How plain is plain? Jargon and readability in online Plain Language Summaries of health research: observational study (Preprint)

Abstract

The idea of making science more accessible to non-scientists has prompted health researchers to involve patients and the public more actively in their research. One aspect of this is the creation of plain language summaries (PLS), short summaries intended to make research findings more accessible to non-specialists. However, whether PLS satisfy the basic requirements of accessible language is unclear.

Our aim was to assess the readability and level of jargon in the PLS of research funded by the largest national clinical research funder in Europe, the UK’s National Institute for Health and Care Research (NIHR). We also aimed to assess whether these things were influenced by internal and external characteristics of the research projects.

We downloaded the PLS in all NIHR National Journals Library reports from mid-2014 to mid-2022 (n=1241) and analyzed them using the Flesch Reading Ease (FRE) formula and a jargon calculator. Independent variables related to each PLS were research topic, funding programme, project size, length, publication year, and readability and jargon scores of the original funding proposal.

Readability scores ranged from 1.1 to 70.8, with an average FRE score of 39.0 (95% Confidence interval (CI) 38.4 to 39.7). 35 (2.8%) of the PLS had a FRE score classified as “Plain English” or better. Jargon scores ranged from 76.4 to 99.3, with an average score of 91.7 (95% CI 91.5 to 91.9). 269 (21.7%) of the PLS had a jargon score suitable for general comprehension. Variables such as research topic, funding programme, and project size significantly influenced readability and jargon scores. The biggest differences related to scores from the original proposals: proposals with a PLS that were in the 20% most readable were almost three times more likely to have a more readable final PLS (incidence rate ratio (IRR) 2.88, 95% CI 1.86 to 4.45). Those with the 20% least jargon in the original were more than ten times as likely to have low levels of jargon in the final PLS (IRR 13.87, 95% CI 5.17 to 37.2). There was no observable trend over time.

Many of the PLS published in the NIHR’s National Journals Library are hard to read due to their complexity and use of jargon. There are significant variations in readability and jargon scores depending on the research topic, funding program, and other factors. Notably, the readability of the original funding proposal seemed to significantly impact the final report's readability. Ways of improving the accessibility of PLS are needed, as is greater clarity over who and what they are for.