Multivariate Logistic Regression Analysis of Risk Factors and Clinical Countermeasures of Neurogenic Bladder After Spinal Cord Injury

Aims/Background Neurogenic bladder (NB) is a common complication after spinal cord injury (SCI), significantly impacting patients’ social functioning and quality of life. Urinary tract infections and neurological prognosis following SCI have been thoroughly investigated; however, studies assessing NB risk factors are limited, hindering its effective prevention. This study developed and validated a high-accuracy predictive model based on multivariate logistic regression analysis to identify key risk factors for NB after SCI and provide a basis for early intervention. Methods This study included 115 SCI patients hospitalised between January 2022 and July 2024. Patients were divided into the NB and non-NB groups. General data were collected and analysed using univariate and multivariate logistic regression methods. A prediction model for NB post-SCI was developed and validated using the Hosmer-Lemeshow test and receiver operating characteristic (ROC) curve analysis with the bootstrap resampling method. Furthermore, NB patients underwent routine rehabilitation, and improvements in urinary incontinence, single voiding volume, and micturition frequency were recorded before and after treatment. Results In this study, the incidence of NB in patients with SCI was 30.43%. Significant differences in NB incidence were observed based on age, American Spinal Cord Injury Association (ASIA) grade, urine culture results, C-reactive protein (CRP) levels, urination mode, pain scores, and sexual dysfunction (p < 0.05). Multivariate logistic regression identified old age, ASIA grade A, increased CRP levels, urinary incontinence/indwelling catheter, and higher pain scores as significant risk factors (p < 0.05). The formula is expressed as: NB prediction index after SCI = 0.058 × age + 0.045 × ASIA grade + 0.394 × CRP + 2.132 × urinary incontinence/indwelling catheter + 0.608 × pain score + 5.032. The prediction model had an overall accuracy rate of 97.14%, with a Hosmer-Lemeshow chi-square value of χ 2 =10.904 (p = 0.207), indicating a good fit. ROC curve analysis showed an area under the curve (AUC) of 0.904 (95% CI: 0.832–0.976, p < 0.05), with sensitivity of 88.60% and specificity of 82.50%. Post-treatment, daily urinary incontinence, and urination frequency decreased, while single voiding volume increased (p < 0.05). Conclusion The occurrence of NB post-SCI is affected by various factors, such as age, ASIA grade, CRP levels, urination modes, and pain severity. The risk prediction model showed strong predictive value, offering a promising early risk management and clinical decision-making method.