Chao Zhang1, Pengli Jia2, Liu Yu3, Chang Xu2. 1. Center for Evidence-based Medicine and Clinical Research, Taihe Hospital, Hubei University of Medicine, Shiyan, China. 2. Chinese Evidence-Based Medicine Center/Chinese Cochrane Center, West China Hospital, Sichuan University, Chengdu, China. 3. Gansu Provincial Maternity and Child-care Hospital, Gansu, China.
Abstract
OBJECTIVE: Dose-response meta-analysis (DRMA) is widely applied to investigate the dose-specific relationship between independent and dependent variables. Such methods have been in use for over 30 years and are increasingly employed in healthcare and clinical decision-making. In this article, we give an overview of the methodology used in DRMA. METHODS: We summarize the commonly used regression model and the pooled method in DRMA. We also use an example to illustrate how to employ a DRMA by these methods. RESULTS: Five regression models, linear regression, piecewise regression, natural polynomial regression, fractional polynomial regression, and restricted cubic spline regression, were illustrated in this article to fit the dose-response relationship. And two types of pooling approaches, that is, one-stage approach and two-stage approach are illustrated to pool the dose-response relationship across studies. The example showed similar results among these models. CONCLUSION: Several dose-response meta-analysis methods can be used for investigating the relationship between exposure level and the risk of an outcome. However the methodology of DRMA still needs to be improved.
OBJECTIVE: Dose-response meta-analysis (DRMA) is widely applied to investigate the dose-specific relationship between independent and dependent variables. Such methods have been in use for over 30 years and are increasingly employed in healthcare and clinical decision-making. In this article, we give an overview of the methodology used in DRMA. METHODS: We summarize the commonly used regression model and the pooled method in DRMA. We also use an example to illustrate how to employ a DRMA by these methods. RESULTS: Five regression models, linear regression, piecewise regression, natural polynomial regression, fractional polynomial regression, and restricted cubic spline regression, were illustrated in this article to fit the dose-response relationship. And two types of pooling approaches, that is, one-stage approach and two-stage approach are illustrated to pool the dose-response relationship across studies. The example showed similar results among these models. CONCLUSION: Several dose-response meta-analysis methods can be used for investigating the relationship between exposure level and the risk of an outcome. However the methodology of DRMA still needs to be improved.