Dealing with complexity in systematic reviews through outcome categorisation and selection

Article type
Authors
Brennan S1, McKenzie J2
1Australasian Cochrane Centre, Monash University
2Statistical Methods Group, Australasian Cochrane Centre
Abstract
Objectives:
1. To illustrate the complexity that arises from diversity of outcomes in reviews.
2. To explore approaches for dealing with diversity of outcomes that facilitate summary and synthesis.

Description:
Complexity in systematic reviews can arise from several facets: complexity of the interventions; diversity of settings, conditions, and outcomes; and different types of study designs. Systematic reviews evaluating interventions of importance to policy, public health, and health service delivery often have complexity arising in all facets. Systematic reviews of clinical interventions, while often restricted to a specific condition and study design, may have complexity arise in the interventions evaluated, and the diversity of outcomes encountered across the studies.

In this workshop, we will focus on outcomes. Grouping of conceptually similar outcomes into outcome categories (domains), and specifying methods to reduce the potential multiplicity of outcomes that may be encountered within a study, facilitates the ability to undertake a quantitative synthesis. We will discuss logic models (conceptual frameworks) and taxonomies as methods for developing outcome domains, and decision rules for selecting an outcome (when multiple are available) within outcome domains.

The workshop will use small group work and discussion to explore the impact of different approaches on facilitating synthesis and interpretation of review findings.