DEVELOPMENT AND VALIDATION OF THE MINNESOTA FECAL INCONTINENCE QUALITY OF LIFE (FIQL) SCALE

Background and Aims. Fecal incontinence (FI) symptoms can markedly impair quality of life (QoL). The Fecal Incontinence Quality of Life (FIQL) scale, which has 29 items divided into 4 subscales (lifestyle, coping/behavior, depression/self-perception, and embarrassment), is widely used in research studies but not in clinical practice. In addition, it was developed and validated in patients seeking care for FI but not in unselected persons with FI, raising concerns about its applicability to this larger population.

Aims. To create an abridged version of the FIQL and to evaluate the criterion validity of the shortened instrument.

Methods. The severity of FI and QoL were respectively evaluated with the Fecal Incontinence Severity Index [FISI]) and 4 instruments (SF36, FIQL, and 2 versions of abridged FIQL) in 903 persons with FI in the community (n=170), in a pelvic floor clinic practice (n=437), and in clinical trials (n=111). Stepwise forward linear regression was used to identify the optimal number and combination of questions from the FIQL questionnaire used to develop the full FIQL score. Linear regression was used to assess the relationship between FIQL and abridged FIQL scales with the SF36.

Results. FI was more severe in clinic and trial patients than community persons (P<.001) (Table). For example, the FISI score was 17.9 in community women but ranged from 30.1 to 35.9 in patients seeking FI treatment. We compared 2 abridged scales with 4 and 11 items with the original FIQL scale. The 4-item scale had 1 question from each FIQL subscale while the 11-item version contained multiple items from each subscale. In the entire cohort, the 11-item version predicted the total instrument (R² = .97, P = .01) better than the 4-item version (R²⁼88, P = .01). Among individual groups (Figure), the spread between the predicted and actual values was greater for the 4 than the 11-item version in all populations. Controlling for study population (ie, clinic or community), age, sex, and the FISI, the standardized beta for the 4 item is .89 (P = .01) and the 11-item is .97 (P = .01), which is consistent with the unadjusted evaluation. In 2 models with SF36 as the dependent variable (1) the 11-item FIQL score explained 41% of the variance (P=.005). Adding the total FIQL to this model did not increase the variance explained. (2) The total FIQL + 4 item FIQL scores also explained 40% of the variance (P<0.0001). But, in this model, only the total FIQL, not the 4-item FIQoL score was significant. This suggests that the 4-item scale does not capture the information relative to SF36.

Conclusions. There are differences in FI between the community and clinic. Our findings confirm the content, construct, and criterion validity of a user-friendly, abridged, 11-item Minnesota FIQL instrument that provides a single score rather than subscale scores as generated by the FIQL.