Statistical analyses

The primary analysis is a comparison of upper gastrointestinal cancer mortality between the screening group and the controls. Mortality rate will be calculated as the ratio of the number of death due to upper gastrointestinal cancer to the person-year at risk for each group. Person-years will be estimated from the time of randomization to the date of upper gastrointestinal cancer death or censoring at the end of the study, which will be decided at a later time. The secondary outcomes including incidence rate and detection rates will be estimated in a similar way. SAS software (version 9.2; SAS Institute, Cary, NC, USA) will be used for these statistical analyses. χ² tests and t-tests are used to compare categorical and continuous variables between the two groups, respectively. The Cox proportional hazards regression model is chosen to analyze the difference of incidence and mortality between screening group and control group.

We will create a decision tree-Markov model to evaluate the cost-effectiveness of endoscopic screening for upper gastrointestinal cancer in China, using the software of TreeAge Pro 2016 (TreeAge Software, Inc., MA, USA) (23). The perspective for the evaluation was societal, and the model was run with a cycle length of 1 year for 36 cycles (i.e. from 40 to 75 years old) in order to cover the entire upper gastrointestinal cancer experiences of the vast majority of cohort members. Base-case values and ranges of parameters used in the model will be estimated from a variety of sources: 1) the results of our prospective evaluation study of screening for upper gastrointestinal cancer; 2) surveillance data of Chinese National Cancer Center; 3) published literatures; 4) government documents; and 5) our survey. Costs and utility scores of upper gastrointestinal cancer-related diseases will be estimated based on our survey. Given the uncertainty about some parameters, deterministic and probabilistic sensitivity analyses will be used to assess the robustness of the model results.

The study is approved by the independent ethics committee of National Good Clinical Practice Center for Anticancer Drugs, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (2015SQ00223), and the study used the Protocol Registration System in Chinese Clinical Trial Registry (identifier: ChiCTR-EOR-16008577).