Comparison of linguistic characteristics of plain language summaries of systematic reviews between medical and non-medical sciences: a cross-sectional study

Date & Time
Wednesday, September 6, 2023, 12:15 PM - 12:25 PM
Location Name
St James
Session Type
Oral presentation
Understanding and using evidence
Oral session
Communicating evidence
Bralić N1, Buljan I2
1Department of Research in Biomedicine and Health, University of Split School of Medicine, Split, Croatia, Croatia
2Department of Psychology, Faculty of Humanities and Social Sciences, University of Split, Split, Croatia, Croatia

Background: Both medical professionals and patients should be able to comprehend health information presented in scientific research. People with low health literacy should be able to access health information easily, with reading level recommendation of maximum 6th-grade reading level. Cochrane is not the only organization developing plain language summaries of systematic reviews (PLSSRs). Although there has been research conducted on medical PLSSRs, little is known about the differences in characteristics between PLSSRs published by medical organizations and those published by non-medical organizations.
Objectives: This cross-sectional study aimed to compare the readability, linguistic, and textual characteristics and conclusiveness of PLSSRs between medical and non-medical sciences.
Methods: We analysed all PLSSRs of the last updated versions of systematic reviews published in English until November 10th, 2022, by Cochrane Collaboration, Campbell Collaboration, and the International Initiative for Impact Evaluation (3ie). PLSs were acquired from organizations’ web pages manually or using the web scraping procedure.
Results: A total of 8,474 medical (by Cochrane) and 163 non-medical (152 by Campbell Collaboration and 11 by 3ie) PLSSRs were included in the analysis. The median years of education needed to understand medical PLSSRs was 15.51 (95% CI: 15.47-15.58), and for non-medical PLSs was 15.23 (95% CI: 14.94-15.50). Non-medical PLSSRs were longer, written with more confidence (median difference [MDiff] 2.43 on a scale of 0-100 [95% CI: 0.08-4.75]) and a more positive emotional tone (MDiff=9.9 [95% CI: 5.94-13.82]) than medical PLSSRs. The next step is to assess the conclusiveness of the PLSSRs using the machine learning approach.
Conclusions: There is a great discrepancy in the number of PLSSRs between medical and non-medical areas. However, even on a small sample, we found that writers of non-medical PLSSRs are more confident, write in a more positive tone, and use more words than writers of medical PLSSRs. Patient, public, and/or healthcare consumer involvement: As this was a methodological study, no patients were directly involved. However, improved readability of plain language summaries of systematic reviews would lead to better decision making, both for patients and other stakeholders.