A model for predicting motor urge urinary incontinence

Background: While the historical interview has been shown to diagnose stress urinary incontinence (UI) with reasonable accuracy, it is less accurate in the diagnosis of urge or mixed UI. Objectives: To construct an optimal model for the diagnosis of motor urge UI, and to refine this model into a simplified instrument that can be used to diagnose motor urge UI during a routine incontinence evaluation. Methods: A model was constructed to allow a more accurate diagnosis of motor urge UI using historical data. Initially, an optimal model was developed that used three key symptoms, age, gender, a history of neurologic disorder, obstruction diagnosed via voiding pressure study, and the urethral resistance algorithm to diagnose motor urge UI. A simplified model was then constructed using factors such as symptoms of motor urge UI, age, and gender that were readily accessible to the nurse when completing a routine UI evaluation. This simplified model was used to develop an instrument for the clinical diagnosis of motor urge UI. Results: While the agreement between clinical and urodynamic diagnosis was relatively high among patients with genuine stress UI (93% accuracy rate), it was considerably less among patients with urge and mixed UI, yielding accuracy rates of 63% and 35%, respectively. An optimal model for diagnosing motor urge UI was constructed and provided an overall accuracy rate of 91%. A simplified model was then constructed and evaluated for performance by least squares fit test. It revealed an R-2 Of 0.85 and an adjusted R-2 Of 0.84. Conclusions: A combination of age, gender, and three key symptoms (diurnal frequency, nocturia, and symptom of urge incontinence) provide an accurate and clinically useful model for the diagnosis of motor urge UI. Additional research is recommended to test the validity and reliability of the instrument derived from this model.