Published in:
01-11-2016 | Original Research
Development and Validation of the Spanish Numeracy Understanding in Medicine Instrument
Authors:
Elizabeth A. Jacobs, MD AMPP, Cindy M. Walker, PhD, Tamara Miller, PhD, Kathlyn E. Fletcher, MD, MA, Pamela S. Ganschow, MD, Diana Imbert, BS, Maria O’Connell, BA, Joan M. Neuner, MD, MPH, Marilyn M. Schapira, MD, MPH
Published in:
Journal of General Internal Medicine
|
Issue 11/2016
Login to get access
ABSTRACT
BACKGROUND
The Spanish-speaking population in the U.S. is large and growing and is known to have lower health literacy than the English-speaking population. Less is known about the health numeracy of this population due to a lack of health numeracy measures in Spanish.
OBJECTIVE
we aimed to develop and validate a short and easy to use measure of health numeracy for Spanish-speaking adults: the Spanish Numeracy Understanding in Medicine Instrument (Spanish-NUMi).
DESIGN
Items were generated based on qualitative studies in English- and Spanish-speaking adults and translated into Spanish using a group translation and consensus process. Candidate items for the Spanish NUMi were selected from an eight-item validated English Short NUMi. Differential Item Functioning (DIF) was conducted to evaluate equivalence between English and Spanish items. Cronbach’s alpha was computed as a measure of reliability and a Pearson’s correlation was used to evaluate the association between test scores and the Spanish Test of Functional Health Literacy (S-TOFHLA) and education level.
PARTICIPANTS
Two-hundred and thirty-two Spanish-speaking Chicago residents were included in the study.
KEY RESULTS
The study population was diverse in age, gender, and level of education and 70 % reported Mexico as their country of origin. Two items of the English eight-item Short NUMi demonstrated DIF and were dropped. The resulting six-item test had a Cronbach’s alpha of 0.72, a range of difficulty using classical test statistics (percent correct: 0.48 to 0.86), and adequate discrimination (item-total score correlation: 0.34–0.49). Scores were positively correlated with print literacy as measured by the S- TOFHLA (r = 0.67; p < 0.001) and varied as predicted across grade level; mean scores for up to eighth grade, ninth through twelfth grade, and some college experience or more, respectively, were 2.48 (SD ± 1.64), 4.15 (SD ± 1.45), and 4.82 (SD ± 0.37).
CONCLUSIONS
The Spanish NUMi is a reliable and valid measure of important numerical concepts used in communicating health information.