MetaPROM – Enhancing the interpretation of Meta-analyses of Patient-Reported Outcome Measures through a Microsoft Excel-based statistical software program

ID: 

18550

Session: 

Short oral session 2: Considerations for meta-analyses

Date: 

Wednesday 13 September 2017 - 11:00 to 12:30

Location: 

All authors in correct order:

Devji T1, Guyatt G1, Svendrovski A2, da Costa B3, Johnston B1
1 McMaster University, Canada
2 Systematic Overviews through advancing Research Technology, Canada
3 University of Bern, Switzerland
Presenting author and contact person

Presenting author:

Tahira Devji

Contact person:

Abstract text
Background: Interpretation of the magnitude of treatment effects for most continuous outcomes and, particularly for patient-reported outcome measures (PROMs), presents challenges. Having decided which PROMs to include in a meta-analysis, review authors must ensure the results they present are optimally interpretable to their target audiences. RevMan is currently limited to two presentation formats when performing meta-analysis of PROM data (i.e. mean difference (MD) and standardised mean difference (SMD)). Although statistical software such as Stata and R offer packages to perform meta-analysis with greater flexibility, allowing investigators to compute pooled estimates using alternative presentation formats that may enhance interpretability, these software can be complex for the average systematic reviewer, as they require programming and advanced statistical knowledge.

Objectives: To summarise available presentation approaches embedded within a novel Microsoft Excel-based tool for enhancing the interpretability of pooled estimates of PROMs.

Methods: MetaPROM performs fixed and random effects meta-analysis for continuous PROM data, and is particularly useful when the included trials report results using different PROMs. MetaPROM facilitates the use of a series of common and emerging statistical presentation formats including SMD, MD in natural units of the most familiar instrument, MD in MID units, ratio of means, relative risk, odds ratio, risk difference and the number needed to treat.

Results: We illustrate the application of these approaches in meta-analysis of PROM data with an example using data from a systematic review of paroxetine vs. placebo for the treatment of major depression. We discuss the relative merits and limitations of each alternative and offer guidance for meta-analysts and guideline developers.

Conclusions: MetaPROM offers various presentation approaches to enhance interpretability of pooled estimates of PROMs using flexible, user-friendly, and soon to be widely available software.