Visual analogue scales for grading faecal incontinence and quality of life: their relationship with the Jorge-Wexner score and Rockwood scale.

BACKGROUND: The purpose of this study was to determine the probability of substituting two validated methods frequently used for scoring severity and quality of life of faecal incontinence (FI) by a single score represented in a visual analogue scale (VAS).
METHODS: One hundred and three consecutive unselected patients affected by FI of different aetiologies were prospectively included in the study. Evaluation of the continence status and quality of life referred to the last month was scored by (1) Jorge-Wexner score, (2) Rockwood FIQL scale, (3) Visual Analogue Scale for Faecal Incontinence (VASFI), and (4) Visual Analogue Scale for Quality of Life (VASQL). Patients unable to respond accurately to one or more of the questionnaires were excluded. The Bland and Altman method was applied to evaluate the agreement between Jorge-Wexner and VASFI. Correlation between both the analogue scales (VASFI and VASQL) and each one with each of the four FIQL scale scores was first determined by an analysis of simple correlation with each subscale and thereafter by multiple regression analysis following the backward strategy.
RESULTS: Comparison of VASFI with the Jorge-Wexner score shows that the two methods are neither concordant nor interchangeable, and in a given patient, the differences between both may fluctuate between 5.4 and -10.6 (95 % confidence interval). Correlation of VASFI and VASQL shows that the correlation between both the analogue scales is middling (r (2) = 0.543, β coefficient -0.538). Correlation of VASFI and FIQL subscales shows that in simple regression analysis, VASFI correlates with lifestyle, coping/behaviour, and embarrassment, but after multiple regression analysis, VASFI correlates significantly only with embarrassment. Correlation of VASQL and FIQL subscales shows that in simple regression analysis, VASQL correlates with lifestyle, coping/behaviour, and embarrassment, but after multiple regression analysis, VASQL only correlates significantly with coping/behaviour.
CONCLUSIONS: This study shows that a VAS for FI cannot replace the Jorge-Wexner score and a VAS for quality of life cannot substitute all the four subscales of FIQL. Severity of FI and its impact on quality of life expressed in a VAS only have a fair correlation, showing that they do not assess the same issues, which is also supported by the finding that VASFI correlates significantly only with the embarrassment subscale of FIQL.