Rasch analysis of the Indian vision function questionnaire.

AIM

To investigate the psychometric properties of the three scales (general functioning, psychosocial impact, visual symptoms) of the Indian vision function questionnaire (IND-VFQ) using the Rasch measurement model.

METHODS

236 visually impaired patients referred to vision rehabilitation centres were administered the 33-item IND-VFQ. Rasch analysis was used to investigate the scales for the following properties: precision by person separation (ie, discrimination between strata of patient ability, recommended minimum value 2.0), unidimensionality (ie, measurement of a single construct) and targeting (ie, matching of item difficulty to patient ability).

RESULTS

Only the general functioning scale possessed adequate measurement precision (person separation 3.49). However, it lacked unidimensionality as some items did not contribute towards the measurement of a single construct indicating a secondary dimension. This comprised seven mobility items, which formed a separate valid subscale with good targeting (-0.57 logits). Deleting these items restored unidimensionality but a misfitting item required removal. Following this the 13 items fit and were visual functioning related. However, targeting was suboptimal (-1.13 logits).

CONCLUSIONS

The general functioning scale of the IND-VFQ consists of two separate unidimensional constructs: visual functioning and mobility. Both these Rasch scaled versions with good psychometric properties are effective tools for the assessment of visually impaired patients in India.