REMR: Stata module to implement robust error meta-regression method for dose–response meta-analysis

remr performs dose-response meta-analysis using inverse variance weighted least squares (WLS) regression with cluster robust error variances. This approach is a special case of the one-stage generalised least squares for trend approach where the covariance need not be imputed from the data. This method allows the model to include a non-zero intercept to absorb any bias that is not accounted for by the terms in the model but notably the intercept now becomes a coefficient rather than a constant given the use of WLS. Weights for the reference doses are imputed, these are equal to the maximum of the within study weights based on the inverse of the variance of the effect sizes in the study. This minimises the possible deviation of the regression intercept from the origin without having to force the regression model through the origin. This method does not require knowledge of the correlation structure of the data within a study, because it stacks included effects as a cluster by study and uses the cluster-robust analysis to obtain a robust standard error, thus treating observations as independent across clusters but correlated within each cluster.