Find in Library

BackgroundOur research team developed two prediction models of postpartum stress urinary incontinence (PSUI) , one for primiparas, and the other for multiparae, aiming at early identifying women at high risk of PSUI, and providing effective interventions, but they have not yet been externally validated.ObjectiveTo externally validate the risk prediction models of PSUI previously developed by us to assess their clinical applicability.MethodsThis study was conducted between July and September 2020. Participants were 6-month postpartum women (validation group) who were selected from the electronic medical record system of Shenzhen Hospital of Southern Medical University, and the University of Hong Kong-Shenzhen Hospital. Information about age, height, pre-pregnancy weight, abortion history, and delivery history was collected from the electronic medical record system of the two hospitals. A telephone follow-up was conducted to investigate the incidence of stress urinary incontinence within 6-month postpartum. The area under the receiver operating characteristic curve (AUC) was computed to estimate the value of the predictive models in discriminating PSUI. Hosmer-Lemeshow goodness-of-fit test was used to examine the calibration of the prediction models.ResultsA total of 298 cases were included, and 203 of them (68.1%) were primiparas (158 with PSUI, and other 45 without) , other 95 (31.9%) were multiparae (72 with PSUI, and other 23 without) . The AUC of the risk prediction model for PSUI in primiparas was 0.719〔95%CI (0.643, 0.795) 〕, and that of the risk prediction model for PSUI in multiparae was 0.833〔95%CI (0.738, 0.928) 〕. Hosmer-Lemeshow goodness-of-fit test suggested that the PSUI risk prediction model for primiparas had poor calibration (χ2=34.11, P<0.001) , while that for multiparae had satisfactory calibration (χ2=9.62, P=0.293) .ConclusionThe PSUI risk prediction model for primiparas could effectively distinguish PSUI, but its applicability needs to be further improved. The PSUI risk prediction model for multiparae had acceptable performance, which may be used and promoted as an evaluation tool for early pelvic floor rehabilitation in multiparae.