Amy Ming-Fang Yen

Evidence-based medicine (EBM) has been a sine qua non for evaluation of intervention program of each discipline of medicine since 1970. It has become basic information on which the efficacy and effectiveness of most of decision-making policies and strategies are based since 1990. It forms the bedrock for three components of synthesis science, meat-analysis, decision analysis, and also economic appraisal. However, as the effectiveness of prevention program is affected by a number of determinants it is therefore rather heterogeneous and varies from place to place. In addition, evaluation of a service screening program in a small local area also suffers from small sample size as well as the absence of a control group as like in the randomized controlled trial. The use of stochastic process to elucidate the sequel of diseases enables one to project the effectiveness and cost-effectiveness of intervention programs. In addition, Bayesian analysis is useful for analyzing small-area data by taking other data sources as prior information from the population subjected to the same health services to increase precision without comprising the validity.

Here, we demonstrate two examples applied in small area. First, the administration of H. Pylori eradication to prevent gastric cancer in Matsu was followed by the elucidation of disease progression of gastric neoplasm using stochastic process and the economic evaluation among different intervention programs. The second example was the evaluation of long-term effectiveness of population-based breast cancer service screening program in a small geographic area in Finland borrowing prior evidence of breast cancer screening from neighboring countries with Bayesian approach underpinning.