Evidence-based medicine

Editor, – I refer to the article 'Are we there yet? - Travel along the information highway seeking evidence-based medicine' (Aust Prescr 2001;24:116-9). I enjoyed this problem-based article on influenza vaccination but was surprised that the authors did not suggest consulting the Australian Immunisation Handbook as their first search strategy. To solve the problem I pulled the 7th edition (2000) off the shelf, looked up the index on influenza, flicked to page 140, skimmed to recommendation 4 regarding pregnant women on page 144 and found:

'Influenza vaccine is safe for pregnant women. Pregnant women who fall into one of the above risk categories should be vaccinated. In addition, there is evidence from a number of studies that pregnant women, particularly during the second and third trimester, are at increased risk of influenza-associated complications. The US Centers for Disease Control estimates that an average of 1-2 hospitalisations among pregnant women could be prevented for every 1,000 pregnant women immunised. It is therefore recommended that all women who will be in the second or third trimester of pregnancy during the influenza season be vaccinated in advance, so that they will be protected during that period.'

Time: 45 seconds!

The Australian Immunisation Handbook is also available on the internet at: http://www.health.gov.au/pubhlth/immunise/publications.htm (albeit as a 2.6 meg PDF file)!

To me, this exercise shows the clear value of independent immunisation/therapeutic guidelines produced by expert colleagues who have distilled the evidence into authoritative recommendations. It also shows the deficiencies of the Commonwealth Department of Health web search engine which apparently does not currently index their own PDF documents!

Dr Ken Harvey
Senior Lecturer
School of Public Health
La Trobe University
Bundoora, Vic.

Dr Peteris Darzins and Ms Majella Pugh, co-authors of the article, comment:

Dr Ken Harvey reports being surprised that the authors did not suggest consulting the Australian Immunisation Handbook (NHMRC)¹ as their first search strategy. However, not everyone has the latest version of the Handbook on their shelf. Even so, Dr Harvey has overlooked Table 2 which shows that the very first place the medical practitioner conducting the search looked was in the NHMRC web site. It is interesting that the search conducted by browsing the NHMRC web site, and also by using the search terms 'vaccination' and 'guidelines', separately, did not lead to the immunisation guidelines. This shows that information retrieval by electronic means from readily accessible sources is still seriously limited. This may be because the needed information is simply not in the electronic databases or, if it is there, it cannot be readily found by people who are not accustomed to using that particular database.

We agree with Dr Harvey that more attention could be devoted to proper indexing of databases. Poorly indexed databases have a number of deleterious effects. First, they do not provide the information searchers are looking for. Second, they provide a strong negative incentive to searchers to look for information in the databases when next they want to find something. In our opinion, it would be preferable to have fewer, readily accessible, items in the databases, rather than masses of information that is not readily accessible. Proper structuring of databases requires discipline and the active involvement of content experts in deciding what should or should not be included. Many web sites sacrifice function for form and appear to be designed by computing experts without adequate supervision from content experts. It is time those who care about evidence-based medicine invest the required effort to attend to this serious barrier to the optimal provision of health care.