Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Comment
  • Published:

A data-driven future for paediatric surgery

Pediatric Research (2025)Cite this article

Global childhood mortality has declined substantially over recent decades, with under-five deaths reduced by more than 60% worldwide.1 Despite this progress, congenital anomalies remain a leading cause of neonatal and early childhood mortality, contributing disproportionately to the global burden of disease.2 A persistent gap between outcomes in high- and low-income countries underscores the urgent need for strategies that enhance prognostic accuracy and inform both clinical practice and healthy system planning. Concurrently, the past decade has seen rapid growth in the application of machine learning (ML) across medicine. Unlike traditional statistical methods, ML algorithms are able to model complex, non-linear relationships among numerous predictors, offering the potential for improved predictive performance in heterogeneous populations.3 Within this context, Serban et al.4 present in this issue of Paediatric Research a global, multi-centre study applying Random Forest (RF) methodology to an existing data set to predict short-term mortality in children with gastrointestinal anomalies, including oesophageal atresia, gastroschisis, omphalocele, Hirschsprung disease, and congenital diaphragmatic hernia. While the immediate clinical applicability of the findings warrants careful evaluation at the patient level, this study represents an important step toward integrating ML-based predictive models into paediatric surgery and surgical outcomes research more broadly.

The study builds upon data published by the Global PaedSurg Collaboration, an international initiative collecting prospective data on paediatric surgical outcomes from 264 hospitals across 74 countries.5 The Global PaedSurg Collaborative represents a much-needed pivot in paediatric surgical research. Using a Lasso-based regression model, it focused on collaborative cumulation of data to draw conclusions based on a large dataset. Serban et al. aimed to expand on this data set using RF modelling. This is an ML algorithm used for both classification (predicting categories such as survival/death) and regression (predicting continuous values such as length of stay) analysis. It is an ensemble method, combining many models to make a stronger one. It works by using decision trees (hundreds or thousands of them) as the building blocks, with each tree being trained on a random sample of the data set. At each split in a tree, the algorithm only considers a random subset of variables. For classification modelling, each tree ‘votes’ on the outcome, and the majority wins; for regression analysis, the predictions are averaged across the trees. It would be like asking hundreds or thousands of clinicians whether a patient is at risk of dying within 30 days. Each one only sees a random set of patient details and gives an opinion. By taking the majority vote, the overall decision is usually very reliable. This method has been used in a range of papers focusing on surgical outcomes through a range of specialities.6,7

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to the full article PDF.

USD 39.95

Prices may be subject to local taxes which are calculated during checkout

References

  1. Wang, H. et al. Global, regional, national, and selected subnational levels of stillbirths, neonatal, infant, and under-5 mortality, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 388, 1725–1774 (2016).

    Article  Google Scholar 

  2. Wright, N. J. et al. Mortality from gastrointestinal congenital anomalies at 264 hospitals in 74 low-income, middle-income, and high-income countries: a multicentre, international, prospective cohort study. Lancet 398, 325–339 (2021).

    Article  Google Scholar 

  3. A Random Forest based predictor for medical data classification using feature ranking. Inform. Med. Unlocked 15, 100180 (2019).

  4. Serban, A. M. et al. Short-term mortality prediction in children with gastrointestinal congenital anomalies using a random forest classifier. Pediatr. Res. https://doi.org/10.1038/s41390-025-04378-2 (2025).

  5. Wright, N. J. Management and outcomes of gastrointestinal congenital anomalies in low, middle and high income countries: protocol for a multicentre, international, prospective cohort study. BMJ Open (2019) https://doi.org/10.1136/bmjopen-2019-030452.

  6. Cabrera, A. et al. Use of random forest machine learning algorithm to predict short term outcomes following posterior cervical decompression with instrumented fusion. J. Clin. Neurosci. 107, 167–171 (2023).

    Article  PubMed  Google Scholar 

  7. Shi, G. et al. A random forest algorithm-based prediction model for moderate to severe acute postoperative pain after orthopedic surgery under general anesthesia. BMC Anesthesiol. 23, 361 (2023).

    Article  PubMed  PubMed Central  Google Scholar 

  8. Carvalho, M., Pinho, A. J. & Brás, S. Resampling approaches to handle class imbalance: a review from a data perspective. J. Big Data 12, 71 (2025).

    Article  Google Scholar 

  9. Song, S. I., Hong, H. T., Lee, C. & Lee, S. B. A machine learning approach for predicting suicidal ideation in post stroke patients. Sci. Rep. 12, 15906 (2022).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

Download references

Acknowledgements

No specific funding for this commentary. Atul Malhotra’s research is supported by multiple sources, including NHMRC (Australia), MRFF (Australia), Monash Health Foundation. National Stem Cell Foundation and Cerebral Palsy Alliance (Australia).

Author information

Authors and Affiliations

  1. Department of Paediatric Surgery and Surgical Simulation, Monash Children’s Hospital, Melbourne, VIC, Australia

    Damir Ljuhar

  2. Department of Paediatrics, Monash University, Melbourne, VIC, Australia

    Damir Ljuhar & Atul Malhotra

  3. Monash Newborn, Monash Children’s Hospital, Melbourne, VIC, Australia

    Atul Malhotra

  4. The Ritchie Centre, Hudson Institute of Medical Research, Melbourne, VIC, Australia

    Atul Malhotra

Authors
  1. Damir Ljuhar
    View author publications

    Search author on:PubMed Google Scholar

  2. Atul Malhotra
    View author publications

    Search author on:PubMed Google Scholar

Contributions

Both authors contributed to the initial draft, editing and final approval of commentary.

Corresponding author

Correspondence to Atul Malhotra.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ljuhar, D., Malhotra, A. A data-driven future for paediatric surgery. Pediatr Res (2025). https://doi.org/10.1038/s41390-025-04545-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • DOI: https://doi.org/10.1038/s41390-025-04545-5

Search

Advanced search

Quick links