Secondary review studies

Primary research studies use one of the study types noted above and are the basis for most scientific research, analysis and decisions on interventions.  Many studies have limitations because they have a small patient base, or limited funding, or limited time.  In most cases researchers will duplicate studies done elsewhere so that a search of the literature may find many different studies on the same topic, some with similar results, and others with conflicting results.  Good potential outcomes may be missed by one or two small studies.  Combining studies can therefore lead to a much more powerful analysis of an intervention or outcome.

Searching the literature can be challenging.  Not all publications may be listed in one place, so multiple databases may need to be searched.  Any search will require the use of specific search terms such as ‘firefighter’ and ‘myocardial infarction’.  Searches may vary depending whether the term ‘firefighter’, ‘fire fighter’ ‘fireman’, ‘firemen’ or plural terms are used.  Care must therefore be taken to ensure a standardised approach to the search.  Reviews may be ‘systematic’ where the search process is exhaustive and systematic over multiple databases abd bibliographies, studies are screened for relevance and appraised for quality.  Such studies require substantial expertise and time, and are generally very expensive.  Systematic reviews provide the best quality evidence available.

Combining the results from a number of studies is often termed ‘meta-analysis’.  Specific statistical techniques are used to enable these results to be combined and analysed systematically.  A meta-analysis of randomised controlled trials is the most powerful research that can be undertaken, with power increasing with the number of subjects, number of trials and size of effect.

In many cases a literature search will identify systematic reviews, so there may be little value in undertaking another systematic review.  In other cases, a literature search will identify so few papers that it is unrealistic to undertake a formal systematic review. 

A search protocol may only identify five papers but a more general search may find a couple of book chapters and some further papers with overlapping topics.  This evidence may well be of value, but the process of finding it was not systematic.  The approach taken to finding evidence on this website has been non-systematic. The evidence is considered useful, and where systematic reviews have been found, this evidence is generally very strong and robust.  Readers are expected to note the source of evidence and decide for themselves whether the evidence is strong or weak.