To the Editor:—
We have read with interest the letter of Hofbauer et al. 1about research in anesthesiology in Europe. This group has made important clinical and experimental contributions. However, their letter has several flaws.
First, it is remarkable that Hofbauer et al. give no references, whereas several articles are published on this topic. 2–6
Second, their claim to have searched 1965 to September 1999 is spurious. Address fields in MEDLINE are only available for publications added to the database after January 1988, equivalent to articles published after mid 1987 in most journals. 7
Third, they rely on identifying anesthesia departments and countries in MEDLINE address fields rather than analyzing all articles in certain journals as done by most other groups. 2–5This may be advantageous because (in Germany) only approximately 70% of the work of anesthesiologists is published in journals devoted to anesthesia or related fields, such as pain and critical care. 8,9Also, only 90% of the articles in major anesthesia journals and one third in pain and critical care are by anesthesiologists (our own unpublished observations). In practice, this is difficult because (1) addresses are often incomplete; (2) addresses are often not in English, even for English-language articles; (3) many abbreviations and most e-mails are useless for searching; (4) information may be implicit (e.g. , region instead of country); (5) MEDLINE sometimes gives more than one department in the address; (6) for many journals in the Russian and Chinese languages (e.g., Anesteziologiia i Reanimatologiia ), MEDLINE does not give address fields; and (7) misspellings occur. For publications by German anesthesia departments in the English language, one can use (ANESTH* OR ANAESTH* OR ANASTH*) AND (GERMAN OR GERMANY OR DEUTSCHE* OR DEUTSCHLAND OR FRG OR BRD OR DDR OR GDR) with about 95% sensitivity and 99% positive predictive value. This results in at least 1,379 English-language publications by German university anesthesiologists from 1988 to 1997. 8There were at least 120 English-language publications by German nonuniversity anesthesiologists in 1988–1997 (own data). Extrapolation for mid 1987 through September 1999 yields approximately 1,800 publications by German anesthesia departments in the English language alone. There are (extrapolated) some 3,000 in these years in the German language. 9The data by Hofbauer et al. of 1,605 publications from Germany in mid 1987 through September 1999 is thus only one third of the total 4,800 because of (1) a hidden language restriction because few German-language publications by Germans explicitly state the nation in the address and (2) omission of the search item ANASTH*, the MEDLINE transcription of the German umlaut in Anästhesie .
Fourth, Hofbauer et al. list a country named England that does not exist. They probably mean the United Kingdom (UK). The UK is difficult to access by MEDLINE searching. Even the search (UK OR (UNITED KINGDOM) OR (GREAT BRITAIN) OR (ENGLAND NOT NEW) OR (WALES NOT (NEW OR PRINCE)) OR (NORTHERN IRELAND) OR SCOTLAND) is less than 50% sensitive for the years in question, necessitating additional city searching. The reason is omission of explicit country identification in domestic (British) journals. The situation has recently improved because most British journals now add UK to British addresses (e.g. , Lancet since 1990, Anaesthesia since 1997, British Journal of Anaesthesia since 1999), but others (e.g. , BMJ ) still maintain that all addresses are British unless stated otherwise.
Fifth, similar problems exist for other countries. In some languages (e.g. , Italian, Spanish), the specialty search term is ANESTE* without an “h.” In WinSPIRS 10(but not PubMed 11), French addresses must be sought as D'ANESTH*. Search routines for various nations—which, however, contain several obvious errors and which we have not tested—can be found in Jorgensen et al. 12
Sixth, Hofbauer et al. use the same raw data for evaluating per capita and per medical school output. This is incorrect because there are publications by nonuniversity departments. For Germany, the amount is approximately 8% of the English- and 20% of the German-language articles. Also, the size of medical schools depends on the population served in education and treatment. For example, because Austria has three medical schools for 8.1 million inhabitants, each serves on average 2.7 million inhabitants. In Switzerland, which has five medical schools (not four, as stated by Hofbauer et al. ) for 7.3 million inhabitants, each serves only 1.5 million people, with Germany in between. Consequently, the Austrian schools (or at least one of them) are likely to be large. This is true for the Vienna school (where Hofbauer et al. work), which in the mid 1990s had approximately 200 physicians and Ph.D.s in its anesthesia department, which to our knowledge was matched by only 1 of the 41 German and none of the 5 Swiss departments. 9,13
Finally, in an article published in a North American journal, such as Anesthesiology, it would have been interesting to give numbers for the United States and Canada for comparison.
To clarify this, we printed all MEDLINE-indexed publications by anesthetists in 1998 and analyzed them by hand (table 1). Although this is only an evaluation for 1 yr, it clarifies the following: (1) Non–English-language publications were still important in the European Union, being 27% overall and even higher for the large non–English-speaking nations (Spain, France, Germany, Italy). (2) Publication numbers by Hofbauer et al. for England–UK in 1987–1999 are impossibly low because the UK surpassed them in 1998 alone. (3) Our tableconfirms the good standing of the Scandinavian and Alpine nations, but the quotient between them and the large European Union nations was much less than what was suggested by Hofbauer et al. (4) Outside Europe, Singapore was the leader in anesthesia with respect to per capita output, although most of its anesthesia publications were published in local journals, such as Singapore Medical Journal . (5) The per capita output of both the USA and Canada matched that of the European Union as a whole closely.