Comparing binary & ordinal definitions of urinary & stool continence outcomes: Data from the National Spina Bifida...

The National Spina Bifida Patient Registry (NSBPR) assesses bladder and bowel incontinence using ordinal categories, but prior NSBPR analyses employed binary classification. Our aims were to 1) perform the first NSBPR analysis of bladder and bowel incontinence as ordinal outcomes to compare to the binary definition and subject variables; 2) explore the correlation of incontinence with undergarment usage, and 3) assess incontinence status following continence surgeries.

Data from NSBPR participants' most recent clinic visit from 2013 to 2020 were analyzed. Ordinal categories of incontinence were compared to previously used binary definitions. Incontinence surgical outcomes were analyzed for those with data at least three months post-operatively. Chi-square tests evaluated associations among categorical variables. Univariate and ordinal logistic regression models were used to test associations of ordinal incontinence status with patient and condition factors. Statistical tests were 2-sided; p values < 0.05 were considered significant.

Analysis of 7217 individuals using ordinal incontinence outcomes showed little difference from previously used binary outcomes. The final multivariable logistic regression models with ordinal multinomial outcomes showed that associations of incontinence with age, sex, race/ethnicity, health insurance, level of lesion, and continence management technique were similar to prior studies. Among those reporting never being incontinent of both bladder and bowel, 14% reported using protective undergarments. Of the 500 individuals who had bladder outlet surgery, 38% reported never being incontinent of urine. Of 1416 individuals who had appendicostomy (ACE) bowel surgery, 48% reported never being incontinent of stool.

Our current analysis showed that ordinal continence outcome classification had similar continence findings as previous studies using the binary definition of continence. Expanding the binary definition of continence to include monthly episodes of incontinence did not greatly increase the proportion of continent individuals and, therefore, would have not likely made meaningful differences in continence outcomes in prior NSBPR analyses. However, it is known that even mild incontinence can affect quality of life, therefore, capturing any level of incontiennce is of clinical importance. Confirmation of the association of continence outcomes with sociodemographic, condition-related, and interventional factors with both approaches further validates previous analyses using the binary definition of continence.

The previously used binary definition of bladder and bowel continence appears robust. Undergarment choice was a poor surrogate for reported incontinence. After bladder and bowel continence surgeries, 38% and 48%, respectively, reported never being incontinent.