This is an archived version of the Handbook. For the current version, please go to or search for this chapter here.

15.6.3  Meta-analysis of resource use and cost data

There are currently no agreed-upon methods for pooling combined estimates of cost-effectiveness (e.g. incremental cost-effectiveness, cost-utility or cost-benefit ratios), extracted from multiple economic evaluations, using meta-analysis or other quantitative synthesis methods. However, in principle, if estimates of measures of resource use and costs in a common metric (and associated measures of uncertainty) are available from two or more included studies, for an intervention and its comparator, these can be pooled using a meta-analysis. In practice, extreme caution is advised when considering whether to undertake a meta-analysis of resource use or cost data as part of a Cochrane review. Prior to any decision to pool estimates using a meta-analysis, particular attention should be given to whether the metric in question has equivalent meaning across studies.


Resource use and costs are sensitive to variability across settings, both within a country and between countries, in features of the local context, such as local prices or aspects of service organization and delivery (Drummond 2001, Sculpher 2004). This may limit the generalizability and transferability of estimates of cost, resource use and, by implication, estimates of cost-effectiveness, across settings. It is also the principal reason that resource use and cost data relating to specific target populations and jurisdictions of interest are regarded as the best available source of data for use in economic evaluations to be used in resource allocation decision processes in the specific setting (Cooper 2005). These issues have generated debate on whether meta-analysis of measures of resource use or costs across wider geographical and political boundaries is likely to generate meaningful results, how the results of such meta-analyses should be interpreted and what additional value the results may have for end-users of Cochrane reviews. (Further discussions around issues of applicability and transferability of health economic evaluations can also be found in texts by Hutubessy et al and Kumaranayake and Walker (Kumaranayake 2002, Hutubessy 2003).


On the other hand, whether specific estimates of resource use or costs are generalizable, or transferable, across settings may be regarded as an empirical question. In circumstances where there is evidence of little variation in resource or cost use between studies, it may be regarded as legitimate to present a pooled estimate. Otherwise it is important that the distribution of costs is clearly presented. Many completed Cochrane reviews include meta-analyses of resource use data. A small number of Cochrane reviews include meta-analyses of cost data, although these are not always accompanied by critical appraisal of the methods used to generate these data.


If meta-analyses of resource use or cost data are undertaken in a Cochrane review, this should always be supported by thorough critical appraisal of the methods used to derive such estimates within the corresponding health economics studies (see Sections 15.5.2, and 15.6.2), alongside use of statistical methods to investigate and incorporate between-study heterogeneity (e.g. I2, chi-squared; random-effects models: see Chapter 9, Section 9.5). Cost estimates collected from multiple studies should be adjusted to a common currency and price year before these data are pooled (see also Section 15.6.1). Authors should consult Chapter 9 for further guidance on the statistical procedures underpinning meta-analysis.


If meta-analyses of resource use or cost data are conducted, a narrative summary should be included in the Results section to comment on the direction and magnitude of results and their precision. Similarly, if two or more health economics studies are included in a review, but a decision is taken not to pool (in a meta-analysis) resource use and/or cost data that have been collected from these studies, this can be stated in the Methods section (see Box 15.6.a for an example of this type of statement).