Moving Beyond Readability Metrics for Health-Related Text Simplification

Area of Science:

• Health Communication
• Natural Language Processing
• Medical Informatics

Background:

• Limited health literacy is a significant barrier to understanding crucial health information.
• Existing text simplification tools often lack demonstrated impact on patient comprehension.
• Health disparities can be exacerbated by difficulties in accessing and understanding health content.

Purpose of the Study:

• To develop and present an evidence-based strategy for creating effective text simplification software.
• To introduce a novel semi-automated text simplification tool leveraging natural language processing.
• To evaluate the effectiveness of traditional readability metrics versus a new metric for assessing word difficulty in healthcare contexts.

Main Methods:

• Utilized modern data sources and natural language processing algorithms for tool development.
• Implemented an evidence-based strategy for designing text simplification software.
• Conducted a comparative study of syllable count and a new metric, 'term familiarity', for measuring word difficulty.

Main Results:

• Initial results from the text simplification tool development are promising.
• Syllable count, commonly used in readability formulas, measures perceived word difficulty but not actual difficulty.
• 'Term familiarity' was found to be a more accurate measure of actual word difficulty.

Conclusions:

• A semi-automated text simplification tool is being developed using a novel, evidence-based approach.
• The 'term familiarity' metric offers a more accurate assessment of word difficulty in health information compared to syllable count.
• This work has the potential to reduce health disparities by improving health literacy through simplified text.