This is an archived version of the Handbook.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

13.7.2  Evaluating the strength of evidence provided by reviews that include non-randomized studies

‘Exposing’ the evidence from NRS on a particular health question enables informed debate about its meaning and importance, and the certainty which can be attributed to it. Critically, there needs to be a debate about the chance that the observed findings could be misleading. Formal hierarchies of evidence all place NRS low down on the list, but above those of clinical opinion (Eccles 1996, National Health and Medical Research Council 1999, Oxford Centre for Evidence-based Medicine 2001). This emphasizes the general concern about biases in NRS, and the difficulties of attributing causality to the observed effects. The strength of evidence provided by a systematic review of NRS is likely to depend on meeting the challenges set out in Section 13.7.1. The ability to meet these challenges will vary with healthcare context and outcome. In some contexts little confounding is likely to occur. For example, little prognostic information may be known when infants are vaccinated, limiting possible confounding (Jefferson 2005).


Whether the debate concludes that there is a need for randomized trials or that the evidence from NRS is adequate for informed decision-making will depend on the cost placed on the uncertainty arising through use of potentially biased study designs, and the collective value of the observed effects. This value may depend on the wider healthcare context. It may not be possible to include assessments of the value within the review itself, and it may become evident only as part of the wider debate following publication.


For example, is evidence from NRS of a rare serious adverse effect adequate to decide that an intervention should not be used? The evidence is uncertain (due to a lack of randomized trials) but the value of knowing that there is the possibility of a potentially serious harm is considerable, and may be judged sufficient to withdraw the intervention. (It is worth noting that the judgement about withdrawing an intervention may depend on whether equivalent benefits can be obtained from elsewhere without such a risk; if not, the intervention may still be offered but with full disclosure of the potential harm.) Where evidence of benefit is not based on randomized trials and is therefore equivocal, the value attached to a systematic review of NRS of harm may be even greater.


In contrast, evidence of a small benefit of a novel intervention from a systematic review of NRS may not be sufficient for decision makers to recommend widespread implementation in the face of the uncertainty of the evidence and the substantial costs arising from provision of the intervention. In these circumstances, decision makers are likely to conclude that randomized trials should be undertaken if practicable and if the investment in the trial is likely to be repaid in the future.


The GRADE scheme for assessing the quality of a body of evidence is recommended for use in ‘Summary of findings’ tables in Cochrane reviews, and is summarized in Chapter 12 (Section 12.2). There are four quality levels: ‘high’, ‘moderate’, ‘low’ and ‘very low’. A collection of studies that can be crudely categorized as randomized trials starts at the highest level, and may be downgraded due to study limitations (risk of bias), indirectness of evidence, heterogeneity, imprecision or publication bias. Collections of observational studies start at a level of ‘low’, and may be upgraded due to a large magnitude of effect, lack of concern about confounders or a dose-response gradient. Review authors will need to make judgements about whether evidence from NRS should be upgraded from a low level or possibly (e.g. in the case of quasi-randomized trials) downgraded from a high level.