Article type
Abstract
Background
Network meta-analyses (NMAs) are a methodological innovation, designed to synthesise results from studies of multiple competing interventions. They are vital in topic areas where many different treatments are available, and most trials are placebo-controlled. NMAs reduce research waste by maximising data use from primary studies, eliminating need for further trials.
Objectives
We completed an NMA in December 2023, synthesising data from nearly 300 studies investigating topical anti-inflammatory treatments for eczema. We aim to describe the challenges faced and how we sought to overcome them.
Take-home points
1)Outcome measurements: we used the Core Outcome Set framework to ensure that we measured the key outcomes for stakeholders. However, there was heterogeneity in the way outcomes were measured. One way we overcame this was by using the Standardised Mean Difference for continuous outcomes and re-expressing the results using scales familiar to readers.
2)Outcome completeness: missing dispersion measures meant some continuous outcome data couldn’t be synthesised. There were also missing data on number of included participants and drop-outs. Where possible, we conducted sensitivity analyses.
3)Review focus: a challenge with large reviews is remaining focussed on important key findings. This review team ensured patients and other stakeholders were involved at the protocol stage through to the writing of the conclusions. We did this by including a person with eczema as an author and consulting a patient panel.
4)Different study designs: many of the included studies randomised treatments within a person – i.e. allocating treatments to different sides of the body. We were able to make statistical adjustments to account for the correlation.
5)Transitivity: for an NMA to be clinically and methodologically appropriate, the studies included should be similar. Authors tried to ensure transitivity was met; where it was unachievable, it was considered when assessing the certainty of evidence.
Conclusions
This large-scale NMA provides crucial findings for stakeholders needing information about the effectiveness and safety of topical treatments for eczema. We were able to combine results from nearly 300 studies to provide synthesised, easily accessible evidence. We hope that our experiences will be of use to other review teams.
Network meta-analyses (NMAs) are a methodological innovation, designed to synthesise results from studies of multiple competing interventions. They are vital in topic areas where many different treatments are available, and most trials are placebo-controlled. NMAs reduce research waste by maximising data use from primary studies, eliminating need for further trials.
Objectives
We completed an NMA in December 2023, synthesising data from nearly 300 studies investigating topical anti-inflammatory treatments for eczema. We aim to describe the challenges faced and how we sought to overcome them.
Take-home points
1)Outcome measurements: we used the Core Outcome Set framework to ensure that we measured the key outcomes for stakeholders. However, there was heterogeneity in the way outcomes were measured. One way we overcame this was by using the Standardised Mean Difference for continuous outcomes and re-expressing the results using scales familiar to readers.
2)Outcome completeness: missing dispersion measures meant some continuous outcome data couldn’t be synthesised. There were also missing data on number of included participants and drop-outs. Where possible, we conducted sensitivity analyses.
3)Review focus: a challenge with large reviews is remaining focussed on important key findings. This review team ensured patients and other stakeholders were involved at the protocol stage through to the writing of the conclusions. We did this by including a person with eczema as an author and consulting a patient panel.
4)Different study designs: many of the included studies randomised treatments within a person – i.e. allocating treatments to different sides of the body. We were able to make statistical adjustments to account for the correlation.
5)Transitivity: for an NMA to be clinically and methodologically appropriate, the studies included should be similar. Authors tried to ensure transitivity was met; where it was unachievable, it was considered when assessing the certainty of evidence.
Conclusions
This large-scale NMA provides crucial findings for stakeholders needing information about the effectiveness and safety of topical treatments for eczema. We were able to combine results from nearly 300 studies to provide synthesised, easily accessible evidence. We hope that our experiences will be of use to other review teams.