Little is known about provider accuracy in assessing health literacy and numeracy of individuals with diabetes and the effect of incorrect assessment on patient outcomes. We conducted a clinic-based study (121 patients, 12 providers) to explore whether there was a discrepancy between patient-actual and provider-perceived literacy and numeracy levels among those (≥18yrs) with type 1 (n=59), type 2 (n=57), and unknown (n=5) diabetes, and to test whether this discrepancy was associated with worse HbA1c. We used multiple linear regression to compare mean HbA1c of patients whose providers overestimated, underestimated, and correctly estimated literacy and numeracy. Literacy was assessed by The Rapid Estimate of Adult Literacy in Medicine-Short Form and numeracy by the Diabetes Numeracy Test. Based on a directed acyclic graph and comparison of precision and bias of multiple models, the final models were adjusted for insurance (public/private), race (white/other), and education (≤high school/≥some undergraduate/≥some postgraduate).
Providers incorrectly identified 33.3% of patients’ literacy and 55.5% of numeracy. Providers overestimated 15.7% of patients’ literacy and 47.1% of numeracy. Providers underestimated 2.5% of literacy and 15.7% of numeracy. Mean Hb1Ac was 8.1%. Patients whose provider incorrectly perceived their literacy and numeracy score had, respectively, an HbA1c that was on average 0.9% (p=0.01) and 0.6% (p=0.05) higher than those whose scores were correctly perceived. Further, patients with overestimated numeracy had a 0.9% higher HbA1c than those whose high numeracy was correctly identified (p=0.001). These findings suggest providers might benefit from accurate knowledge of patients’ numeracy and literacy so that appropriate information can be communicated. Future research is needed to test effects of changing provider communication to minimize overestimation of patient numeracy and literacy levels.
R. Logendran: None. A. Cristello: None. J. DuBois: None. B.S. Friedman: None. E.J. Mayer-Davis: None.