Measurement Comparability of Electronic and Paper Administration of Visual Analogue Scales: A Review of Published Studies

Abstract

Background: Visual analogue scales (VASs) are used in a variety of patient-, observer- and clinician-reported outcome measures. While typically included in measures originally developed for pen-and-paper completion, a greater number of clinical trials currently use electronic approaches to their collection. This leads researchers to question whether the measurement properties of the scale have been conserved during the migration to an electronic format, particularly because electronic formats often use a different scale length than the 100 mm paper standard.

Methods: We performed a review of published studies investigating the measurement comparability of paper and electronic formats of the VAS.

Results: Our literature search yielded 26 studies published between 1997 and 2018 that reported comparison of paper and electronic formats using the VAS. After excluding 2 publications, 23 of the remaining 24 studies included in this review reported electronic formats of the VAS (eVAS) and paper formats (pVAS) to be equivalent. A further study concluded that eVAS and pVAS were both acceptable but should not be interchanged. eVAS length varied from 21 to 200 mm, indicating that 100 mm length is not a requirement.

Conclusions: The literature supports the hypothesis that eVAS and pVAS provide comparable results regardless of the VAS length. When implementing a VAS on a screen-based electronic mode, we recommend following industry best practices for faithful migration to minimise the likelihood of non-comparability with pVAS.

Keywords: Electronic patient-reported outcome; Measurement comparability; Measurement equivalence; VAS; Visual analogue scale; ePRO.

Fig. 1

Typical representations of pVAS and eVAS based on The Stanford Health Assessment Questionnaire Disability Index HAQ-DI [46]

Fig. 2

Summary of eVAS implementations. a Two studies included two different electronic formats. b Seven studies did not report (or enable estimation of) eVAS length

Fig. 3

Summary of correlation analyses comparing eVAS and pVAS scores (n = 41). r Pearson’s r, R² coefficient of determination, ρ spearman’s rho. R² values presented as R values in this figure. Box plot description: box = interquartile range; box divider = median; “+” = mean; start and end of whisker lines: minimum and maximum values, respectively

Fig. 4

Summary of ICC analyses comparing eVAS and pVAS scores (n = 37). Box plot description: box = interquartile range; box divide r = median; “+” = mean; start and end of whisker lines: minimum and maximum values, respectively