Limitations of the S-TOFHLA in measuring poor numeracy: A cross-sectional study

Background: Although the Short Test of Functional Health Literacy in Adults (S-TOFHLA) is widely used, misidentification of individuals with low health literacy (HL) in specific HL dimensions, like numeracy, is a concern. We examined the degree to which individuals scored as "adequate" HL on the S-TOFHLA would be considered as having low HL by two additional numerical measures. Methods: English-speaking adults aged 45-75 years were recruited from a large, urban academic medical center and a community foodbank in the United States. Participants completed the S-TOFHLA, the Subjective Numeracy Scale (SNS), and the Graphical Literacy Measure (GL), an objective measure of a person's ability to interpret numeric information presented graphically. Established cut-points or a median split classified participants and having high and low numeracy. Results: Participants (n = 187), on average were: aged 58 years; 63% female; 70% Black/African American; and 45% had a high school degree or less. Of those who scored "adequate" on the S-TOFHLA, 50% scored low on the SNS and 40% scored low on GL. Correlation between the S-TOFHLA and the SNS Total was moderate (r = 0.22, n = 186, p = 0.01), while correlation between the S-TOFHLA and the GL Total was large (r = 0.53, n = 187, p ≤ 0.01). Conclusions: Findings suggest that the S-TOFHLA may not capture an individuals' HL in the dimension of numeracy. Efforts are needed to develop more encompassing and practical strategies for identifying those with low HL for use in research and clinical practice. Trial registration: NCT02151032 (retrospectively registered: May 30, 2014).