Sign up to save your podcasts
Or
Share Univfy IVF live birth prediction models: real-world validation and superior prediction over SART model
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Univfy IVF live birth prediction models: real-world validation and superior prediction over SART model
This podcast summarizes the research findings reported by Yao et al in Nature Communications 2025 (https://www.nature.com/articles/s41467-025-58744-z).
Check out:
- Dr. Mylene Yao's related LinkedIn post: https://bit.ly/4cLV4ZL
- Univfy press release: https://bit.ly/3RsUHcJ
- The actual research article: https://go.nature.com/4jGaFfx
Created with Google Notebook LM. Content reviewed and approved by Univfy.
Context: The researchers highlight the significant public health issue of infertility and the existing barriers to affordable and accessible in vitro fertilization (IVF) treatment.
This research compares machine learning center-specific (MLCS) models developed by Univfy with the US national registry-based model (SART, Society for Reproductive Technology) for predicting live birth outcomes in IVF. The research, published in Nature Communications 2025 , indicates that MLCS models are superior in accuracy and clinical utility, particularly in minimizing false predictions and providing validated and personalized probability of having a live birth from IVF. Over 25% of patients would receive an underestimation of IVF live birth prognostics from the SART model compared to the Univfy models and the prognosis given by the Univfy models are more appropriate in those cases.
The authors also explain that the improved predictive power of MLCS models has significant implications for patient-centric care, patient counseling, enabling value-based IVF pricing, increasing utilization rates. Collectively, these benefits are also expected to sustain population growth.
View all episodes
By Mylene Yao, MDThis podcast summarizes the research findings reported by Yao et al in Nature Communications 2025 (https://www.nature.com/articles/s41467-025-58744-z).
Check out:
- Dr. Mylene Yao's related LinkedIn post: https://bit.ly/4cLV4ZL
- Univfy press release: https://bit.ly/3RsUHcJ
- The actual research article: https://go.nature.com/4jGaFfx
Created with Google Notebook LM. Content reviewed and approved by Univfy.
Context: The researchers highlight the significant public health issue of infertility and the existing barriers to affordable and accessible in vitro fertilization (IVF) treatment.
This research compares machine learning center-specific (MLCS) models developed by Univfy with the US national registry-based model (SART, Society for Reproductive Technology) for predicting live birth outcomes in IVF. The research, published in Nature Communications 2025 , indicates that MLCS models are superior in accuracy and clinical utility, particularly in minimizing false predictions and providing validated and personalized probability of having a live birth from IVF. Over 25% of patients would receive an underestimation of IVF live birth prognostics from the SART model compared to the Univfy models and the prognosis given by the Univfy models are more appropriate in those cases.
The authors also explain that the improved predictive power of MLCS models has significant implications for patient-centric care, patient counseling, enabling value-based IVF pricing, increasing utilization rates. Collectively, these benefits are also expected to sustain population growth.