Exploring predictors of postpartum pelvic floor hypertonicity in women undergoing spontaneous vaginal delivery: an interpretable machine learning perspective

Introduction Postpartum pelvic floor hypertonicity is an underrecognized form of postpartum pelvic floor dysfunction characterized by increased resting pelvic floor muscle activity and myofascial pain. However, predictors of postpartum pelvic floor hypertonicity following spontaneous vaginal delivery remain insufficiently characterized. This study aimed to develop an interpretable machine learning-based prediction model to explore predictors associated with postpartum pelvic floor hypertonicity. Methods This retrospective prediction modeling study included postpartum women who underwent pelvic floor assessment after spontaneous vaginal delivery at Shenzhen Hospital, Southern Medical University, between January 2024 and February 2025. Candidate predictors were obtained from information available by hospital discharge, whereas postpartum pelvic floor hypertonicity was determined at the subsequent first postpartum pelvic floor assessment using surface electromyography combined with digital vaginal palpation. Participants were randomly divided into train and held-out test sets (7:3). Nine machine learning algorithms were evaluated using nested cross-validation. The final model was interpreted using SHapley Additive exPlanations. Results Among 1,297 women, 1,031 were included in the final analysis and randomly divided into train ( n = 722) and test ( n = 309) sets. A total of 505 women were classified as having postpartum pelvic floor hypertonicity (48.98%). Random forest achieved the numerically highest mean outer-fold area under the receiver operating characteristic curve (0.780 ± 0.018) and was selected as the final model. In the held-out test set, the model achieved an area under the receiver operating characteristic curve of 0.789 (95% confidence interval: 0.739–0.838). SHapley Additive exPlanations analysis identified neuraxial analgesia during labor, gestational diabetes mellitus, gestational hypertension, residence, and first stage of labor as the most influential contributors to model predictions. The absence of neuraxial analgesia during labor, the presence of gestational diabetes mellitus and gestational hypertension, and rural residence contributed to higher model-predicted probabilities, whereas first stage of labor showed a nonlinear contribution pattern. Conclusions Interpretable machine learning provided a useful approach for exploring predictors associated with postpartum pelvic floor hypertonicity following spontaneous vaginal delivery. Clinical Trial Registration Chinese Clinical Trial Registry, identifier ChiCTR2600117411.

Authors

Institutions

Publication Details

Journal
Frontiers in Medicine
Published
2026-11-18
DOI
https://doi.org/10.3389/fmed.2026.1942416
Primary Topic
Pelvic floor disorders treatments
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Exploring predictors of postpartum pelvic floor hypertonicity in women undergoing spontaneous vaginal delivery: an interpretable machine learning perspective

Wenzhi Cai, Ao Huang, Zhikai Ding, Qianwei Chen et al.
5 links
Frontiers in Medicine
Pelvic floor disorders treatments
article

Exploring predictors of postpartum pelvic floor hypertonicity in women undergoing spontaneous vaginal delivery: an interpretable machine learning perspective

Wenzhi Cai, Ao Huang, Zhikai Ding, Qianwei Chen, Si Lu, Jie Li
article en

Abstract

Introduction Postpartum pelvic floor hypertonicity is an underrecognized form of postpartum pelvic floor dysfunction characterized by increased resting pelvic floor muscle activity and myofascial pain. However, predictors of postpartum pelvic floor hypertonicity following spontaneous vaginal delivery remain insufficiently characterized. This study aimed to develop an interpretable machine learning-based prediction model to explore predictors associated with postpartum pelvic floor hypertonicity. Methods This retrospective prediction modeling study included postpartum women who underwent pelvic floor assessment after spontaneous vaginal delivery at Shenzhen Hospital, Southern Medical University, between January 2024 and February 2025. Candidate predictors were obtained from information available by hospital discharge, whereas postpartum pelvic floor hypertonicity was determined at the subsequent first postpartum pelvic floor assessment using surface electromyography combined with digital vaginal palpation. Participants were randomly divided into train and held-out test sets (7:3). Nine machine learning algorithms were evaluated using nested cross-validation. The final model was interpreted using SHapley Additive exPlanations. Results Among 1,297 women, 1,031 were included in the final analysis and randomly divided into train ( n = 722) and test ( n = 309) sets. A total of 505 women were classified as having postpartum pelvic floor hypertonicity (48.98%). Random forest achieved the numerically highest mean outer-fold area under the receiver operating characteristic curve (0.780 ± 0.018) and was selected as the final model. In the held-out test set, the model achieved an area under the receiver operating characteristic curve of 0.789 (95% confidence interval: 0.739–0.838). SHapley Additive exPlanations analysis identified neuraxial analgesia during labor, gestational diabetes mellitus, gestational hypertension, residence, and first stage of labor as the most influential contributors to model predictions. The absence of neuraxial analgesia during labor, the presence of gestational diabetes mellitus and gestational hypertension, and rural residence contributed to higher model-predicted probabilities, whereas first stage of labor showed a nonlinear contribution pattern. Conclusions Interpretable machine learning provided a useful approach for exploring predictors associated with postpartum pelvic floor hypertonicity following spontaneous vaginal delivery. Clinical Trial Registration Chinese Clinical Trial Registry, identifier ChiCTR2600117411.

Frontiers in MedicineVol. 13
Southern Medical University Shenzhen Hospital (CN), Southern Medical University (CN)
Good health and well-being
Pelvic floor disorders treatments
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon