Replication of validation study of the Chew HL screening tool in an adult Italian population

Background Most of the research in the field of health literacy (HL) have been conducted in English-speaking countries. In particular, validation of measurement tools may be a particularly complex task as it is not always possible to completely reproduce the original validation methodology. Methods The original set of 16 Chew items have been first translated into Italian and pre-tested on 32 subjects for comprehensibility and adjusted accordingly. To validate Chew Items, we took the Italian version of the Newest Vital Sign (NVS) as the objective measure of Health Literacy. Additional demographic information was also collected (sex, age, level of education attained). As in the original Chew study, Receiver Operating Characteristic (ROC) curves were calculated for each item, single or in combination, for performance evaluation under two conditions: assessing the high probability of insufficient HL (NVS score 0-1) OR probability AND high probability of insufficient HL (NVS score 0-3). Area Under the Curve (AUC) of 0,8 was considered significant Results Of the 221 subjects asked, a total of 141 agreed to participate in the study (response rate=65%, age range 23-84, mean 56,8, 61% women). In contrast with the original Chew study, no item, single or in combination, showed a sufficient performance (AUC>0,8) for both conditions under study in our population. Conclusions As It was not possible to replicate findings of the original Chew study, results from different studies using this test should be interpreted cautiously if used out of the original context that provided the original validation.