[Economic burden of cancer in China during 1996-2014: a systematic review]

Objective: To explore the current status of research on economic burden of cancer in China from 1996 to 2014. Methods: The key words including cancer, economic burden, expenditure, cost were used to retrieve the literatures published in CNKI and Wanfang (the two most commonly used databases for literature in Chinese) and PubMed during 1996-2014. A total of 91 studies were included after several exclusionary procedures. Information on subjects and data source, methodology, main results were structurally abstracted. All the expenditure data were discounted to year of 2013 value using China's health care consumer price indices. Results: More than half of the included studies were published over the past 5 years, 32 of the studies were about lung cancer. Among the 83 individual-based surveys, 77 were hospital-based and obtained data via individually medical record abstraction, and most of which only considered the direct medical expenditure. Expenditure per cancer patient and expenditure per diem were the most commonly used outcome indicators. Majority of the findings on expenditure per cancer patient ranged from 10 thousands to 30 thousands Chinese Yuan (CNY), with larger disparity in lung and breast cancer (ranged from 10 thousands to 90 thousands CNY), narrower difference in esophageal and stomach cancer (ranged from 10 thousands to 50 thousands CNY), and most stable trend in cervical cancer (almost all the values less than 20 thousands CNY). Without exception, the expenditures per diem for all the common cancers were increasing over the period from 1996 to 2014 (3-7 fold increase). Only 8 population-level economic burden studies were included and the reported expenditure of cancer at national level ranged from 32.6 billions to 100.7 billions CNY. Conclusions: Evidence on economic burden of cancer in China from 1996 to 2014 are limited and weakly comparable, particularly at a population level, and the reported expenditure per patient may be underestimated.