 25.7% respectively. This downward trend existed for age and gender subgroups and the high risk group only with cost reductions varying between 3.5% and 58.4%. The GEE models confirmed statistical significance of the downward trend and of its association with risk profile, where the moderate risk group had cost indices at most 77% of the high risk group. CONCLUSIONS Our study offered empirical evidence of improved cost efficiency of a surveillance programme for GC in the early phase of programme implementation. Mechanisms such as economies of scale and self-learning were found to be involved in the cost reduction. Our findings highlighted the importance of assessing the cost efficiency and offered valuable insights for future programme budgeting and policy making.