Medical Word Recognition Overestimates Comprehension Among Hospitalized Adults

Health literacy screening in clinical settings often relies on brief word-recognition tools, such as the Rapid Estimate of Adult Literacy in Medicine–Short Form (REALM-SF), based on the assumption that recognizing medical terms reflects patient understanding. However, health literacy is multidimensional, and limited evidence exists on whether word recognition accurately reflects comprehension of medical terminology among hospitalized adults. This study examined the concordance between medical word recognition and comprehension and assessed potential misclassification when relying on word-recognition screening alone. In this prospective cross-sectional study, English-speaking adults hospitalized on a general medicine service at a teaching hospital were screened using the REALM-SF. Participants were asked to read each of the seven REALM-SF terms aloud and to define each term. Word recognition scores (0–7) and dictionary-aligned term-definition scores (0–7) were independently scored. We characterized the distributions of recognition and comprehension scores, evaluated concordance between the two measures, examined the ability of word recognition to identify comprehension deficits under prespecified definitions, and modeled associations between recognition, comprehension, and demographic factors. Among 466 participants (mean age 66.8 years), word recognition was high (mean score 6.6), with 80.5% achieving perfect recognition. In contrast, comprehension was substantially lower (mean score 3.9), and only 4.5% correctly defined all seven terms. Comprehension varied widely by term (23.2%–76.0%). Word recognition and comprehension were moderately correlated (r = 0.39, p < 0.001), yet word recognition demonstrated low sensitivity for identifying comprehension deficits. In adjusted analyses, higher word recognition was associated with greater comprehension, though overall model fit improvement was modest. Among hospitalized adults, medical word recognition substantially overestimates comprehensionof common medical terms. Reliance on brief word-recognition screening tools alone may misclassify patients’ understanding and provide false reassurance, underscoring the need for direct checks of comprehension, such as teach-back, in inpatient care.