Analysis of operation performance of general hospitals in Shenzhen, China: a super-efficiency data envelopment analysis. (October 2016)

• Record Type: Journal Article
• Authors: Chen, Shanquan; Wu, Yushan; Chen, Yao; Zhu, Haidi; Wang, Zheng; Feng, Da; Feng, Zhanchun; Yao, Lan; Xiang, Li; Wong, Eliza Lai-yi; Fung, Hong; Yeoh, Eng-kiong; Liu, Zhiyong
• Is Part Of: Lancet. Volume 388(2016)Supplement 1
• Page Start: S57
• Publication Date: 2016-10
• Subjects: Medicine -- Periodicals; Medicine -- Periodicals; Medicine; Medicine; Electronic journals; Periodicals; 610.5
• DOI: 10.1016/S0140-6736(16)31984-5
• Languages: English
• ISSNs: 0140-6736

Abstract

Abstract: Background: Hospitals are the main and last component of health-care systems. We aimed to assess the performance of Shenzhen's general hospitals during 2010–15, and to suggest ways in which to improve operation performance. Methods: Our study included 64 general hospitals in Shenzhen, of which six were third-level hospitals, 18 second-level hospitals, and 40 first-level hospitals. Five input variables (number of physicians, number of nurses, number of beds, the number of pieces of equipment costing >¥100 000, and fixed assets) and five output variables (number of outpatient visits, number of inpatient cases, length of stay, bed occupation rate, and total revenue) were filtered as evaluation measures. A super-efficiency data envelopment analysis (super-DEA) model was used to assess annual operation performance. Malmquist productivity index was used to analyse intertemporal change in operation performance and for regression analysis to explore ways in which to increase hospitals' efficiency. Findings: 19 (48%) first-level hospitals had an average super-efficiency score lower than 1, indicating an inefficient performance, compared with eight (44%) second-level hospitals and no third-level hospitals. 18 (46%) first-level hospitals were technically inefficient, compared with three (17%) second-level hospitals and one (13%) third-level hospital. For pure technological efficiency, 15 (38%) first-level hospitals had a score lower than 1, indicating inefficiency, compared with three (17%) second-level hospitals and one (13%) third-level hospital. Ten (25%) first-level hospitals were inefficient in scale, compared with one (6%) second-level hospital and no third-level hospitals. The regression analysis showed that the significant contributing factor for first-level hospitals was the improvement of technological efficiency (p<0·0001), whereas the factor for second-level and third-level hospitals was the improvement of allocative efficiency (both p<0·0001). Interpretation: With use of the super-DEA model, our analysis shows that third-level hospitals are efficient, but almost half of first-level and second-level hospitals are inefficient. To increase efficiency or improve inefficiency, the preferred measure is to improve the allocative efficiency for third-level and second-level hospitals, and to improve the technological efficiency for first-level hospitals. Funding: Health and Family Planning Commission of Shenzhen Municipality.