2019. Illustrating the NPDFC Ambulatory Care Survey for Low Literate Patients          
JA Shea, Philadelphia VA Center for Health Equity Research and Promotion, KL Ravenell, Philadelphia VA Center for Health Equity Research and Promotion, CAN Henry, University of Pennsylvania, KM Fosnocht, University of Pennsylvania Health System, JJ Murphy, Philadelphia VA, VJ McDonald, Philadelphia VA Center for Health Equity Research and Promotion, AC Aguirre, University of Pennsylvania, DA Asch, Philadelphia VA Center for Health Equity Research and Promotion

Objectives: An illustrated patient satisfaction instrument (IPSI) was created to address needs of low literate populations.  The IPSI was compared to the standard 62-item NPDFC printed Ambulatory Care Satisfaction Survey in terms of 1) completion time; 2) missing data; and 3) equivalence of subscale scores.

Methods: The IPSI or the NPDFC was randomly assigned to 1709 patients in primary care waiting areas of the Philadelphia VAMC and University of Pennsylvania Health System.  Patients also completed the Rapid Estimate of Adult Literacy in Medicine (REALM).  Low literacy was defined as 8th grade or less.  T-tests compared low literate subgroups and high literate subgroups on completion time, amount of missing and illogical data, and subscale scores.  Differences were expected between forms for the low literate comparisons favoring the IPSI but not between the high literate samples.

Results:  44% of those taking the IPSI and 37% of those taking the NPDFC had low literacy scores. For the low literate samples, completion times were 21.4 minutes for the IPSI and 18.4 minutes for the NPDFC (p = .01).  For the high literate sample, the times were 19.7 and 15.9, respectively (p = .007).  For both low literate groups, the median amount of missing data on the item level was 6%.  For both high literate groups, the median was 5%.  On many of the 11 subscales for both forms, the high literate groups had lower (less dissatisfied) mean scores than the low literate groups.  On one subscale (ACCESS), subscale scores were higher (more dissatisfied) for both low and high literate groups taking the IPSI (p = .04 and p = .03) compared to the NPDFC.

Conclusions: The IPSI performed similarly to the NPDFC.  Different subscale scores require more study and a detailed look at item differences will highlight where illustrations made a difference.

Impact: More than a third of veterans have low health literacy. If patient satisfaction information is to represent the broad veteran population accurately, it must be collected in a format that is accessible to the substantial proportion of veterans with lower literacy skills.  By relying on illustrations, the IPSI may offer advantages to the text instrument.