Interactive, computer-based, self-reported, visual function questionnaire: the PalmPilot-VFQ

Abstract

Purpose: To evaluate the testing performance, reliability, and validity of a self-administered visual function questionnaire designed for a Palm Pilot in comparison with the interviewer-administered-National Eye Institute Visual Function Questionnaire-25 (NEI-VFQ-25) (interviewer-NEI-VFQ-25) and self-administered-NEI-VFQ-25 (self-NEI-VFQ-25).

Method: The interviewer-NEI-VFQ-25 was administered to 135 sequential patients who visited a retina clinic, followed on separate days by the Palm Pilot-Visual Function Questionnaire (PalmPilot-VFQ) and self-NEI-VFQ-25. Rasch analysis of ordinal difficulty ratings for the PalmPilot-VFQ was used to estimate interval measures of perceived visual ability. Reliability was determined by calculating Cronbach's alpha and test-retest intraclass correlation coefficients (ICCs). Concurrent validity was determined by calculating correlations of the PalmPilot-VFQ score with that of a general vision question. For evaluating convergent validity, the PalmPilot-VFQ was compared with the interviewer-NEI-VFQ-25, habitual-correction visual acuity (HCVA), and with two visual analogue scale (VAS) questions. Performance time and testability were compared among the three questionnaires.

Results: Rasch analysis eliminated two items in the PalmPilot-VFQ due to poor-fit statistics. The final items showed internal consistency (Cronbach's alpha=0.89) and test-retest reliability (ICC=0.79), as well as an excellent separation index (3.23 and 4.01) for item parameters with significant concurrent correlation (P<0.0001). On evaluating convergent validity, the PalmPilot-VFQ showed strong correlations with interviewer-NEI-VFQ-25, with the HCVA of the better-seeing eye, and with the VAS questions (P=0.0001). Ninety percent of the 135 patients (HCVA >20/200 in the better-seeing eye) could perform the PalmPilot-VFQ with their habitual correction or high-plus spectacles, but in significantly less time than either interviewer-NEI-VFQ-25 or self-NEI-VFQ-25 (P<0.0001).

Conclusion: The PalmPilot-VFQ seems to be a reliable, valid, interactive, computer-based, self-administered questionnaire that can be used routinely by physicians to evaluate functional vision disability in populations with a high prevalence of macular disease.