Can gatekeepers help contain health-care costs for diabetes in one eastern city in China: a multistage difference-in-difference analysis
Luying Zhang,Xinyi Liu,Wen Chen
DOI: https://doi.org/10.1016/S0140-6736(20)32425-9
IF: 202.731
2020-01-01
The Lancet
Abstract:Background
Primary care has been recognised as an effective way to meet health demands and reduce health-care costs. As a key part of its health-care reforms, China has been striving to strengthen primary care and rebuild referral systems, with family physicians being considered as gatekeepers tasked with providing continuous treatment and management of chronic diseases. However, little is known about the effect of these gatekeepers on the health-care system in China. Our study aims to examine whether gatekeepers can effectively reduce the costs associated with patients with diabetes.
Methods
We chose an eastern city in China to do our study, given that it introduced a so-called gatekeeper system in 2015, as one of the earliest pilots in China. All registered patients with diabetes were identified from the official health management information system. By linking patient identifications, we extracted individual-level data from the medical insurance claim database, including patient demographic characteristics, health-care use, and expenditure from 2014 to 2017. Inclusion criteria for the target population were patients insured by social medical insurance and those who had health-care use records and diabetes management follow-up records in the information system for every single year. Because the gatekeeping system was rolled out in a gradual way, with patients with diabetes who newly register with general practitioners being included each year, we used a multistage difference-in-difference analysis to evaluate the net difference in annual medical costs caused by the reform.
Findings
Our sample included 37 955 people with diabetes (mean age 67·8 years, IQR 61–76; 20 538 [54%] men and 17 417 [46%] women) who voluntarily registered with their family physicians between Jan 1, 2014 and Dec 31, 2017. Of these patients, 26 668 (70·2%) enrolled with the gatekeeping system for 3 years, 5426 (14·3%) enrolled for 2 years, and 2187 (5·8%) enrolled for 1 year, and a further 3674 (9·7%) were not enrolled with the gatekeeper system and acted as a control group. According to the regression model results, compared with patients who were not enrolled, patients enrolled with gatekeepers obtained a mean of 0·25 more follow-up visits (0·15 [SE 0·05] more in year 1, 0·22 [SE 0·08] more in year 2, and 0·35 [SE 0·14] more in year 3) and 1·9 more community outpatient visits (1·74 [SE 0·13] more in year 1, 2·00 [SE 0·19] more in year 2, and 2·09 [SE 0·27] more in year 3), and their medical cost slightly decreased by 2·30% (0·90% [SE 0·01] less in year 1, 2·65% [SE 0·01] less in year 2, and 3·32% [SE 0·02] less in year 3) compared with the control group, with a dose-response effect. All of these results were statistically significant.
Interpretation
Family physicians voluntarily enrolled as so-called soft gatekeepers can be effective at increasing health-care use at the primary care level by patients with diabetes and at containing medical costs. The decrease in medical costs could be attributed to the substitution of hospital care and retail pharmacy services with primary care. These findings have substantial policy implications for China, and act as an incentive to further gatekeeper reforms. Further research is needed to provide evidence for the long-term use of this system.
Funding
National Natural Science Foundation of China Grant (number 71603055), China Medical Board Health Policy, and System Science Open Competition Grant (number GNL 15-225).