• Edit
• Help

Improving Health Outcomes through Computer Generation of Reader-Appropriate Texts

Many studies have shown that the readability of the health information provided to consumers does not match their reading levels. While the average US adult is known to read at 7^th grade level, most health instructions tend to require a 10^th grade reading level or higher. This project aims to address this problem, by developing software tools that can provide more readable health information through automatic translation of text. The domain focus of the project will be diabetes self-care materials.

More specifically this project aims to:

1. Develop tools to automatically assess the readability of health texts.
2. Develop tools to translate complex health texts to more comprehensible versions targeted at consumers of different readability levels, through the use of vocabulary knowledge and grammar transformation rules.

Team Members

• Qing Zeng-Treitler (PI)
• Lucila Ohno-Machado (Co-PI)
• John Guttag (Subcontract PI)
  Professor, Dept. of EECS, MIT
• Dorothy Curtis
  Research Scientist, CSAIL, MIT
• Sasikiran Kandula
  Software Engineer, DSG
• Freddy Bafuka
  Master's Candidate, MIT
• Chantal Friedman
  Administrator, DSG

Software Tools

Preliminary versions of the software tools listed above are currently under development.

Readability Tool
This GUI-enabled tool can be used to assess the readability of a document. In addition to standard readability formulae like Flesh-Kincaid grade level and SMOG, this tool implements a readability assessment method developed in-house. This technique evaluates the document's readability on the basis of the document's syntactic and semantic features, use of medical terminology, cohesion and physical layout. Sample screenshots can be viewed here.

Translation Tool
This tool provides a simplified version of the health-related text by providing more comprehesible alternatives(individual terms or phrases) for the target audience. The current version of the tool also gives Spanish language alternatives to hard medical terms. Sample screenshots can be viewed here.

Publications

• Q Zeng-Treitler , Goryachev S, Kim H, Keselman A, Rosendale S. Making Texts in Electronic Health Records Comprehensible to Consumers: A Prototype Translator. In: AMIA Fall Symposium, 2007.
• Kim H , Goryachev S, Rosemblat G, Browne A, Keselman A, Zeng-Treitler Q. Beyond Surface Characteristics: A New Health Text-Specific Readability Measurement. In: AMIA Fall Symposium, 2007.
• Kandula S, Zeng-Treitler Q. Creating a Gold Standard for the Readability Measurement of Health Texts. In: AMIA Fall Symposium, 2008