Literature Analysis of the Application of Discrete Choice Experiment in the Treatment Preferences of Patient with Type 2 Diabetes
- VernacularTitle:离散选择实验应用于2型糖尿病患者治疗偏好的文献分析
- Author:
Shimeng LIU
1
;
Shunping LI
2
;
Yi YANG
1
;
Chengxiang TANG
3
;
Yingyao CHEN
4
Author Information
1. School of Public Health，Fudan University，Shanghai 200032，China;Key Laboratory of Health Technology Assessment of National Health Commission （Fudan University ），Shanghai 200032，China
2. School of Health Management，Cheeloo College of Medicine，Shandong University，Jinan 250012，China
3. School of Public Administration，Guangzhou University，Guangzhou 510006，China
4. School of Public Health，Fudan University，Shanghai 200032，China
- Publication Type:Journal Article
- Keywords:
Type 2 Diabetes;
Patients;
Treatment preferences;
Discrete choice experiment
- From:
China Pharmacy
2020;31(20):2524-2531
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVE：To provide reference for the clinical treatment and pharmacoeconomics research of type 2 diabetes patients. METHODS ：Using“Discrete choice ”“Discrete ranking ”as Chinese keywords ，“Discrete choice ”“Discrete ranking ” “Conjoint analysis ”“Diabetes mellitus ”“Type 2”“Type 2 diabetes mellitus ”“Non-insulin-dependent diabetes mellitus ”as English keywords，Chinese and English literatures were retrieved from domestic and foreign databases as CNKI ，Wanfang database ， PubMed，Web of Science during the inception to Dec. 2019. The application status of discrete choice experiment （DCE）was analyzed and summarized from the aspect of attributes and levels ，DCE choice sets ，DCE data quality ，sample size ，econometrics analysis and the preference results. RESULTS & CONCLUSIONS ：A total of 295 related literatures were retrieved ，involving 30 valid literatures. The attributes as drug administration ，glucose control and hypoglycemic events were included more frequently. D-efficient/ D-optimal and orthogonal experiment designs were used more frequently to create the DCE choice sets. DCE data quality could be checked by the internal validity tests. The rules of thumb was usually used to calculate the sample size. Conditional Logit model ，multinomial Logit model or mixed Logit model were used more frequently to analyze the preference data. ZH187） Compared with mild hypoglycemic events ，patients’treatment E-mail：19111020032@fudan.edu.cn choices were more likely to be affected by blood glucose control. However ， when hypoglycemic events occurred at：yychen@shmu.edu.cn night or the degree of hypoglycemia was serious ， the ·treatment preference of patients would change. Although most studies included the drug administration related attributes ，they were not major factors influencing patients ’treatment preferences and were closely related to patients ’previous medication history. DCE had been widely used in the study of type 2 diabetes in foreign countries ，but there were few relevant studies in China. The data quality of DCE was difficult to control. Although the trend of building complex econometric models was gradually rising ，most studies had not fully introduced the design details such as sample size determination method ，option set design principle and quality control option. In addition ，there were some deficiencies such as too many attributes ，too large or too small horizontal spacing. It is suggested that the ten criteria of related research in ISPOR report by Bridges JF and other soholars can be used for reference in DCE design to improve the rigor of design and ensure the credibility of preference results.
