Article type
Year
Abstract
Background: overviews of systematic reviews (overviews) aim to systematically identify and summarize several systematic reviews on a topic. Many published methods papers and guidance describe steps for undertaking overviews of interventions, yet it is uncertain whether these can be applied to other types of overviews. Despite this, other types of overviews are frequently being conducted and used to inform healthcare decision making and policy.
Objectives: to describe the key differences between conducting an overview of interventions compared to an overview that is restricted to observational primary studies.
Methods: we compared guidance for overviews of interventions to a recently conducted overview for an aetiology question. The overview was commissioned by the National Health and Medical Research Council (NHMRC) to inform an update of the 2009 Australian guidance on the health benefits and harms of alcohol consumption. We selected a single review for each outcome. It included 38 systematic reviews (for 53 outcomes) on the health effects of varying levels and/or patterns of alcohol consumption. In addition, it aimed to identify gaps in evidence where no systematic reviews were found for an outcome. The methodological decision process for this overview is shown in the flow chart.
Results: we found additional considerations in a number of steps of the overview: objective, selection criteria, inclusion, quality and certainty of evidence. The selection of a single review for inclusion was complicated by criteria such as the adjustment for confounders and lack of Cochrane Reviews. In fact, some of the arguably ‘best’ evidence available for these types of overviews may come from grey literature sources and evidence that may not follow strict systematic review criteria, for example work conducted by the World Cancer Research Fund. The lack of quality assessment in identified reviews was particularly high, with only 18% that assessed the quality of primary studies. No identified reviews conducted an assessment of the certainty of evidence, so we had to conducted a full assessment for each outcome.
Conclusions: we have reported the different methodology considerations for undertaking an overview including observational primary studies, compared to an overview of interventions. We hope this is useful for those undertaking similar overviews, particularly those intending to inform public health guidance and policy.
Patient or healthcare consumer involvement: decision makers often use overviews to address broad research questions. These are usually developed by working across different interest groups, including patients and consumers, researchers, commissioners and subject matter experts. Due consideration of the challenges and trade-offs inherent in this methodological approach, is important to ensure the needs of all stakeholders, including patients and consumers, are met.
Objectives: to describe the key differences between conducting an overview of interventions compared to an overview that is restricted to observational primary studies.
Methods: we compared guidance for overviews of interventions to a recently conducted overview for an aetiology question. The overview was commissioned by the National Health and Medical Research Council (NHMRC) to inform an update of the 2009 Australian guidance on the health benefits and harms of alcohol consumption. We selected a single review for each outcome. It included 38 systematic reviews (for 53 outcomes) on the health effects of varying levels and/or patterns of alcohol consumption. In addition, it aimed to identify gaps in evidence where no systematic reviews were found for an outcome. The methodological decision process for this overview is shown in the flow chart.
Results: we found additional considerations in a number of steps of the overview: objective, selection criteria, inclusion, quality and certainty of evidence. The selection of a single review for inclusion was complicated by criteria such as the adjustment for confounders and lack of Cochrane Reviews. In fact, some of the arguably ‘best’ evidence available for these types of overviews may come from grey literature sources and evidence that may not follow strict systematic review criteria, for example work conducted by the World Cancer Research Fund. The lack of quality assessment in identified reviews was particularly high, with only 18% that assessed the quality of primary studies. No identified reviews conducted an assessment of the certainty of evidence, so we had to conducted a full assessment for each outcome.
Conclusions: we have reported the different methodology considerations for undertaking an overview including observational primary studies, compared to an overview of interventions. We hope this is useful for those undertaking similar overviews, particularly those intending to inform public health guidance and policy.
Patient or healthcare consumer involvement: decision makers often use overviews to address broad research questions. These are usually developed by working across different interest groups, including patients and consumers, researchers, commissioners and subject matter experts. Due consideration of the challenges and trade-offs inherent in this methodological approach, is important to ensure the needs of all stakeholders, including patients and consumers, are met.
PDF