Article type
Year
Abstract
Background:
Properly conducted systematic reviews and meta-analyses are regarded as the highest level of evidence for assessing interventions; however, there are several limitations and potential sources of bias arising from their retrospective nature that jeopardise the validity of their results. Prospectively planned meta-analysis (PMA) is an emerging approach that addresses many of these limitations and sources of bias. To move forward in the development and employment of this methodology, it is important to get an overview of the prevalence of PMA, and the different types of PMA in different fields of medicine.
Objectives:
The aim of this study was to describe the current landscape of PMA in health research by systematically identifying all planned, ongoing and published PMA.
Methods:
We planned to include all studies evaluating an outcome relevant to human health that fulfilled the key features of a PMA as outlined in the Cochrane Handbook. The search strategy was derived by screening previously identified PMA to ascertain the terms used to describe PMA, and by consulting experts in the field. We systematically searched PubMed, Embase, the Cochrane Database of Systematic Reviews and PROSPERO, performed web searches and contacted experts in the field to identify all planned, published or ongoing PMAs. Two review authors independently conducted the full-text screening and extracted data.
Results and conclusions:
We identified 1056 entries for title and abstract screening; 274 of these were included for full-text screening. Most PMA-related publications were published in recent years (69% of entries were published in the last five years). There was great variation in the terminology, reporting and definitions of PMA. Particularly for non-interventional meta-analyses, such as those including genomic association studies, cohort studies and brain imaging studies, the current definitions of PMA are not always sufficient to classify them. We propose more detailed and refined definitions of PMA, and better reporting standards, to improve the overall quality and understanding of PMA.
Patient involvement: None.
Relevance:
The current project aims to describe the landscape of prospectively planned meta-analyses and offer methods to improve these to reduce sources of bias in reviews that inform healthcare guidelines and ultimately improve outcomes for patients and consumers.
Properly conducted systematic reviews and meta-analyses are regarded as the highest level of evidence for assessing interventions; however, there are several limitations and potential sources of bias arising from their retrospective nature that jeopardise the validity of their results. Prospectively planned meta-analysis (PMA) is an emerging approach that addresses many of these limitations and sources of bias. To move forward in the development and employment of this methodology, it is important to get an overview of the prevalence of PMA, and the different types of PMA in different fields of medicine.
Objectives:
The aim of this study was to describe the current landscape of PMA in health research by systematically identifying all planned, ongoing and published PMA.
Methods:
We planned to include all studies evaluating an outcome relevant to human health that fulfilled the key features of a PMA as outlined in the Cochrane Handbook. The search strategy was derived by screening previously identified PMA to ascertain the terms used to describe PMA, and by consulting experts in the field. We systematically searched PubMed, Embase, the Cochrane Database of Systematic Reviews and PROSPERO, performed web searches and contacted experts in the field to identify all planned, published or ongoing PMAs. Two review authors independently conducted the full-text screening and extracted data.
Results and conclusions:
We identified 1056 entries for title and abstract screening; 274 of these were included for full-text screening. Most PMA-related publications were published in recent years (69% of entries were published in the last five years). There was great variation in the terminology, reporting and definitions of PMA. Particularly for non-interventional meta-analyses, such as those including genomic association studies, cohort studies and brain imaging studies, the current definitions of PMA are not always sufficient to classify them. We propose more detailed and refined definitions of PMA, and better reporting standards, to improve the overall quality and understanding of PMA.
Patient involvement: None.
Relevance:
The current project aims to describe the landscape of prospectively planned meta-analyses and offer methods to improve these to reduce sources of bias in reviews that inform healthcare guidelines and ultimately improve outcomes for patients and consumers.