Abstract
Background Pneumonia is the leading cause of preventable mortality under five years of age. Appropriate case management is as essential as disease prevention interventions, especially in primary care settings. Computer science has been used accurately and widely for pneumonia diagnosis; however, prognosis studies are relatively low. Herein, we developed a machine learning-based clinical decision support system tool for childhood pneumonia to provide prognostic support for case management.
Methods We analyzed data from 437 children admitted to our clinic with a pneumonia diagnosis between 2014 and 2020. Pediatricians encoded the raw dataset according to candidate features. Before the experimental study of the machine learning algorithms of Pycaret, SMOTE-Tomek was utilized for managing imbalanced datasets. The feature selection was made by examining the SHAP values of the algorithm with the highest performance and re-modeled with the most important clinical features. We optimized hyperparameters and employed ensemble methods to develop a robust predictive model.
Results Optimized models predicted pneumonia prognosis with %77-88 accuracy. It was shown that severity could be determined over %84 by five clinical features: hypoxia, respiratory distress, age, Z score of weight for age, and antibiotic usage before admission.
Conclusions In this experimental study, we demonstrated that contemporary data science methods, such as oversampling, feature selection, and machine learning tools, are promising in predicting the critical care need of patients. Even in small-size samples like our study, ML methods can reach current wisdom.
Highlights
Pneumonia accounts for 14% of mortality in children under 5, with over 740,000 deaths in 2019 alone.
WHO and UNICEF’s GAPPD aims to cut mortality rates by focusing on vaccinations, sanitation, breastfeeding, and addressing pediatric HIV.
Accurate diagnosis and timely treatment can reduce pneumonia mortality by up to 28%, yet diagnosing can be challenging.
Many in underdeveloped regions lack access to essential equipment and trained staff, exacerbating mortality rates.
Data science and machine learning offer promising solutions for pneumonia management, especially in LMICs, with a focus on prognostic support.
Competing Interest Statement
The authors have declared no competing interest.
Clinical Protocols
https://pnm.pediatridefteri.com/
Funding Statement
This study did not receive any funding
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
This study design and procedures were approved by Hacettepe University Clinical Research Ethics Committee with protocol number GO-20/1182.
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Data Availability
All data produced in the present study are available upon reasonable request to the authors
