Variability-in-meta-analysis-estimates-of-continuous-outcomes

This repository is part of the methodological work in evidence synthesis developed in the Centre for Reviews and Dissemination (CRD; University of York) and the Bristol Medical School (University of Bristol). It investigates the variability in meta-analysis estimates of continuous outcomes using different standardization and re-expression methods.

Study analysing different standardization methods and how to re-express those estimates using scale-specific re-expression methods

Our work entitled "Variability in meta-analysis estimates of continouous outcomes using different standardization and scale-specific re-expression methods" tried to clarify (1) the impact of using different data standardization and scale-specific re-expression methods in meta-analyses using standardized mean differences (SMDs), and (2) to give some recommendations to promote methodological guide and transparency in meta-analyses. The full text is available in: https://www.jclinepi.com/article/S0895-4356(23)00297-4/pdf.

When you import the dataset in your environment...

You may need some clarification on the meaning of the variables. Let's crumble our dataset:

• studyID refers to each of the included studies in the meta-analysis.
• class is the first level of specificity referring to the intervention. In this level we can find Physical activity or Usual care.
• agent is the second level of specificity referring to the intervention. In this level we can find the specific intervention that a study performed (e.g., Ambulation consisted on movements oriented to real life and/or walkings across the wards).
• component is the third level of specificity. It is a dismantling of each intervention.
• pre_n is the study sample at baseline.
• pre_meanis the mean value of physical performance measured by the Short Physical Performance Battery (SPPB) or the Barthel Index (BI) at baseline.
• pre_sd is the standard deviation (SD) of the mean value at baseline.
• post_n is the study sample at post-intervention time point.
• post_meanis the mean value of physical performance measured by the Short Physical Performance Battery (SPPB) or the Barthel Index (BI) at post-intervention time point.
• post_sd is the standard deviation of the mean value at post-intervention time point.
• ward referes to the type of hospital ward where the intervention was conducted.
• measure corresponds to the scale used to measure the physical performance of the patients.
• lower_better refers if a lower mean value means a greater improvement in the outcome. This is not the case.
• outcome is the type of analysed outcome.
• y is the mean change from baseline.
• sd is the standard deviation of y.
• se is the standard error of y.
• post_sd_pooled corresponds to the pooled standard deviation of a specific scale and study at post-intervention time point.
• pre_sd_pooled corresponds to the pooled standard deviation of a specific scale and study at baseline.
• y_int is the rescaled mean change from baseline using an internal SD reference by scale.
• sd_int is the rescaled standard deviation of y_int.
• y_ext is the rescaled mean change from baseline using an external SD reference by scale.
• sd_ext is the rescaled standard deviation of y_ext.
• diff refers to the mean difference between intervention and control arms of the same study.
• se_diff is the corresponding standard error of diff values.
• diff_ext is the rescaled mean difference using an external sD reference by scale.
• se_diff_ext is the rescaled variance of diff_ext.
• diff_int is the rescaled mean difference using an internal sD reference by scale.
• se_diff_int is the rescaled variance of diff_int.
• diff_pre is the rescaled mean difference using the pooled SD of each study at baseline.
• se_diff_pre is the rescaled variance of diff_pre.
• diff_post is the rescaled mean difference using the pooled SD of each study at post-intervention time point.
• se_diff_post is the rescaled variance of diff_post.