Lumping versus splitting in systematic reviews: Feasibility for researchers versus relevance for practice and policy?




Poster session 1 Wednesday: Evidence production and synthesis


Wednesday 13 September 2017 - 12:30 to 14:00


All authors in correct order:

De Buck E1, Van Remoortel H2, Vande veegaete A2, Govender T3, Vandekerckhove P4, Young T5
1 Centre for Evidence-Based Practice Belgian Red Cross; Department of Public Health and Primary Care, Faculty of Medicine, KU Leuven, Belgium
2 Centre for Evidence-Based Practice Belgian Red Cross, Belgium
3 Division of Health Systems and Public Health, Stellenbosch University, South Africa
4 Belgian Red Cross; Department of Public Health and Primary Care, Faculty of Medicine, KU Leuven; Faculty of Medicine and Health Sciences, Ghent University, Belgium
5 Centre for Evidence-Based Healthcare, Stellenbosch University, South Africa
Presenting author and contact person

Presenting author:

Emmy De Buck

Contact person:

Abstract text
Background: A clearly defined research question is key in developing a systematic review (SR). However, within international development there is a tendency to work with very broad SR questions. This consideration in defining research questions is known as 'splitting versus lumping', where 'splitting' is focusing on a single, well-defined intervention, and 'lumping' broadens the scope at the intervention, outcome and study type level.

Objectives: To reflect on the pros and cons of lumping versus splitting in a mixed-methods SR on the effectiveness and implementation of WASH (water, sanitation, hygiene) promotion programmes to promote behaviour change in low- and middle-income countries.

Methods: The researchers’ perspective was considered based on total time spent to develop the SR, the number of included studies, the number of outcomes for which data were extracted. The field perspective was obtained by 2 face-to-face consultations with funders, field practitioners and policy makers: one during the protocol phase, and one after the SR results were analysed/synthesised.

Results: From the researchers’ perspective, the following favours splitting: time and resource availability, and capacity to deal with complexity at various stages in the SR. From the field perspective there is more often a tendency to lumping for innovation (the SR should not only confirm what is already known from practice), correspondence with real-life situation (in reality not 1 isolated intervention is implemented), and relevance of factors influencing implementation. An argument for splitting would be the simplicity of the analysis and presentation of results.

Conclusions: It is crucial for researchers to consult with different stakeholders beforehand if they want to develop a policy-relevant SR. However, this should be balanced against time and resources available.