Welcome to the first newsletter for the Cochrane Statistical Methods Group

Issue 1 December 2023

Cochrane Colloquium London

The Statistical Methods Group meeting was held during the Cochrane Colloquium in London on 5th September, our first face to face meeting since 2018 in Edinburgh. It was great to see many there despite the early start. 

We gave a short overview of groups activities over the past year, including training activities that the group have delivered and our very successful webinar series.

We gave a summary of a survey we ran during the summer asking for members’ feedback on potential research and training topics as well as more general feedback on the group’s activities. Thank you to everyone who provided helpful responses. Following this feedback we are aiming to communicate with members more frequently through this new newsletter and would welcome your suggestions of items to include in future issues. Please do use the mailing list (smglist@lists.cochrane.org) to post ideas and questions for the community to discuss. We would also highlight the archive of learning resources and webinars below.  

We provided an overview of the new random effects methods to be implemented in RevMan Web. Specifically, we introduced new methods to: estimate the heterogeneity variance (and its confidence interval), calculate a confidence interval for the meta-analysis effect, and, calculate a prediction interval. We outlined the process followed by the Statistical Methods Group to develop these recommendations, and elaborated on the next steps regarding the transition and implementation periods.

Dr. Skipka presented on a method used by the Institute for Quality and Efficiency in Health Care (IQWiG) to summarize the results of several studies in a qualitative manner when heterogeneity is excessive to meaningfully estimate a meta-analysis effect. The concept of conclusive effects was presented. This concept includes: number of studies, prediction interval, direction of estimations, weight of studies, and statistical significance of the study results. If a direction of the effects can be recognized on the basis of this concept, recommendations can be made regarding benefit or harm.

For those who could not make it to the Colloquium this year we have provided access to the presentations here

Webinar series

A reminder that we run regular webinars through Cochrane Learning Live with details of upcoming webinars posted on the smglist. Please do send us suggestions of presenters and topics.