A multidimensional analysis-based risk prediction model for stress urinary incontinence in middle-aged and elderly women.

backgroundUrinary incontinence (UI) is a common condition in middle-aged and elderly women, characterized by involuntary urine leakage. It significantly affects quality of life, causing embarrassment, depression, social isolation, and economic burdens. Although the prevalence of UI among adult Chinese women has declined to approximately 16.0% from 30.9% fifteen years ago, the rate of medical consultation remains low, around 10%.

objectiveThis study aimed to develop a multidimensional predictive model for stress urinary incontinence (SUI) to assist in the elderly women identification of high-risk individuals.

methodsWe conducted a retrospective review of clinical data from female patients undergoing urodynamic evaluations at the [BLINDED FOR REVIEW] from September 2020 to March 2025. After excluding incomplete records, 202 participants were included and randomly allocated into a training set (n = 151) and a validation set (n = 51) at a ratio of 7:3. Clinical variables including demographic characteristics, lifestyle factors, obstetric history, and relevant comorbidities (hypertension, diabetes, urethral prolapse, etc.) were analyzed. Predictive variables were selected using Least Absolute Shrinkage and Selection Operator (LASSO) regression in the training set. Subsequently, multivariable logistic regression analysis was conducted to construct the prediction model and nomogram. The nomogram provides an intuitive tool for individualized risk assessment, enabling early identification and targeted interventions. The discrimination ability was evaluated by the area under the receiver operating characteristic (ROC) curve (AUC), calibration was assessed via the Hosmer-Lemeshow test and calibration curves, and clinical utility was measured through decision curve analysis (DCA).

resultsBaseline characteristics revealed significant differences between SUI and non-SUI groups regarding BMI (P = 0.003), smoking (P = 0.001), alcohol consumption (P = 0.001), urinary tract symptoms (P < 0.001), and diabetes mellitus (P = 0.007). LASSO regression identified six critical predictors from 20 candidate variables: BMI ≥ 28 kg/m² (OR = 15.18, P = 0.007), urethral prolapse (OR = 12.57, P = 0.009), parity ≥ 3 (OR = 21.92, P = 0.033), diabetes mellitus (OR = 16.37, P = 0.011), history of urinary tract diseases (OR = 9.01, P = 0.004), and heavy physical labor (OR = 6.90, P = 0.05). These variables were confirmed as independent predictors by multivariate logistic regression (all P ≤ 0.05). The area under the ROC curve (AUC) was 0.94 (95% CI: 0.89-0.99) in the training set and 0.77 (95% CI: 0.63-0.92) in the validation set, indicating robust discrimination ability. Calibration curves showed high consistency between predicted and observed outcomes (Hosmer-Lemeshow test, P = 0.618). DCA demonstrated substantial net clinical benefits within the threshold probability range of 10-60%. The nomogram provided intuitive individualized risk assessment.

conclusionThis study successfully developed a multidimensional predictive model for SUI with favorable discrimination, calibration, and clinical applicability. The model serves as a valuable tool for early identification and targeted interventions in women at high risk of SUI. Further prospective studies and external validations are warranted to assess its applicability across diverse populations. A nomogram-based multidimensional model accurately predicts the risk of stress urinary incontinence, aiding early identification and targeted intervention in middle-aged and elderly women.