Context To compare family physicians' ability to retain information when reading
a review article on paper vs on screen, and in their mother tongue vs in English.
Methods Randomized trial of 114 Scandinavian family physicians who read a review
article in October or November 2000 from the Journal of
Trauma for 10 minutes either on paper and in English, on screen and
in English, on paper in their mother tongue, or on screen in their mother
tongue. To assess comprehension, they immediately completed a questionnaire
with 6 open questions about 13 key facts from the review article. Sum score
was on a scale from 0 (no correct answers) to 13 points (all questions answered
correctly).
Results There was no significant difference between readers of paper vs screen
versions, with a median (interquartile range [IQR]) of 4 (2-6) vs 4 (2-5),
respectively (P = .97). Physicians who read in their
mother tongue scored significantly higher than those who read in English,
with a median (IQR) of 4 (3-6) vs 3 (2-4) (P = .01).
Conclusion The medium (paper vs screen) did not influence the ability of family
physicians to retain medical information. They best retained medical information
when reading in their mother tongue.
Two features are characteristic for the developments in medical publishing
recently; a trend toward English as the leading language of publication1 and the introduction of the Internet. Concern has
been raised over the future of national and paper journals.2-4
No comparisons of the effect of reading the same medical text in different
media and languages have been made. The objective of this study was to assess
whether the medium (paper vs screen) or the language (English vs mother tongue)
influenced the ability of Scandinavian family physicians to retain information
from a review article.
A review article from the Journal of Trauma5 on how to handle mild and moderate head injuries was
selected. A questionnaire with 6 open questions about 13 key facts of the
article was prepared. A total score ranging from 0 (no correct answers) to
13 points (all questions answered correctly) could be achieved. Given a difference
of 2 points between means, a significance level of .05, a power of 80%, and
an estimated SD of 3, we needed at least 96 physicians in the analysis (48
in each group). In addition, the questionnaire included questions about the
frequency and preference of reading on paper vs screen, as well as in mother
tongue vs English on a 5-point Likert scale. Year of birth, sex, and nationality
of the physician were reported.
The review article was translated into 3 Scandinavian languages. Paper
and screen versions were prepared pairwise, so that the English and mother
tongue versions had the same layout in each country, equal to the regular
layout of the national medical journal (eg, prepared paper and screen versions
[in Danish and English] of the article in Denmark had the same layout as the
paper or Internet version of the Journal of the Danish Medical
Association). We used laptops with 13-in LCD display and pdf format
in Denmark, and personal computers with 17-in screens and html format in Norway
and Sweden.
Family physicians passing the study room during breaks at the annual
assemblies of the national societies of family medicine in October or November
2000 were asked to participate. Those 122 physicians who agreed were informed
about the aims of the study. Eight declined after having received this information:
114 were included.
By drawing opaque, closed envelopes from a box, the participants were
randomized into 4 groups: to read in their mother tongue on paper vs screen,
or to read in English on paper vs screen. Each participant was given exactly
10 minutes to read the article. The questionnaire was completed immediately
afterward. We used 2-tailed Mann-Whitney U (Wilcoxon)
tests for comparison among groups. The statistical package SPSS version 10.0
(SPSS Inc, Chicago, Ill) was used.
Of the 114 respondents, 2 were non-Scandinavian physicians and 1 did
not complete the questionnaire, leaving 111 physicians (32 Danes, 36 Norwegians,
and 43 Swedes) for the analyses. Age, sex, nationality, and reading habits
are summarized in Table 1.
Table 1. Characteristics of Study Physicians (N = 111)
Table 2 presents the main
results. We found no significant differences in retaining information between
those who had been presented the article on paper vs those who had read it
on screen (median [interquartile range (IQR)], 4 [2-6] vs 4 [2-5]; P = .97). Those who had read the paper in their mother tongue scored
higher than those who had read in English (median [IQR], 4 [3-6] vs 3 [2-4]; P = .01). The tendency to score higher with the mother
tongue applied to all subgroups.
Table 2. Median Score and Interquartile Range (IQR) (Scale 0-13) After Reading a Review Paper for 10 Minutes
The physicians younger than 40 years scored higher than physicians aged
40 years or older (median [IQR], 6 [3.75-8] vs 3 [2-5]; P = .001). The men physicians scored higher than the women physicians
(median [IQR], 4 [3-6] vs 3 [1.5-5]; P = .05).
We deliberately chose a review article and a short period of reading
because it is thought that paper articles are easier to browse than computer
screen articles. The only indication that paper may be easier to read was
observed among the Danish physicians. The quality of the computer screen and
the format of the text may explain the difference because we used pdf versions
on laptops in Denmark. Reading on a computer screen compared with reading
paper does not seem to influence the ability to retain information.
A large proportion of the physicians claimed to read medical information
in English at least weekly. Nevertheless, we found that more of the medical
information was retained if it was read in the mother tongue than in English.
The median difference on our scale represents 25% less retention of information,
which is of practical relevance. Scientific publication in the mother tongue
of the readers offers obvious advantages.
We have not found support in the literature for the sex difference.6,7 The subject of the review article could
possibly have interested the men readers more than the women readers. The
relevance of the information to the reader must be taken into account when
assessing the ability to retain information.
Physicians who attend national assemblies are likely to read more and
be more proficient in English. The difference between reading in the mother
tongue and in English might have been even larger in a representative sample
of family physicians.
1.Sousa Escandón MA, Gonzalez Guitián C, Gonzalez Fernández MM. Which language will Medline speak in the next millenium [in Spanish]?
Arch Esp Urol.2000;53:93-99.Google Scholar 2.Nylenna M, Hagve T-A. Small journals and non-English language journals. In: Godlee F, Jefferson T, eds. Peer Review in
the Health Sciences. London, England: BMJ Books; 1999.
3.Beller FK. The future of German language in science [in German].
Gynakol Geburtshilfliche Rundsch.2000;40:50-54.Google Scholar 4.Delamothe T, Smith R. PubMed Central: creating an Aladdin's cave of ideas.
BMJ.2001;322:1-2.Google Scholar 5.Ingebrigtsen T, Romner B, Kock-Jensen C. Scandinavian guidelines for initial management of minimal, mild, and
moderate head injuries.
J Trauma.2000;48:760-766.Google Scholar 6.Richardson JTE. Gender differences in imagery, cognition, and memory. In: Logie RH, Denis M, eds. Mental Images in Human
Cognition. Amsterdam, the Netherlands: Elsevier; 1991:271-303.
7.Hyde JS, McKinley NM. Gender differences in cognition: results from meta-analyses. In: Caplan PJ. Gender Differences in Human Cognition. New York, NY: Oxford University Press; 1997:30-51.