Evaluating the validity of health literacy items in English- and Spanish-speaking Hispanics from a community registry

Introduction: Individuals with inadequate literacy have poor health-related outcomes such as lack of medication adherence, elevated hospitalization rates and high healthcare costs. Well validated measures such as the Short Test of Functional Health Literacy in Adults (STOFHLA) are time-consuming and not practical to implement in clinical settings. However, assessing health literacy is critical for ensuring effective patient-physician discussions. Prior studies have found that single-item measures may be sufficient to identify low-literacy individuals, but little is known about the performance of these items among Spanish speakers. This study evaluated the validity of three such items to screen for inadequate health literacy in Spanish- and English-speaking Hispanics.

Methods: Hispanic members of a community research registry ages 18-64 year were invited to complete a survey in English or Spanish. We measured health literacy with the STOFHLA and three single-item questions: self-rated reading ability, confidence completing medical forms, and help needed with reading hospital materials. To evaluate validity for detecting inadequate literacy, we compared each item to the STOFHLA (gold standard) using area under the receiver operator curves (AUROC). Chew et al. defined performance as good if AUROC >0.80.

Results: In 129 Hispanic adults (73.6% Spanish speakers), we found more inadequate health literacy, as measured by STOFHLA, among Spanish speakers compared to English speakers (16.3% and 6.1%). For the single-item measures, rates of reporting inadequate literacy were lower among Spanish than English speakers (e.g., very poor/poor/ok reading ability was 18.9% and 26.4%; some/a little bit of confidence completing medical forms was 9.5% and 26.5%). When compared to the STOFHLA, only the reading ability item was effective at identifying inadequate literacy in Spanish and English (AUROC of 0.819 and 0.822, respectively). Confidence with medical forms was not effective at identifying inadequate literacy among Spanish speakers, but was effective among English speakers (AUROC of 0.615 and 0.887, respectively). The help reading item was not effective for either Spanish or English group (AUROC of 0.395 and 0.363, respectively).

Conclusion: Self-rated reading ability is the best item to screen for low health literacy among Hispanics regardless of language preference. Rapid, practical methods to identify low-literacy patients are needed in busy clinical settings. Adding this screening item during medical history intake and as a discrete field in electronic medical records would enable healthcare organizations to identify and deliver interventions to low-literacy individuals at risk for poor health outcomes.