Online prediction tool for female pelvic floor dysfunction: development and validation

Area of Science:

• Urogynecology
• Digital Health
• Predictive Modeling

Background:

• Pelvic floor dysfunction (PFD) encompasses conditions like pelvic organ prolapse, stress urinary incontinence, and overactive bladder.
• Accurate and accessible diagnostic tools are crucial for timely intervention and management of PFD.

Purpose of the Study:

• To develop and validate a prediction model for PFD and its common subtypes.
• To create a patient-completed online tool for PFD assessment.

Main Methods:

• A retrospective cohort study of over 3,500 women was used for model development and internal validation.
• Penalized logistic regression incorporated 42 questionnaire items and demographic data.
• External validation was performed on 449 women to confirm model generalizability.

Main Results:

• The model demonstrated high accuracy in predicting PFD (97% sensitivity, 74% specificity, C-index 0.96).
• Accurate predictions were also achieved for pelvic organ prolapse, stress urinary incontinence, and overactive bladder.
• External validation confirmed the model's robust performance across different cohorts.

Conclusions:

• The developed online tool accurately differentiates women with and without PFD.
• This tool serves as a valuable self-assessment resource for women and primary care providers.
• The model facilitates the diagnosis of common pelvic floor disorders.