A paper by Oster in Medical Care back in 1990 looked at this important topic. They are writing in a USA context but the findings appear to be of interest to us in the UK.

In recent years, many new over-the-counter (OTC) medications have resulted from the granting of OTC status to drug entities that previously were available only by prescription (Rx). While the benefits to consumers of Rx-to-OTC switches may be substantial, they also involve some degree of risk, as usage typically expands and physician supervision diminishes. Oster’s study explored the balance of benefits and risks. The study used the technique of decision analysis to examine drug regulatory policy.

Histamine H2 receptor antagonists (H2-blockers), which were available only by prescription, were presented as a case study and were examined to determine how OTC availability of these agents would be likely to alter the patterns, effectiveness, and risks of self-treatment for acid-peptic disorders.

A team of clinicians at the Centre for Evidence-Based Medicine in Oxford,
has set up "Evidence-Based On-Call" to meet the challenge of providing clinically relevant evidence-based advice at the bedside.

The Oxford team will use high-quality techniques to search the literature for the best journal articles on specific on-call problems. EBOC has developed an internet authorising tool allowing clinicians with EBM experience to create one-page summaries of these articles.

Clinicians from all specialities and at all levels of experience may participate in Evidence-Based On-Call.

The group are looking for contributors and reviewers:

Their website at http://cebm.jr2.ox.ac.uk/eboc/eboc.html gives a sample guideline and some CATs. Contact Chris Ball at:

Back to top

Andrew Booth has recently given a useful outline of what makes a systematic review.

He suggests there are four components to the review that make it systematic:

1. The literature search

2. The selection of the materials

3. The appraisal

4. The synthesis

A bona fide systematic review is only such if all elements have been done systematically - the methodology is not therefore study dependent (i.e. not necessarily RCTs only). The considerations that come into play are as follows:

1. All systematic reviews should give evidence of the widest possible strategy for identifying studies. This is done by listing all databases searched, strategies where available, performing funnel plots to look for publication bias etc.

With a qualitative review one obviously cannot prove that one has a statistically valid sample so one would look for the presence of data saturation - i.e. to search the literature persistently until you no longer find any new types of models. Citation searching is particularly important in this context.

2. Selection of materials should be done according to an a priori (i.e. up front) set of criteria. This includes a precise definition of studies to be included. Selection of criteria after retrieving articles is methodologically suspect.

4. The added value of the review, when done systematically, is in bringing together comparable studies and commenting on commonalties and differences. Guidance on how to do this for qualitative studies is found in a useful publication Metaethnography (full ref in Cochrane Review Methodology Database).

If your review meets the above then it is arguably a systematic review. If not then you could describe it according to those characteristics that it possesses, e.g. a systematic search and qualitative review, systematic search and concept analysis etc.

Often the problem lies in the synthesis if you are trying to analyse different sorts of studies - this will only be possible if you have subgroups of similar enough studies. "The mixing of diverse studies can make for a strange fruit salad: mixing apples and oranges may seem reasonable enough but when sprouts, turnips or even an old sock are added it can cast doubt on the meaning of any aggregate estimates".

This comes from the context of meta-analysis but the same principles apply.