Readability of Open Access and Non–Open Access Articles Reporting Research Studies in Primary Health Care: A Cross-sectional Study

Shuhei Ichikawa,1 Kae Uetani,2,3 Yoshihito Goto,3 Takanori Fujita4


Open access journals can enhance accessibility to the scientific articles. However, publishing in open access journals may not assure enough accessibility unless readers can understand what the article says. If articles in open access journals are more readable for nonprofessional readers, open access journals can enhance accessibility to contents of their articles. Therefore, we investigated whether articles in open access journals are more readable than journals with subscription models via a cross-sectional study.


We identified clinical trials and observational studies published in journals in the primary health care field from April 2010 to December 2016 using PubMed. Journals with 20 or more target articles were extracted, and the most recent 20 articles were analyzed. Journals were divided into 3 groups: full open access, hybrid open access, and completely subscription. The primary outcome was readability in the Introduction section estimated by Gunning Fog score. Gunning Fog score ranges from 6 (level for sixth-grade student) to 17 (level for college graduates); lower score indicates that the text is more readable. Secondary outcomes were number of sentences, number of words, number of words that have 3 or more syllables, mean sentence length, and average syllables per words. Journal Impact Factor was set as a covariate because high-impact journals generally set a more limited article length, which would contribute to more efficient sentences. Country of journal publisher was set as a covariate because familiarity with the English language could affect readability of the articles. Bayesian estimation with Markov Chain Monte Carlo was used to estimate whether articles in full open access journals and hybrid open access journals were more readable than those in subscription journals.


One hundred forty articles were extracted from 7 journals: Ann Fam Med, Fam Med, J Gen Intern Med, Br J Gen Pract, Fam Pract, J Am Board of Fam Med, and BMC Fam Pract. Fog scores and other parameters in each group are shown in the Table. Mean Fog scores in each type of journals was 19, which exceeded the college graduate level (17). Mean (SD) FOG scores in full open-access journals (19.0 [2.1]; estimates 0.69; 95% CI, −0.40 to 1.78) and hybrid open access journals (19.0 [2.5]; estimates, 0.12; 95% CI, −0.96 to 1.23) were not significantly different from subscription journals (mean [SD], 19.6 [2.4]).


We found no difference in readability among full open access, hybrid open access, and completely subscription journals.

1Department of Education and Research in Family and Community Medicine, Mie University Graduate School of Medicine, Mie, Japan, jwellyfish@outlook.com, ichikawa-s@clin.medic.mie-u.ac.jp; 2Medical Communication Department, Statcom Co Ltd, Tokyo, Japan; 3Department of Health Informatics, Kyoto University School of Public Health, Kyoto, Japan; 4KEIO University Platform for Healthcare Industries, Tokyo, Japan

Conflict of Interest Disclosures:

None reported.


Mr Ichikawa received a grant for younger researcher from Mie University in 2016.

Role of the Funder/Sponsor:

Mie University had no role for this research, except for funding.