BACKGROUND: The most commonly used tool for assessing member satisfaction with health plans in the United States is the Consumer Assessment of Health Plans (CAHPS). Yet its complexity exceeds the abilities of many patient populations. An illustrated version of the CAHPS instrument was designed for low-literacy audiences; the illustrations were tested and were then revised toreflect respondents' feedback. METHODS: Ilustrattions were developed to support the central themes in each of 63 CAHPS text items. The cognitive testing approach was used to create a systematic method to analyze key features of an item and develop the corresponding interview procedure. RESULTS: Following cognitive interviews on subsets of illustrations and items with more than 900 patients, all but 7 (11%) of the 63 items met the criterion that no more than 25% of the sample who saw the illustration could be rated as having "limited/no understanding." By the final pilot, a median of 66% had "full understanding," 20% had "partial understanding," and 14% had "limited/no understanding" of the 43 illustrations needing revision. CONCLUSIONS: The protocol helped improve the respondents' understanding of the illustrations. The experiences and lessons learned can guide design of illustration-enhanced materials for low-literacy populations.
BACKGROUND: The most commonly used tool for assessing member satisfaction with health plans in the United States is the Consumer Assessment of Health Plans (CAHPS). Yet its complexity exceeds the abilities of many patient populations. An illustrated version of the CAHPS instrument was designed for low-literacy audiences; the illustrations were tested and were then revised toreflect respondents' feedback. METHODS: Ilustrattions were developed to support the central themes in each of 63 CAHPS text items. The cognitive testing approach was used to create a systematic method to analyze key features of an item and develop the corresponding interview procedure. RESULTS: Following cognitive interviews on subsets of illustrations and items with more than 900 patients, all but 7 (11%) of the 63 items met the criterion that no more than 25% of the sample who saw the illustration could be rated as having "limited/no understanding." By the final pilot, a median of 66% had "full understanding," 20% had "partial understanding," and 14% had "limited/no understanding" of the 43 illustrations needing revision. CONCLUSIONS: The protocol helped improve the respondents' understanding of the illustrations. The experiences and lessons learned can guide design of illustration-enhanced materials for low-literacy populations.