Assessing the Readability of Medical Documents: A Ranking Approach

An Analysis of the Readability of Online Sarcoidosis Resources

Introduction Sarcoidosis is an inflammatory disease characterized by the formation of noncaseating granulomas in multiple organ systems. The presentation can vary widely; although some patients with sarcoidosis can be asymptomatic, sarcoidosis can also present in others with symptomatic multiorgan system involvement. Considering the potential severity of the disease, patients need to be well-informed about sarcoidosis to better manage their health. This study aims to assess the readability levels of online resources about sarcoidosis. Methods We conducted a retrospective cross-sectional study. The term "sarcoidosis" was searched online using both Google and Bing to find websites written in English. Each website was categorized by type: academic, commercial, government, nonprofit, and physician. The readability scores for each website were calculated using six different readability tests: the Flesch-Kincaid reading ease (FKRE), Flesch-Kincaid grade level (FKGL), Gunning fog score (GFS), Simple Measure of Gobbledygook (SMOG), automated readability index (ARI), and Coleman-Liau index (CLI). FKRE gives a score that corresponds to the difficulty of the text, while the remaining tests give a score that corresponds to a grade level in terms of reading ability. A one-sample t-test was used to compare all test scores with the national recommended standard of a sixth-grade reading level. Our null hypothesis was that the readability scores of the websites searched would not differ statistically significantly from the sixth-grade reading level and that there would be no significant differences across website categories. To evaluate the difference between the categories of websites, ANOVA testing was used. Results Thirty-four websites were analyzed. Each of the six readability tests for the websites had an average score, which corresponded to being significantly harder to read than the nationally recommended sixth-grade reading level (p<0.001). None of the mean readability scores showed a statistically significant difference across the five different website categories. Conclusions This is the first study, to our knowledge, to examine the readability of online English resources on sarcoidosis and calculate standardized readability scores for them. It implies that the online English material for sarcoidosis is above the health literacy recommended reading levels for patients. There is a need to simplify the material to be easier to read for patients.

Assessing the Readability of Online Patient Education Materials in Obstetrics and Gynecology Using Traditional Measures: Comparative Analysis and Limitations

BACKGROUND

Patient education materials (PEMs) can be vital sources of information for the general population. However, despite American Medical Association (AMA) and National Institutes of Health (NIH) recommendations to make PEMs easier to read for patients with low health literacy, they often do not adhere to these recommendations. The readability of online PEMs in the obstetrics and gynecology (OB/GYN) field, in particular, has not been thoroughly investigated.

OBJECTIVE

The study sampled online OB/GYN PEMs and aimed to examine (1) agreeability across traditional readability measures (TRMs), (2) adherence of online PEMs to AMA and NIH recommendations, and (3) whether the readability level of online PEMs varied by web-based source and medical topic. This study is not a scoping review, rather, it focused on scoring the readability of OB/GYN PEMs using the traditional measures to add empirical evidence to the literature.

METHODS

A total of 1576 online OB/GYN PEMs were collected via 3 major search engines. In total 93 were excluded due to shorter content (less than 100 words), yielding 1483 PEMs for analysis. Each PEM was scored by 4 TRMs, including Flesch-Kincaid grade level, Gunning fog index, Simple Measure of Gobbledygook, and the Dale-Chall. The PEMs were categorized based on publication source and medical topic by 2 research team members. The readability scores of the categories were compared statistically.

RESULTS

Results indicated that the 4 TRMs did not agree with each other, leading to the use of an averaged readability (composite) score for comparison. The composite scores across all online PEMs were not normally distributed and had a median at the 11th grade. Governmental PEMs were the easiest to read amongst source categorizations and PEMs about menstruation were the most difficult to read. However, the differences in the readability scores among the sources and the topics were small.

CONCLUSIONS

This study found that online OB/GYN PEMs did not meet the AMA and NIH readability recommendations and would be difficult to read and comprehend for patients with low health literacy. Both findings connected well to the literature. This study highlights the need to improve the readability of OB/GYN PEMs to help patients make informed decisions. Research has been done to create more sophisticated readability measures for medical and health documents. Once validated, these tools need to be used by web-based content creators of health education materials.

Automated Strategy Feedback Can Improve the Readability of Physicians' Electronic Communications to Simulated Patients

Modern communication between health care professionals and patients increasingly relies upon secure messages (SMs) exchanged through an electronic patient portal. Despite the convenience of secure messaging, challenges include gaps between physician and patient expertise along with the asynchronous nature of such communication. Importantly, less readable SMs from physicians (e.g., too complicated) may result in patient confusion, non-adherence, and ultimately poorer health outcomes. The current simulation trial synthesizes work on patient-physician electronic communication, message readability assessments, and feedback to explore the potential for automated strategy feedback to improve the readability of physicians' SMs to patients. Within a simulated secure messaging portal featuring multiple simulated patient scenarios, computational algorithms assessed the complexity of SMs written by 67 participating physicians to patients. The messaging portal provided strategy feedback for how physician responses might be improved (e.g., adding details and information to reduce complexity). Analyses of changes in SM complexity revealed that automated strategy feedback indeed helped physicians compose and refine more readable messages. Although the effects for any individual SM were slight, the cumulative effects within and across patient scenarios showed trends of decreasing complexity. Physicians appeared to learn how to craft more readable SMs via interactions with the feedback system. Implications for secure messaging systems and physician training are discussed, along with considerations for further investigation of broader physician populations and effects on patient experience.

Internet-Based Patient Education Materials Regarding Diabetic Foot Ulcers: Readability and Quality Assessment

BACKGROUND

While diabetic foot ulcers (DFU) are a common complication of diabetes, little is known about the content and readability of online patient education materials (PEM) for DFU. The recommended reading grade level for these materials is grades 6-8.

OBJECTIVE

The aim of this paper was to evaluate the quality and readability of online PEM on DFU.

METHODS

A Google search was performed using 4 different search terms related to DFU. Two readability formulas were used to assess the readability of the included PEM. These included the Flesch-Kincaid grade level and the Flesch-Reading ease score. The DISCERN tool was used to determine quality and reliability.

RESULTS

A total of 41 online PEM were included. The average Flesch-Reading ease score for all PEM was 63.43 (SD 14.21), indicating a standard difficulty level of reading. The average reading grade level was 7.85 (SD 2.38), which is higher than the recommended reading level for PEM. The mean DISCERN score was 45.66 (SD 3.34), and 27% (11/41) of the articles had DISCERN scores of less than 39, corresponding to poor or very poor quality.

CONCLUSIONS

The majority of online PEM on DFU are written above the recommended reading levels and have significant deficiencies in quality and reliability. Clinicians and patients should be aware of the shortcomings of these resources and consider the impact they may have on patients' self-management.

Predicting Health Material Accessibility: Development of Machine Learning Algorithms

Background

Current health information understandability research uses medical readability formulas to assess the cognitive difficulty of health education resources. This is based on an implicit assumption that medical domain knowledge represented by uncommon words or jargon form the sole barriers to health information access among the public. Our study challenged this by showing that, for readers from non-English speaking backgrounds with higher education attainment, semantic features of English health texts that underpin the knowledge structure of English health texts, rather than medical jargon, can explain the cognitive accessibility of health materials among readers with better understanding of English health terms yet limited exposure to English-based health education environments and traditions.

Objective

Our study explores multidimensional semantic features for developing machine learning algorithms to predict the perceived level of cognitive accessibility of English health materials on health risks and diseases for young adults enrolled in Australian tertiary institutes. We compared algorithms to evaluate the cognitive accessibility of health information for nonnative English speakers with advanced education levels yet limited exposure to English health education environments.

Methods

We used 113 semantic features to measure the content complexity and accessibility of original English resources. Using 1000 English health texts collected from Australian and international health organization websites rated by overseas tertiary students, we compared machine learning (decision tree, support vector machine [SVM], ensemble tree, and logistic regression) after hyperparameter optimization (grid search for the best hyperparameter combination of minimal classification errors). We applied 5-fold cross-validation on the whole data set for the model training and testing, and calculated the area under the operating characteristic curve (AUC), sensitivity, specificity, and accuracy as the measurement of the model performance.

Results

We developed and compared 4 machine learning algorithms using multidimensional semantic features as predictors. The results showed that ensemble classifier (LogitBoost) outperformed in terms of AUC (0.858), sensitivity (0.787), specificity (0.813), and accuracy (0.802). Support vector machine (AUC 0.848, sensitivity 0.783, specificity 0.791, and accuracy 0.786) and decision tree (AUC 0.754, sensitivity 0.7174, specificity 0.7424, and accuracy 0.732) followed. Ensemble classifier (LogitBoost), support vector machine, and decision tree achieved statistically significant improvement over logistic regression in AUC, sensitivity, specificity, and accuracy. Support vector machine reached statistically significant improvement over decision tree in AUC and accuracy. As the best performing algorithm, ensemble classifier (LogitBoost) reached statistically significant improvement over decision tree in AUC, sensitivity, specificity, and accuracy.

Conclusions

Our study shows that cognitive accessibility of English health texts is not limited to word length and sentence length as had been conventionally measured by medical readability formulas. We compared machine learning algorithms based on semantic features to explore the cognitive accessibility of health information for nonnative English speakers. The results showed the new models reached statistically increased AUC, sensitivity, and accuracy to predict health resource accessibility for the target readership. Our study illustrated that semantic features such as cognitive ability–related semantic features, communicative actions and processes, power relationships in health care settings, and lexical familiarity and diversity of health texts are large contributors to the comprehension of health information; for readers such as international students, semantic features of health texts outweigh syntax and domain knowledge.

Developing and Testing Automatic Models of Patient Communicative Health Literacy Using Linguistic Features: Findings from the ECLIPPSE study

Patients with diabetes and limited health literacy (HL) may have suboptimal communication exchange with their health care providers and be at elevated risk of adverse health outcomes. These difficulties are generally attributed to patients' reduced ability to both communicate and understand health-related ideas as well as physicians' lack of skill in identifying those with limited HL. Understanding and identifying patients with barriers posed by lower HL to improve healthcare delivery and outcomes is an important research avenue. However, doing so using traditional methods has proven difficult and infeasible to scale. This study using corpus analyses, expert human ratings of HL, and natural language processing (NLP) approaches to estimate HL at the individual patient level. The goal of the study is to better understand HL from a linguistic perspective and to open new research areas to enhance population management and individualized care. Specifically, this study examines HL as a function of patients' demonstrated ability to communicate health-related information to their providers via secure messages. The study develops an NLP-based HL model and validates the model by predicting patient-related events such as medical outcomes and hospitalizations. Results indicate that the developed model predicts human ratings of HL with ~80% accuracy. Validation indicates that lower HL patients are more likely to be nonwhite and have lower educational attainment. In addition, patients with lower HL suffered more negative health outcomes and had higher healthcare service utilization.

Employing computational linguistics techniques to identify limited patient health literacy: Findings from the ECLIPPSE study

OBJECTIVE

To develop novel, scalable, and valid literacy profiles for identifying limited health literacy patients by harnessing natural language processing.

DATA SOURCE

With respect to the linguistic content, we analyzed 283 216 secure messages sent by 6941 diabetes patients to physicians within an integrated system's electronic portal. Sociodemographic, clinical, and utilization data were obtained via questionnaire and electronic health records.

STUDY DESIGN

Retrospective study used natural language processing and machine learning to generate five unique "Literacy Profiles" by employing various sets of linguistic indices: Flesch-Kincaid (LP_FK); basic indices of writing complexity, including lexical diversity (LP_LD) and writing quality (LP_WQ); and advanced indices related to syntactic complexity, lexical sophistication, and diversity, modeled from self-reported (LP_SR), and expert-rated (LP_Exp) health literacy. We first determined the performance of each literacy profile relative to self-reported and expert-rated health literacy to discriminate between high and low health literacy and then assessed Literacy Profiles' relationships with known correlates of health literacy, such as patient sociodemographics and a range of health-related outcomes, including ratings of physician communication, medication adherence, diabetes control, comorbidities, and utilization.

PRINCIPAL FINDINGS

LP_SR and LP_Exp performed best in discriminating between high and low self-reported (C-statistics: 0.86 and 0.58, respectively) and expert-rated health literacy (C-statistics: 0.71 and 0.87, respectively) and were significantly associated with educational attainment, race/ethnicity, Consumer Assessment of Provider and Systems (CAHPS) scores, adherence, glycemia, comorbidities, and emergency department visits.

CONCLUSIONS

Since health literacy is a potentially remediable explanatory factor in health care disparities, the development of automated health literacy indicators represents a significant accomplishment with broad clinical and population health applications. Health systems could apply literacy profiles to efficiently determine whether quality of care and outcomes vary by patient health literacy; identify at-risk populations for targeting tailored health communications and self-management support interventions; and inform clinicians to promote improvements in individual-level care.

Predicting the readability of physicians' secure messages to improve health communication using novel linguistic features: Findings from the ECLIPPSE study

Background

Low literacy skills impact important aspects of communication, including health-related information exchanges. Unsuccessful communication on the part of physician or patient contributes to lower quality of care, is associated with poorer chronic disease control, jeopardizes patient safety and can lead to unfavorable healthcare utilization patterns. To date, very little research has focused on digital communication between physicians and patients, such as secure messages sent via electronic patient portals.

Method

The purpose of the current study is to develop an automated readability formula to better understand what elements of physicians' digital messages make them more or less difficult to understand. The formula is developed using advanced natural language processing (NLP) to predict human ratings of physician text difficulty.

Results

The results indicate that NLP indices that capture a diverse set of linguistic features predict the difficulty of physician messages better than classic readability tools such as Flesch Kincaid Grade Level. Our results also provide information about the textual features that best explain text readability.

Conclusion

Implications for how the readability formula could provide feedback to physicians to improve digital health communication by promoting linguistic concordance between physician and patient are discussed.

BENTO: A Visual Platform for Building Clinical NLP Pipelines Based on CodaLab

CodaLab is an open-source web-based platform for collaborative computational research. Although CodaLab has gained popularity in the research community, its interface has limited support for creating reusable tools that can be easily applied to new datasets and composed into pipelines. In clinical domain, natural language processing (NLP) on medical notes generally involves multiple steps, like tokenization, named entity recognition, etc. Since these steps require different tools which are usually scattered in different publications, it is not easy for researchers to use them to process their own datasets. In this paper, we present BENTO, a workflow management platform with a graphic user interface (GUI) that is built on top of CodaLab, to facilitate the process of building clinical NLP pipelines. BENTO comes with a number of clinical NLP tools that have been pre-trained using medical notes and expert annotations and can be readily used for various clinical NLP tasks. It also allows researchers and developers to create their custom tools (e.g., pre-trained NLP models) and use them in a controlled and reproducible way. In addition, the GUI interface enables researchers with limited computer background to compose tools into NLP pipelines and then apply the pipelines on their own datasets in a "what you see is what you get" (WYSIWYG) way. Although BENTO is designed for clinical NLP applications, the underlying architecture is flexible to be tailored to any other domains.

Clinical Text Data in Machine Learning: Systematic Review

Background

Clinical narratives represent the main form of communication within health care, providing a personalized account of patient history and assessments, and offering rich information for clinical decision making. Natural language processing (NLP) has repeatedly demonstrated its feasibility to unlock evidence buried in clinical narratives. Machine learning can facilitate rapid development of NLP tools by leveraging large amounts of text data.

Objective

The main aim of this study was to provide systematic evidence on the properties of text data used to train machine learning approaches to clinical NLP. We also investigated the types of NLP tasks that have been supported by machine learning and how they can be applied in clinical practice.

Methods

Our methodology was based on the guidelines for performing systematic reviews. In August 2018, we used PubMed, a multifaceted interface, to perform a literature search against MEDLINE. We identified 110 relevant studies and extracted information about text data used to support machine learning, NLP tasks supported, and their clinical applications. The data properties considered included their size, provenance, collection methods, annotation, and any relevant statistics.

Results

The majority of datasets used to train machine learning models included only hundreds or thousands of documents. Only 10 studies used tens of thousands of documents, with a handful of studies utilizing more. Relatively small datasets were utilized for training even when much larger datasets were available. The main reason for such poor data utilization is the annotation bottleneck faced by supervised machine learning algorithms. Active learning was explored to iteratively sample a subset of data for manual annotation as a strategy for minimizing the annotation effort while maximizing the predictive performance of the model. Supervised learning was successfully used where clinical codes integrated with free-text notes into electronic health records were utilized as class labels. Similarly, distant supervision was used to utilize an existing knowledge base to automatically annotate raw text. Where manual annotation was unavoidable, crowdsourcing was explored, but it remains unsuitable because of the sensitive nature of data considered. Besides the small volume, training data were typically sourced from a small number of institutions, thus offering no hard evidence about the transferability of machine learning models. The majority of studies focused on text classification. Most commonly, the classification results were used to support phenotyping, prognosis, care improvement, resource management, and surveillance.

Conclusions

We identified the data annotation bottleneck as one of the key obstacles to machine learning approaches in clinical NLP. Active learning and distant supervision were explored as a way of saving the annotation efforts. Future research in this field would benefit from alternatives such as data augmentation and transfer learning, or unsupervised learning, which do not require data annotation.

Using natural language processing and machine learning to classify health literacy from secure messages: The ECLIPPSE study

Limited health literacy is a barrier to optimal healthcare delivery and outcomes. Current measures requiring patients to self-report limitations are time-consuming and may be considered intrusive by some. This makes widespread classification of patient health literacy challenging. The objective of this study was to develop and validate "literacy profiles" as automated indicators of patients' health literacy to facilitate a non-intrusive, economic and more comprehensive characterization of health literacy among a health care delivery system's membership. To this end, three literacy profiles were generated based on natural language processing (combining computational linguistics and machine learning) using a sample of 283,216 secure messages sent from 6,941 patients to their primary care physicians. All patients were participants in Kaiser Permanente Northern California's DISTANCE Study. Performance of the three literacy profiles were compared against a gold standard of patient self-reported health literacy. Associations were analyzed between each literacy profile and patient demographics, health outcomes and healthcare utilization. T-tests were used for numeric data such as A1C, Charlson comorbidity index and healthcare utilization rates, and chi-square tests for categorical data such as sex, race, poor adherence and severe hypoglycemia. Literacy profiles varied in their test characteristics, with C-statistics ranging from 0.61-0.74. Relations between literacy profiles and health outcomes revealed patterns consistent with previous health literacy research: patients identified via literacy profiles indicative of limited health literacy: (a) were older and more likely of minority status; (b) had poorer medication adherence and glycemic control; and (c) exhibited higher rates of hypoglycemia, comorbidities and healthcare utilization. This represents the first successful attempt to employ natural language processing to estimate health literacy. Literacy profiles can offer an automated and economical way to identify patients with limited health literacy and greater vulnerability to poor health outcomes.

Manual wheelchair tilt-rest skill: a cross-sectional survey of awareness and capacity among wheelchair users

Purpose: The tilt-rest skill consists of tipping the wheelchair back and allowing it to rest against a solid object with the wheel locks applied (e.g., for pressure redistribution, neck comfort or hands-free activities). The objective of this study was to determine the proportion of experienced manual wheelchair users who are aware of this skill and who can perform it. Materials and methods: We conducted a cross-sectional survey of 49 manual wheelchair users using a questionnaire developed for the purpose. The tilt-rest skill was attempted by those who reported that they were capable of performing it. Results: Participants' mean (SD) age was 55.1 (18.2) years, 38 (77.6%) were male, their median (IQR) duration of wheelchair use was 2 (7.2) years and their mean (SD) daily time spent in the wheelchair was 9.5 (4.6) hours. Twenty-seven (55.1%) participants were aware of the skill, 19 (38.8%) reported being able to perform the skill and 16 of 47 (34.0%) were able to demonstrate the skill. Multivariate modelling with the question "Can you complete the tilt-rest skill?" as the dependent measure revealed an inverse relationship with age - Odds Ratio (95% Confidence Interval) of 0.476 (0.293, 0.774) (p = .0028) for each 10 year increase in age. Conclusions: Only just over half of manual wheelchair users are aware of the tilt-rest skill and one-third of users can perform it. Older people are less likely to report being able to complete the skill. These findings have implications for wheelchair skills training during the wheelchair-provision process. Implications for Rehabilitation Only just over half of manual wheelchair users are aware of the tilt-rest skill and only about one-third of users can perform it. Older people are less likely to report being able to complete the skill. These findings have clinical implications for wheelchair skills training during the, specifically that clinicians responsible for manual wheelchair-provision process should ensure that appropriate wheelchair users have the opportunity to learn this skill.