- Healthcare professionals need a better metric for testing patient readability for EHR notes, especially as patients gain more access to their health data via clinician notes and the patient portal, says a recent study published in the Journal of Medical Internet Research.
According to a research team from the University of Massachusetts, healthcare professionals have praised patient access to EHR notes as a successful method for improving patient engagement and patient-centered healthcare.
To ensure EHR note readability, healthcare professionals have developed several readability formulas, including the Flesch-Kincaid Grade Level (FKGL), Simple Measure of Gobbledygook (SMOG), and Gunning-Fog Index (GFI).
However, some experts are skeptical of these formulas’ effectiveness in testing actual patient EHR note readability.
“In general, these aforementioned metrics rely on the assumption that the longer the words and the sentences, the more difficult the text is,” the researchers explained. “However, this assumption may not hold true for EHR narratives, which contain lists of clinical events (eg, medication list), abbreviations, and incomplete and short sentences, unduly lowering the readability score.”
The researchers collected 140 Wikipedia articles and 242 EHR notes on diabetes, and assigned 15 laypeople to read and assess readability for all articles. The research team also applied FKGL, SMOG, and GFI readability formulas on each article.
All three readability formulas predicted that EHR notes were easier to read than Wikipedia articles. However, patients reported the opposite, stating that Wikipedia articles were 21 percent easier to read than EHR notes. According to patient reports, word use, sentence structure, and frequency and difficulty of medical jargon made EHR notes more difficult to read.
“Writing style such as choice of words and textual cohesion might also account for some of the variances in the difficulty in perceptions of EHR notes,” the researchers said. “In spite of the differences, these correlations suggested that medical jargon was a substantial contributor to readers’ perceived difficulty of both genres of text.”
According to the researchers, there was limited correlation between patient-reported readability and formulas for both EHR notes and Wikipedia articles, meaning that patient assessments and formula assessments did not show the same level of text readability. This suggested that readability formulas are not an accurate method to predict readability, the research team said.
These results should encourage clinicians to write EHR notes in simple and easy-to-read language, the researchers noted. As the healthcare industry increasingly values patient portals and initiatives such as OpenNotes, it will be important that EHR notes be written in language that patients can understand and apply to their health.
Additionally, these results should encourage industry professionals to reassess how they measure EHR note readability, making these measures more accurate to ensure patients can glean actionable information from these materials.
“Higher readability could facilitate patient comprehension. Existing formulas are widely used in the health domain, but our analysis suggested that they were not accurate when applied on complex documents,” the researchers wrote. “A better metric should incorporate features beyond simple word and sentence length, such as the complexity of the concepts involved in the document.”
Ultimately, clinicians must focus on how patient reading materials will affect patient engagement efforts. Making engagement materials such as EHR notes and patient portals more readable, providers can ensure patients are empowered with better information to improve their own health.
“As more institutions allow patients access to their own EHR records, patients are also interested in reading them,” the research team concluded. “Research has shown that patients may need help in understanding them. An accurate readability metric for the EHR notes can encourage physicians to write notes in a simpler language. It may make patient portals more useful.”