Context Patients with limited literacy are at higher risk for poor health outcomes; however, physicians' perceptions are inaccurate for identifying these patients. Objective To systematically review the accuracy of brief instruments for identifying patients with limited literacy. Data Sources Search of the English-language literature from 1969 through February 2010 using PubMed, Psychinfo, and bibliographies of selected manuscripts for articles on health literacy, numeracy, reading ability, and reading skill. Study Selection Prospective studies including adult patients 18 years or older that evaluated a brief instrument for identifying limited literacy in a health care setting compared with an accepted literacy reference standard. Data Extraction Studies were evaluated independently by 2 reviewers who each abstracted information and assigned an overall quality rating. Disagreements were adjudicated by a third reviewer. Data Synthesis Ten studies using 6 different instruments met inclusion criteria. Among multi-item measures, the Newest Vital Sign (English) performed moderately well for identifying limited literacy based on 3 studies. Among the single-item questions, asking about a patient's use of a surrogate reader, confidence filling out medical forms, and self-rated reading ability performed moderately well in identifying patients with inadequate or marginal literacy. Asking a patient, “How confident are you in filling out medical forms by yourself?” is associated with a summary likelihood ratio (LR) for limited literacy of 5.0 (95% confidence interval [CI], 3.8-6.4) for an answer of “a little confident” or “not at all confident”; a summary LR of 2.2 (95% CI, 1.5-3.3) for “somewhat confident”; and a summary LR of 0.44 (95% CI, 0.24-0.82) for “quite a bit” or “extremely confident.” Conclusion Several single-item questions, including use of a surrogate reader and confidence with medical forms, were moderately effective for quickly identifying patients with limited literacy.
