机器学习模型预测埃塞俄比亚中裂谷地区疑似肺结核患者的肺结核
Machine learning model development for predicting pulmonary tuberculosis in symptomatic patients from Ethiopia’s Mid-Rift Valley
研究开发并测试了八种监督机器学习模型,以预测埃塞俄比亚中裂谷地区疑似肺结核患者的肺结核状态。随机森林模型表现最佳,AUC为0.94(95% CI 0.90–0.97),准确率为0.88(95% CI 0.84–0.91),灵敏度为0.83(95% CI 0.69–0.92),特异度为0.89(95% CI 0.84–0.92)。
- Article
- Open access
- Published: 13 November 2026
- Dereje Bekele ORCID: orcid.org/0000-0002-4573-34811,2,
- Helen Nigussie1,
- Mujib Abdulkadir1,
- Tariku Dejene3,
- Getu Diriba4,
- Gadissa Gutema1,4,
- Desye Melese Sisay5,6,
- Melat Mezemir7,
- Dawit Wolday4,8,
- Tobias F. Rinke de Wit9,
- Gilman Siu10 &
- …
- Ketema Tafess11,12
Scientific Reports (2026) Cite this article
We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.
Abstract
Tuberculosis continues to pose a serious health threat to high-burden countries, including Ethiopia. Complex interactions among multiple risk factors challenge disease control. Conventional epidemiological methods often fail to capture the complex interactions among risk factors; as a result, targeted interventions to control and prevent the disease are often less effective. The present study attempted to develop and test eight supervised machine learning models to address the methodological gap using comprehensive data collected from tuberculosis-suspected individuals attending public hospitals in the Mid-Rift Valley area of Ethiopia. A cross-sectional study using consecutive sampling was carried out between December 30, 2023, and January 2025 at five hospitals in the Mid-Rift Valley, Ethiopia. The required sociodemographic, clinical, and behavioral data, along with sputum samples, were collected from study subjects with signs and symptoms suggestive of TB using a structured questionnaire. Sputum samples collected from 2025 individuals suspected of having TB at OPDs were tested using the GeneXpert MTB/RIF assay for TB diagnosis. Data were split into training (70%), validation (15%), and test (15%) sets before applying the models. Eight supervised ML models were trained and tuned. Model performance was assessed using metrics: Area Under the Curve (AUC), accuracy, sensitivity, and specificity. 95% confidence intervals were computed using bootstrap resampling (1000 replicates; percentile interval method) for AUC and F1-score, and the Wilson score interval for binomial proportions. We interpreted the best model using SHAP (SHapley Additive exPlanations) and permutation importance. The data were analyzed using Jupyter Notebook, Python Version 3.12.4. Tuberculosis was diagnosed in 280 (13.8%) of the 2025 participants in the study. The Random Forest model performed best at predicting tuberculosis positivity. The final test set showed an AUC of 0.94 (95% CI 0.90–0.97), with an accuracy of 0.88 (95% CI 0.84–0.91), sensitivity of 0.83 (95% CI 0.69–0.92), and specificity of 0.89 (95% CI 0.84–0.92). HIV status, age, Body Mass Index (BMI), Bacillus Calmette-Guérin (BCG) vaccination history, night sweats, and weight loss were identified as strong predictors of tuberculosis disease in the Random Forest model. The Random Forest algorithm showed the best performance in predicting tuberculosis status, based on actual patient data in a high-burden setting. It also showed good promise as a clinical decision-support tool. However, further external validation in independent populations is required before routine clinical implementation.
Abbreviations
- AIDS:
-
Acquired immunodeficiency syndrome
- ANN:
-
Artificial neural networks
- AUC:
-
Area under curve
- BCG:
-
Bacillus Calmette–Guérin
- BMI:
-
Body mass index
- CART:
-
Classification and regression tree
- CXR:
-
Chest X-ray
- DNA:
-
Deoxyribonucleic acid
- EDA:
-
Exploratory data analysis
- FN:
-
False negative
- FP:
-
False positive
- HIV:
-
Human immunodeficiency virus
- KNN:
-
K-nearest neighbors
- LR:
-
Logistic regression
- MDR-TB:
-
Multidrug-resistant TB
- ML:
-
Machine learning
- NAAT:
-
Nucleic acid amplification test
- NB:
-
Naïve Bayes
- NPV:
-
Negative predictive value
- OPD:
-
Outpatient department
- PPV:
-
Positive predictive value
- PTB:
-
Pulmonary tuberculosis
- RF:
-
Random forest
- ROC:
-
Receiver operating characteristic
- SHAP:
-
SHAPley Additive exPlanations
- SMD:
-
Standardized mean difference
- SVM:
-
Support vector machine
- TB:
-
Tuberculosis
- TN:
-
True negative
- TP:
-
True positive
- WHO:
-
World Health Organization
- XDR-TB::
-
Extensively drug-resistant TB
- XGBoost:
-
Extreme gradient boosting
Acknowledgements
We extend our sincere gratitude to the Department of Microbial, Cellular, and Genetics at Addis Ababa University for providing the institutional support and academic environment essential for this research. We are also deeply grateful to the Oromia Regional Health Bureau for their invaluable collaboration, including granting the administrative clearance required to conduct this study in the region’s health facilities, as well as their financial support for the data collection. Our heartfelt thanks go to the dedicated team of laboratory technologists across the five participating hospitals for their diligence and professionalism in collecting the data and samples. Our sincere gratitude goes to Dr. Mekdimu Mezemir Damte for his invaluable guidance and multifaceted support. Finally, and most importantly, we would like to express our sincere appreciation to all the study participants who voluntarily gave their time and provided samples.
Funding
This research received no external funding.
Author information
Authors and Affiliations
Department of Microbial Sciences and Genetics, College of Natural and Computational Sciences, Addis Ababa University, Addis Ababa, Ethiopia
Dereje Bekele, Helen Nigussie, Mujib Abdulkadir & Gadissa Gutema
Communicable and Non-Communicable Diseases Prevention and Control Directorate, Oromia Region Health Bureau, Addis Ababa, Ethiopia
Dereje Bekele
Center for Population and Gender Studies, School of Development Studies, Addis Ababa University, Addis Ababa, Ethiopia
Tariku Dejene
Ethiopian Public Health Institute, Addis Ababa, Ethiopia
Getu Diriba, Gadissa Gutema & Dawit Wolday
Department of Biotechnology, School of Bioscience and Technology, Wollo University, Dessie, Ethiopia
Desye Melese Sisay
Center for Innovative Drug Development and Therapeutic Trials for Africa (CDT-Africa), College of Health Sciences, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia
Desye Melese Sisay
Health Promotion and Disease Prevention Directorate, Addis Ababa City Administration Health Bureau, Addis Ababa, Ethiopia
Melat Mezemir
Department of Biochemistry and Biomedical Sciences, McMaster University, Hamilton, Canada
Dawit Wolday
Department of Global Health, Amsterdam Institute for Global Health and Development (AIGHD), Amsterdam University Medical Center, University of Amsterdam, Amsterdam, The Netherlands
Tobias F. Rinke de Wit
Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Hong Kong
Gilman Siu
Department of Applied Biology, School of Applied Natural Science, Adama Science and Technology University, Adama, Ethiopia
Ketema Tafess
Institute of Pharmaceutical Science, Adama Science and Technology University, Adama, Ethiopia
Ketema Tafess
Authors
- Dereje Bekele
- Helen Nigussie
- Mujib Abdulkadir
- Tariku Dejene
- Getu Diriba
- Gadissa Gutema
- Desye Melese Sisay
- Melat Mezemir
- Dawit Wolday
- Tobias F. Rinke de Wit
- Gilman Siu
- Ketema Tafess
Corresponding author
Correspondence to Dereje Bekele.
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethics approval and consent to participate
The College of Natural & Computational Sciences Institutional Review Board (CNCS-IRB) at Addis Ababa University approved the study. A permission letter was obtained from the Oromia Regional Health Bureau. We obtained informed consent from both participants and the legal guardian for those under 15 years of age. To ensure strict confidentiality, the study used only code numbers throughout. These procedures followed the Declaration of Helsinki.
Consent for publication
Not applicable.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Bekele, D., Nigussie, H., Abdulkadir, M. et al. Machine learning model development for predicting pulmonary tuberculosis in symptomatic patients from Ethiopia’s Mid-Rift Valley. Sci Rep (2026). https://doi.org/10.1038/s41598-026-78644-6
Received: 13 December 2025
Accepted: 27 October 2026
Published: 13 November 2026
DOI: https://doi.org/10.1038/s41598-026-78644-6
Keywords
来源:Nature 健康护理主题 · nature.com