Browsing by keyword "readability"
Now showing items 1-3 of 3
-
Assessing COVID-19 Health Information on Google Using the Quality Evaluation Scoring Tool (QUEST): Cross-sectional and Readability AnalysisBackground: The COVID-19 pandemic spurred an increase in online information regarding disease spread and symptomatology. Objective: Our purpose is to systematically assess the quality and readability of articles resulting from frequently Google-searched COVID-19 terms in the United States. Methods: We used Google Trends to determine the 25 most commonly searched health-related phrases between February 29 and April 30, 2020. The first 30 search results for each term were collected, and articles were analyzed using the Quality Evaluation Scoring Tool (QUEST). Three raters scored each article in authorship, attribution, conflict of interest, currency, complementarity, and tone. A readability analysis was conducted. Results: Exactly 709 articles were screened, and 195 fulfilled inclusion criteria. The mean article score was 18.4 (SD 2.6) of 28, with 7% (14/189) scoring in the top quartile. National news outlets published the largest share (70/189, 36%) of articles. Peer-reviewed journals attained the highest average QUEST score compared to national/regional news outlets, national/state government sites, and global health organizations (all P<.05). The average reading level was 11.7 (SD 1.9, range 5.4-16.9). Only 3 (1.6%) articles were written at the recommended sixth grade level. Conclusions: COVID-19-related articles are vastly varied in their attributes and levels of bias, and would benefit from revisions for increased readability.
-
Assessing the Readability of Medical Documents: A Ranking ApproachBACKGROUND: The use of electronic health record (EHR) systems with patient engagement capabilities, including viewing, downloading, and transmitting health information, has recently grown tremendously. However, using these resources to engage patients in managing their own health remains challenging due to the complex and technical nature of the EHR narratives. OBJECTIVE: Our objective was to develop a machine learning-based system to assess readability levels of complex documents such as EHR notes. METHODS: We collected difficulty ratings of EHR notes and Wikipedia articles using crowdsourcing from 90 readers. We built a supervised model to assess readability based on relative orders of text difficulty using both surface text features and word embeddings. We evaluated system performance using the Kendall coefficient of concordance against human ratings. RESULTS: Our system achieved significantly higher concordance (.734) with human annotators than did a baseline using the Flesch-Kincaid Grade Level, a widely adopted readability formula (.531). The improvement was also consistent across different disease topics. This method's concordance with an individual human user's ratings was also higher than the concordance between different human annotators (.658). CONCLUSIONS: We explored methods to automatically assess the readability levels of clinical narratives. Our ranking-based system using simple textual features and easy-to-learn word embeddings outperformed a widely used readability formula. Our ranking-based method can predict relative difficulties of medical documents. It is not constrained to a predefined set of readability levels, a common design in many machine learning-based systems. Furthermore, the feature set does not rely on complex processing of the documents. One potential application of our readability ranking is personalization, allowing patients to better accommodate their own background knowledge.
-
Readability Formulas and User Perceptions of Electronic Health Records Difficulty: A Corpus StudyBACKGROUND: Electronic health records (EHRs) are a rich resource for developing applications to engage patients and foster patient activation, thus holding a strong potential to enhance patient-centered care. Studies have shown that providing patients with access to their own EHR notes may improve the understanding of their own clinical conditions and treatments, leading to improved health care outcomes. However, the highly technical language in EHR notes impedes patients' comprehension. Numerous studies have evaluated the difficulty of health-related text using readability formulas such as Flesch-Kincaid Grade Level (FKGL), Simple Measure of Gobbledygook (SMOG), and Gunning-Fog Index (GFI). They conclude that the materials are often written at a grade level higher than common recommendations. OBJECTIVE: The objective of our study was to explore the relationship between the aforementioned readability formulas and the laypeople's perceived difficulty on 2 genres of text: general health information and EHR notes. We also validated the formulas' appropriateness and generalizability on predicting difficulty levels of highly complex technical documents. METHODS: We collected 140 Wikipedia articles on diabetes and 242 EHR notes with diabetes International Classification of Diseases, Ninth Revision code. We recruited 15 Amazon Mechanical Turk (AMT) users to rate difficulty levels of the documents. Correlations between laypeople's perceived difficulty levels and readability formula scores were measured, and their difference was tested. We also compared word usage and the impact of medical concepts of the 2 genres of text. RESULTS: The distributions of both readability formulas' scores (P < .001) and laypeople's perceptions (P=.002) on the 2 genres were different. Correlations of readability predictions and laypeople's perceptions were weak. Furthermore, despite being graded at similar levels, documents of different genres were still perceived with different difficulty (P < .001). Word usage in the 2 related genres still differed significantly (P < .001). CONCLUSIONS: Our findings suggested that the readability formulas' predictions did not align with perceived difficulty in either text genre. The widely used readability formulas were highly correlated with each other but did not show adequate correlation with readers' perceived difficulty. Therefore, they were not appropriate to assess the readability of EHR notes.
