Giving European SLRs the upper hand: Are manual searches of local journals necessary?

Article type
Authors
Bobrowska A1, Evans JS1, Steeves S1
1Costello Medical
Abstract
Background:
Literature reviews in medicine and healthcare are usually performed by searching English language databases such as MEDLINE and Embase. As a consequence, there may be a bias towards English language publications in literature reviews that have not supplemented such database searches with country or language-specific searches. A possible method of accounting for this includes hand-searching of local journals.
Objectives:
We investigated the extent to which hand-searches of European journals might be advantageous alongside English language database searches.
Methods:
A list of journals categorised under the "Medicine" subject area in the Scimago Journal & Country Rank database and with a 2018 ranking was retrieved for EU/EEA countries and Switzerland. For countries with >100 journals, we identified which journals were not indexed in MEDLINE. The country with the most non-indexed journals was analysed to see which subject categories those journals were assigned to. The overall count for each category was then calculated to determine whether any specific areas would be particularly susceptible to data being missed (in that country).
Results:
We found 7 countries with >100 journals: Netherlands (585), Germany (498), Switzerland (196), France (172), Italy (164), Spain (157) and Poland (117). The proportion of journals not indexed in MEDLINE ranged from 4 to 16% in all countries but Germany, where 33% (162/498) journals were not indexed in MEDLINE. The German non-indexed journals were classified under 38 different categories within the “Medicine” subject area (each journal could be assigned more than 1 category). The most represented individual categories were "Medicine (miscellaneous)" (40), "Surgery" (17) and "Public health, environmental and occupational health" (17). When grouped thematically, the most represented topic areas were public health (21), internal and emergency medicine (20), surgery (17), psychiatry (16) and orthopaedics/sports medicine (16). These non-indexed journals made up 29 to 56% of all German journals in these topic areas and 5 to 9% of all European Journals indexed in Scimago in these topic areas.
Limitations:
Firstly, we based our list of journals on those indexed in Scimago, which may itself not represent all available journals. Secondly, our analysis did not consider the relative contribution that non-indexed journals make to the literature, either in terms of the number of articles published in these journals or their quality. Nevertheless, we suggest that our results give an indication of where it may be valuable to include additional journal hand-searches in literature reviews.
Conclusions:
Our results indicate that over most European journals, the majority are indexed in MEDLINE and failing to conduct country-level specific searches is in general unlikely to result in large numbers of articles being missed. However, specialised searches may be necessary where a review’s scope is limited to certain countries or on topics where there is a higher proportion of non-indexed journals.