Skip to main content
Download PDF
  • Systematic Review
  • Open access
  • Published:

Applications of machine learning approaches for pediatric asthma exacerbation management: a systematic review

BMC Medical Informatics and Decision Making volume 25, Article number: 170 (2025) Cite this article

Abstract

Background

Pediatric asthma is a common chronic respiratory disease worldwide, and its acute exacerbation events significantly impact children’s health and quality of life. Machine learning, an advanced data analysis technique, has shown great potential in healthcare applications in recent years. This systematic review aims to assess the application of ML techniques in pediatric asthma exacerbation and explore their effectiveness and potential value.

Methods

Studies from four electronic databases, including PubMed, EBSCO, Elsevier, and Web of Science, from Jan 2000 to Jan 2025, were searched. Studies applying the ML methods for pediatric asthma exacerbation and published in English were eligible. The risk of bias and applicability of the included studies was assessed using the Effective Public Health Practice Project (EPHPP) quality assessment tool.

Results

A total of 23 studies were selected for inclusion in this review, covering different ML models such as decision trees, neural networks, and support vector machines. These studies focused on analyzing risk factors for asthma exacerbation, diagnosing and predicting, optimizing and allocating healthcare resources, and comprehensive asthma management. The results show that ML techniques have significant advantages in the application of pediatric asthma exacerbation and in the provision of personalized health care.

Conclusions

ML techniques show great promise for application in pediatric asthma exacerbations. With further research and clinical validation, these techniques are expected to provide strong support for diagnosis, personalized treatment, and long-term management of pediatric asthma exacerbation.

Clinical trial number

Not applicable, Prospero registration number CRD42024559232.

Peer Review reports

Background

Asthma is one of the most common chronic respiratory diseases and has a significant impact on sufferers and their families. The symptoms of asthma include coughing, wheezing, chest tightness and shortness of breath. These symptoms vary from person to person, ranging from mild to severe, and may occur frequently or infrequently. An asthma exacerbation occurs when these symptoms worsen [1, 2]. Asthma exacerbation in adolescents and children aged 6–11 years according to the 2024 Global Strategy for Asthma Management and Prevention guidelines is defined as follows: Exacerbations of asthma are episodes characterized by a progressive increase in symptoms of shortness of breath, cough, wheezing or chest tightness and progressive decrease in lung function. Exacerbations may occur in patients with a preexisting diagnosis of asthma or, occasionally, as the first presentation of asthma. Among other things, the guidelines defined asthma exacerbation in children under 5 years of age as an acute or sub-acute deterioration in symptom control that is sufficient to cause distress or risk to health and necessitates a visit to a healthcare provider or requires treatment with systemic corticosteroids [3]. These exacerbations often result in school absences, parental absenteeism, unplanned emergency department (ED) visits, and hospitalizations, severely affecting the health-related quality of life (HRQL) of children and their parents [4]. Despite interventions, nearly half of pediatric asthma patients experience exacerbations annually, with 1/6 requiring ED visits and 1/20 hospitalization. These visits account for over 1.8 million ED visits and more than 60% of the total cost of asthma care [5,6,7,8].

In recent years, the use of machine learning (ML) algorithms in pediatric asthma exacerbations has shown significant potential. ML algorithms can assist physicians in more accurately diagnosing asthma by analyzing clinical data, such as electronic health records (EHRs), and pulmonary function test results, as well as predicting the risk of asthma exacerbations by integrating multiple sources of data, including clinical indicators, environmental factors, and socioeconomic factors [9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31]. It also provides patients with more personalized and precise medical services through its unique ability to process and analyze large and complex datasets and capture highly nonlinear relationships and complex interactions in the data [32,33,34,35]. However, there is still a lack of systematic review on the applications of ML in pediatric asthma exacerbation management.

This study aims to comprehensively analyze the current research progress of ML techniques in pediatric asthma exacerbation management, identify the critical risk factors, evaluate the effectiveness of these techniques, and explore the potential applications of ML in the diagnosis and prediction of pediatric asthma exacerbations, personalized treatment, and long-term health management. The knowledge synthesis from this study may provide a scientific basis for clinical decision-making, policy formulation, and health education, potentially improving the quality and efficiency of care in the future.

Methods

Search strategy

This systematic review complies with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards. Institutional review board approval was not required, as publicly available data were used and no human subjects were involved. A comprehensive search was performed in PubMed, EBSCO, Elsevier, and Web of Science, covering the period from January 2000 to January 2025. Studies published before 2000 were ineligible since they were considered less relevant to modern asthma care [36]. The search strategy was centered around the terms “Asthma,” “Asthma exacerbation/attack/deterioration,” and “Machine learning/Deep learning” and included their appropriate derivatives and synonyms such as “Asthmas” OR “Bronchial Asthma” OR “Asthma, Bronchial”, “Learning, Machine” OR “Transfer Learning” OR “Learning, Transfer”, “Deep Learning” OR “Learning, Deep” OR “Hierarchical Learning” OR “Learning, Hierarchical”. Additionally, we examined the reference lists of included articles to identify any additional relevant studies not retrieved by the automatic searches. The whole search strategy [see Additional file 1].

Inclusion and exclusion criteria

Eligibility criteria for inclusion were: (1) The study subjects included children and/or adolescents under 18 years; (2)The study subjects had asthma exacerbation/attack/deterioration and/or asthma exacerbation/attack/deterioration included in the primary or secondary outcome; (3) The language of the article was in English; (4) Research or application of ML for asthma exacerbation/attack/deterioration was conducted; (5) Observational studies (including retrospective, prospective, cohort studies, case-control studies, etc.) and randomized controlled trials (RCT) were eligible.

We excluded the following: (1) Books or dissertations or thesis or conference abstracts or comments or patents or awarded grant or editorial, or case reports; (2) Systematic reviews or Meta-analysis; (3) Non-full text articles.

Screening and data extraction

Two authors (Chunni Zhou and Liu Shuai) independently scanned abstracts, titles, and citations retrieved by electronic and hand searches against the inclusion criteria to assess eligibility. Two review authors independently reviewed the full-text studies retrieved to determine final eligibility. Disagreements were discussed and resolved by consensus, and if necessary, a third author (Meng Li) was involved.

Data were extracted by one reviewer (Chunni Zhou) and checked for consistency by the other two reviewers (Liu Shuai and Meng Li). Data extracted included first author, country, year, study design, data collection period, study population, sample size, type of ML algorithm, definition of asthma exacerbation/attack/deterioration, outcome events, and study results. For the ML algorithms, we also extracted validation methods and performance metrics. In addition, we read through each study to generalize and categorize the research objectives. Due to significant methodological heterogeneity among the studies, a meta-analysis was not conducted. Instead, a narrative synthesis of the results was performed, and complete details of the included studies are reported in Table 1.

Table 1 Main characteristics of the identified studies reported in these studies
Full size table

Quality assessment

The Effective Public Health Practice Project (EPHPP) quality assessment tool for quantitative studies [37] was used to assess each research regarding potential biases and global study quality. Studies were given a global rating of strong, moderate, or weak based on the score. The tool was used for removing confounders, blinding, intervention integrity, and analysis, as these were irrelevant to the study designs assessed in this review. This left the following areas: selection bias, study design, data collection methods, and withdrawals and dropouts. One author (Chunni Zhou) conducted this assessment, and a discussion was undertaken with the second author.

Results

Study selection

Figure 1 shows the study selection process. A total of 675 papers were identified from four databases (PubMed (136), Elsevier (47), Web of Science (333), and Ebsco (159)). After excluding 335 duplicates, 340 papers were screened by titles and abstracts, leading to 31 potentially eligible papers. Then, full-text screening of these articles confirmed eligibility for 16. Additionally, examining the references of these articles yielded 7 more, totaling 23 articles included in the review (Fig. 1).

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram

Full size image

Study characteristics

The publication year of these papers they were ranged from 2006 to 2024. Eight studies were published between 2006 and 2015 [9,10,11,12,13,14,15, 28] and 15 between 2016 and 2024 [16,17,18,19,20,21,22,23,24,25,26,27, 29,30,31]. Among these studies, 14 studies were from America [9,10,11, 14, 16, 17, 19, 23,24,25,26,27, 29, 30], three were from the Netherlands [13, 15, 18], and one each was from Canada [12], Greece [20], Korea [21], Japan [22], China [28] and Poland [31] (Fig. 2).

Fig. 2
figure 2

Study characteristics of the included studies. Note (a) study setting of the included studies; (b) study design of the included studies; (c) study population of the included studies; (d) publication years of the included studies; (e) sample size of the included studies

Full size image

Nine were prospective studies [9, 12,13,14,15, 17, 18, 21, 31], including two prospective cohort studies [17, 18] and two prospective longitudinal studies [13, 15], with follow-up times of one [13, 15, 18], and three years [17]. Thirteen were retrospective studies [10, 11, 16, 19, 20, 22, 24,25,26,27,28,29,30], including one retrospective cohort study [27] and one case-crossover study [16], with follow-up periods of two [27] and eleven years [16]. One study was a randomized controlled trial with a 1-year follow-up [23]. In the randomized controlled trial, the intervention positively affected exacerbation outcomes [23] (Fig. 2).

The age range for populations varied across the studies: five included children aged 2–18 years [9, 10, 14, 19, 30], three included children aged 0–17 years [23, 25, 31], two included children aged 5–18 years [26, 27], two included children aged 6–18 years [15, 18], one each included children aged six months to 15 years [22], 1–14.5 years [20], 1–17 years [12], 2–21 years [24], 5–12 years [11], 6–14 years [21], 6–16 years [13] and 6–17 years[29] and three did not specify the age of the pediatric participants [16, 17, 28] (Fig. 2).

The sample size varied from 14 to 54981 (eight studies ≤ 100 [13, 15, 17, 18, 20, 21, 28, 31], six studies ≤ 1,000 [11, 12, 14, 23, 25, 30], four studies ≤ 10,000 [9, 10, 24, 26], four studies > 10,000 [16, 19, 22, 27] and one study not explained [29]). Full details of the included studies are reported in Table 1 (Fig. 2).

Definition of exacerbation and outcome

Ten studies did not specify a definition of asthma exacerbation [10, 14, 16, 17, 20, 21, 24, 28,29,30,31], three studies defined asthma exacerbation according to the most recent ATS/ERS [13, 15, 18], three studies defined asthma exacerbation using emergency room visits and hospitalizations [11, 12, 27], two studies defined asthma exacerbation using emergency room visits/hospitalizations or oral corticosteroid [23, 25], two studies defined asthma exacerbation using International Classification of Diseases codes ICD9 or − 10 [22, 26], one study defined asthma exacerbation using concomitant receipt of albuterol and systemic corticosteroids [19], one study defined asthma exacerbation using only emergency room visits [9] and one study defined asthma exacerbation using emergency room visits or hospitalizations or outpatient visit with usage of oral corticosteroids medications [29]. Seventeen of the studies had a primary or secondary outcome of pediatric asthma exacerbation [9,10,11,12,13, 15,16,17,18, 20, 23, 25, 26, 28,29,30,31], while the remaining six studies included populations with pediatric asthma exacerbations [14, 19, 21, 22, 24, 27]. Therefore, the outcomes of the studies were mainly related to emergency room visits or hospitalizations [19, 24, 27], peak expiratory flow rate (PEFR) values [21], asthma control exacerbation [14], and antibiotic variants and adjunctive therapy [22].

ML model related characteristics

Regarding learning algorithms, 23 studies utilized 59 different ML algorithms, which were categorized with the most popular being Bayesian Networks (BN), followed by Random Forests (RF), Decision Trees (DT), Neural Networks (NN) and Support Vector Machines (SVM) (Fig.3).

Fig. 3
figure 3

Histogram of different ML algorithms for asthma exacerbation management. Note: There were a total of 59 models across the 23 studies, which were categorized to give a total of 12 categories of models

Full size image

Because of data limitations, only five studies were included in this study to research the performance of their ML models, and the performance metrics were accuracy, sensitivity, specificity, Area Under the Curve (AUC), Positive Predictive Value (PPV) and Negative Predictive Value (NPV). Overall, most of the values of the six metrics of all the models in the five studies were high, especially the specificity, AUC and NPV were close to 100% for some models, and among all the metrics, the distribution of accuracy was the most concentrated, with most models around 70%, and the remaining metrics have a more dispersed distribution of values. The five models of Dexheimer JW et al [10] have higher values of sensitivity, specificity, AUC and NPV, among which the NPV of BN, Max-Min Hill-Climbing (MMHC) and Gaussian process (GP) reaches 98.9%, but the PPV of the five models are low, especially Artificial Neural Network (ANN), which has a PPV of only 38%. The values of the five models of Gardeux V et al [17] for the six metrics are relatively close to each other, with values between 60% and 70%. The accuracy, specificity, AUC and NPV of the six models of Luo G et al [14] were around 70% and 80%, but the sensitivity and PPV values were lower, with RF having the lowest sensitivity of 38.3%, showing a rectangular pattern (Fig. 4).

Fig. 4
figure 4

Radar charts of ML model performance metrics. Note: Because of data limitations, all five models studied by Dexheimer JW were missing accuracy, with the SVM also missing AUC; the NB model studied by Farion KJ was missing AUC; and the BN model studied by Sanders DL was missing accuracy. BN, Bayesian Network; MMHC, Max-Min Hill-Climbing; ANN, Artificial Neural Network; GP, Gaussian process; SVM, Support Vector Machine; NB, Naive Bayes model; RF, Random Forest; DT, Decision Tree; KNN, K- Nearest Neighbor; DS, Decision stumps; DNN, Deep Neural Network

Full size image

Farion KJ et al [12] and Sanders DL et al [9] each included only one model. In the NB model of Farion KJ et al [12], the highest value of the five reported metrics was 85% (PPV), while the lowest value was 53% (NPV). In contrast, the Bayesian network model of Sanders DL et al [9] reported the highest value as 98.8%(NPV) and the lowest value as 44.7% (PPV).

Regarding algorithm validation methods, nine studies used cross-validation [9, 12,13,14, 19,20,21, 29, 31], six studies used split-sample validation [10, 15, 16, 24,25,26], two studies used bagging validation [18, 27], two studies used holdout validation [11, 17], and four studies did not mention validation methods [22, 23, 28, 30].

Of the 23 studies included, 21 studies dealt with classification tasks using ML models [9,10,11,12,13,14,15, 17,18,19,20,21, 23,24,25,26,27,28,29,30,31], only one study dealt with clustering tasks [22], and one study dealt with association analysis tasks using association rule mining [16].

The study showed that the number of variables entered into the model varied across studies, with 5 studies having a variable count of ≤ 10 variables [9, 18, 21, 22, 27], 13 studies having a variable count of ≤ 50 variables [10, 12, 14,15,16,17, 19, 20, 23,24,25, 28, 29], 1 study having a variable count of > 100 variables [13], and 4 studies having an unknown specific variable count [11, 26, 30, 31]. In the 23 studies, the types of variables input into the model were a mixture of numerical and categorical variables [9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31]. The numerical variables mainly included physiological indicators of patients (such as BMI, and pulmonary function indicators), environmental data, and genetic data. The categorical variables mainly included basic patient characteristics (such as gender, age, and ethnicity), clinical characteristics (such as allergic constitution, comorbid conditions, and asthma severity), treatment-related factors (such as medication use and treatment group), and other clinical diagnostic information (such as symptom severity).

In addition, regarding Explainable Artificial Intelligence (XAI), five studies were found to use feature importance maps to improve the interpretability of model results [11, 18, 19, 24, 27], five studies to use interpretable models and visualize model structures and processes [9, 10, 16, 20, 28] and one study to use the SHapley Additive explanation (SHAP) method to calculate shapley values [30].

Applications of ML algorithms in pediatric asthma exacerbations (categorized by disease management)

Assessment of risk factors

Eight studies have utilized ML to assess risk factors for pediatric asthma exacerbations, exploring genomic data, environmental elements, and socioeconomic status [11, 16, 17, 21, 25,26,27,28]. Two studies focus on genomics [11, 17], two on social factors [25, 27] and three on environmental factors such as indoor and outdoor air pollutants [16, 21, 26, 28]. Some studies have analyzed the effects of genetic polymorphisms and air pollutants on pediatric asthma exacerbations using Random Forest classifiers and association rule mining techniques [11, 16, 28], and one of them proposed two novel data mining methods (pattern-based decision tree (PBDT) and pattern-based class association rule (PBCAR)) to combine patient biosignals and environmental data for the application of asthma exacerbation. others have used deep learning algorithms, such as Long Short-Term Memory (LSTM) modeling, to predict the risk of pediatric asthma exacerbations and assessed the effects of indoor particulate matter concentrations on PEFRs in pediatric asthma [21]. In addition, studies also evaluated the performance of ML models across different socioeconomic groups, aiming to minimize biases [25, 27]. Together, these studies emphasize the importance of individual differences and environmental factors in the management of pediatric asthma exacerbations.

Diagnosis and prediction of pediatric asthma exacerbations

Nine studies have applied ML to diagnose pediatric asthma exacerbations with high accuracy, leveraging clinical data and patient characteristics [9, 10, 12, 13, 15, 18, 20, 30, 31]. Three studies developed and evaluated Bayesian networks for diagnosing patients in line with treatment guidelines in pediatric emergency departments and predicting exacerbations post-medication withdrawal [9, 10, 20]. Three other studies utilized Volatile Organic Compounds (VOCs) and inflammatory markers to diagnose and predict exacerbations, with one noting high accuracy for certain VOC combinations [13, 15, 18]. One study used this to accurately identify pediatric asthma exacerbations from prehospital records by modifying an existing rule-based computable phenotype (CP) and creating a new machine learning-based CP [30]. One study used an AI-assisted home stethoscope and found that the parameters provided by the device were very effective in detecting pediatric asthma exacerbations [31]. Still another study compared the efficacy of various algorithms, including Bayesian networks, ANNs, SVMs, and Gaussian processes, in predicting asthma exacerbations in the ED, concluding that all achieved high accuracy [12].

Optimization and allocation of medical resources

Three studies explored the use of ML for optimizing medical resource allocation in pediatric asthma exacerbations [19, 22, 24]. One study compared four ML models, combining clinical, environmental, and social data to predict demand for hospitalization, with gradient boosting performing the best results [19]. Another applied automated machine learning algorithms (autoML), which outperforming traditional ML [24]. The third analyzed hospitalization patterns in Japan, founding that antibiotic use and the use of other adjunctive treatments differed significantly between hospitals [22].

Comprehensive asthma management

Three studies have applied ML to comprehensive pediatric asthma exacerbations management [14, 23, 29]. One study combined ML algorithms to predict asthma control exacerbations one week in advance by analyzing clinical and environmental data [14]. Two studies evaluated an artificial intelligence-assisted clinical decision support system (AI-CDS) [23, 29], specifically the Asthma Guidance and Prediction System (A-GPS), which uses EHRs to provide clinical information and predict the risk of asthma exacerbation [23]. While all three studies utilized ML to enhance asthma management, the former study emphasized model development and improvement of prediction ability [14] and the latter two assessed AI-CDS in clinical practice [23, 29].

Quality assessment

The EPHPP quality assessment in Fig. 5 rated 5 studies as strong [13, 15, 18, 23, 27], 15 as moderate [10,11,12, 14, 16, 17, 19,20,21,22, 24,25,26, 30, 31] and 3 as weak [9, 28, 29].

Fig. 5
figure 5

Quality assessment. Note The confounders, blinding, intervention integrity, and analyses did not apply to any of the studies and were therefore removed

Full size image

Discussion

This systematic review covers a broader and more recent time frame, spanning from Jan 2000 to Jan 2025, providing insights into the evolving use of ML in pediatric asthma exacerbations. It contains more than just predictions [4, 36, 38, 39], covering risk factor assessment, diagnosis and prediction of pediatric asthma exacerbations, optimization and allocation of medical resources, and comprehensive asthma management, offering a more holistic understanding of ML’s role in this domain.

Findings

The review also highlights that majority of ML applications in the included studies were predictive models for pediatric asthma exacerbations. This trend likely reflects the rapid onset and dynamic nature of asthma exacerbation, which can be life-threatening. Prediction of pediatric asthma exacerbations plays a crucial role in enabling preventive interventions and targeted treatment [40]. Furthermore, disease diagnosis can provide physicians with actionable predictive data to support decision-making, enhance healthcare process efficiency, and reduce costs [40,41,42]. ML algorithms have been applied not only in predicting acute exacerbations of chronic obstructive pulmonary disease (COPD) [43,44,45] and acute kidney injury [46, 47], but also in detecting and diagnosing conditions such as asthma, heart disease, and diabetes [48,49,50]. Additionally, ML aids in creating personalized treatment plans for diseases like cancer and rheumatoid arthritis [51, 52], while optimizing healthcare resources management [53, 54].

Because of the broad definition of pediatric asthma exacerbation, this study included studies that explicitly stated the keyword “asthma exacerbation”. This review also found that the definitions of pediatric asthma exacerbations varied across studies, with most studies defining pediatric asthma exacerbations as hospitalization, emergency visits, and specific medical interventions, and the differing definitions of exacerbations lead to non-comparable findings. In addition, different definitions may have an impact on the diagnosis of pediatric asthma exacerbations, with looser definitions potentially including more potential pediatric asthma exacerbations, thus increasing sensitivity, but also diagnosing asthma exacerbations in patients who do not have an asthma exacerbation, which reduces specificity and leads to more false-positive diagnoses; therefore standardizing the definition of pediatric asthma exacerbation could help improve both the quality of the study and the accuracy of the diagnosis [55, 56].

ML research in pediatric asthma is also different from adult asthma. The pathogenesis of pediatric asthma is relatively complex, and many risk factors are still unknown; therefore, in studies of pediatric asthma, the focus is usually on factors associated with child growth and development, such as family history and genetic predisposition [11, 17, 32]. Still, these factors are often difficult to control, and these factors may be less significant in studies of adult asthma exacerbations, which have focused more on lifestyle and environmental factors, such as air pollution and occupational exposures [57, 58]. Thus, in pediatric asthma, controlling certain environmental factors, such as tobacco smoke exposure, pet hair and dust mites, may reduce the risk of asthma exacerbation [26, 32, 59]. In addition, ML models for pediatric asthma focus more on the explanatory nature of the models so that they can be accepted and used by physicians and parents. The interpretability of black-box models (such as the more complex ML and DL models) can be improved using techniques like feature importance analysis and Local Interpretable Model-agnostic Explanations (LIME), which are post-hoc interpretation methods [60, 61]. Alternatively, one can directly use interpretable models, such as linear models and decision trees, which have simple structures and are easy to understand and interpret. In addition, textual interpretation and visualization of models or results can improve the interpretability of models and results to some extent. There is no unified and objective standard for assessing interpretability, and different methods and application scenarios may require different assessment indicators [62]. In addition, approaches integrating multiple ML algorithms have shown promising results in pediatric asthma exacerbation studies, especially when considering multiple meteorological, environmental, and pollen factors [26]. This suggests that combining different machine-learning algorithms may provide more accurate models for pediatric asthma exacerbation studies [12, 14, 17, 21]. Meanwhile, the emerging development of deep learning models has also shown advantages as they can efficiently process and refine the complex nonlinear relationships between risk factors, thus improving the model accuracy [14, 21, 63].

The inclusion of this systematic review revealed that the number of variables entered into the model varied across studies. Upon comparison, it was found that studies with a smaller number of model inputs included a correspondingly smaller number of study subjects, used fewer types of ML algorithms, and had simple models that were easy to manipulate, with high model interpretability, and that the ML in these types of studies was more focused on the application of clinical practice for pediatric asthma exacerbation; whereas, studies with a larger number of model inputs included a very large number of study subjects, and also used multiple ML algorithms for comparison, however crosswise, the accuracy of the resulting models was also higher, and the study focused more on ML method and prediction performance.

When assessing model performance, using a combination of metrics is essential, as relying on a single metric can be misleading. Different metrics capture different aspects of model performance, providing a more comprehensive evaluation. For instance, using accuracy alone can be very deceptive, particularly with unbalanced dataset. In such cases, accuracy may overestimate model effectiveness by favoring the majority class while masking poor performance on minority classes. While AUC (area under the ROC curve) is robust to class imbalanced and offers a holistic view of model performance, it also has limitations. AUC does not convey how well the model performs at particular decision thresholds, which is critical for practical applications. For example, in pediatric asthma exacerbation studies, the performance of the model under specific decision thresholds can directly affect clinical outcomes. Therefore, it is necessary to refer to metrics such as sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). For example, if the model is used for screening and diagnosis of pediatric asthma exacerbations, then high sensitivity and NPV are essential to reduce underdiagnosis and underreporting, while high specificity and PPV help to minimize misdiagnosis and misreporting [40, 64,65,66].

To effectively evaluate model performance, the choice of metrics should align with the characteristics of the dataset and the objectives of the modeling task. AUC is widely used to comprehensively assess a model’s ability to distinguish between categories. It summarizes the model’s performance under different thresholds and is particularly useful for unbalanced datasets. Sensitivity and specificity are critical when evaluating the performance of a model on positive and negative classes. For situations where false positives and false negatives need to be balanced, metrics such as precision and recall (sensitivity) can be combined into an F1 score to provide a balanced assessment, especially in unbalanced datasets. In addition, accuracy is a straightforward metric for measuring the proportion of correct predictions. However, in unbalanced datasets, accuracy can be misleading, and metrics such as Precision-Recall Curve and Area Under Precision-Recall Curve (AUPRC) provide a more accurate picture of model performance.

Ultimately, the selection of evaluation metrics should be tailored to the specific goals of the modeling task. For instance, AUC measures overall discriminatory power, sensitivity and specificity assess category-specific performance, and theF1 scores balances precision and recall, providing a comprehensive understanding of model effectiveness [67,68,69].

Implications and recommendations

Regarding the diagnosis of pediatric asthma exacerbation, the Global Strategy for Asthma Management and Prevention (Updated 2024) can be referred to for a comprehensive diagnosis by combining information from various aspects such as clinical symptoms, medical history, physical examination, and pulmonary function tests. Accurate diagnosis helps to detect signs of asthma exacerbation and take targeted treatment measures. However, it varies from region to region and should be standardized according to local diagnostic criteria [3, 70].

In the clinical use of ML models to assist decision-making, with the help of explanatory tools such as LIME, SHAP, the results of the model can be interpreted to help doctors and parents understand the model’s decisions. For example, through the SHAP method one can clarify which risk factors have a greater impact on the current prediction results, so as to take more targeted preventive measures, but also can choose some of its own ML algorithms with better interpretability, such as decision trees, logistic regression and so on [60,61,62].

When evaluating the performance of the model, in addition to focusing on metrics such as accuracy and AUC, attention should also be paid to metrics such as the sensitivity, specificity, PPV and NPV of the model. These metrics can more comprehensively reflect the performance of the model in practical applications and help doctors and assess the reliability and usefulness of the model [64,65,66, 69].

Strengths and limitations of the study

The strengths of this systemic review lie in its comprehensive search strategy, adherence to a rigorous systematic review methodology and reporting guidelines, and the independent assessment by researchers during title, abstract, and full-text screening, with data extraction verified by multiple reviewers. However, this review has some limitations. First, non-English studies were excluded, which may limit the generalizability of findings. Second, the definition of asthma exacerbation was not standardized across studies, making comparisons difficult. Third, meta-analysis could not be conducted due to significant heterogeneity in study samples, participants, and outcomes.

Future research

Future research on ML in pediatric asthma exacerbations holds considerable promise. Enhancing data quality and diversity is crucial, with the inclusion of broader datasets encompassing pediatric asthma genetic information, environmental factors, and lifestyle habits. Additionally, algorithmic advancements, especially in deep learning, will drive further personalization of treatment, leading to improved efficacy. The integration of real-time monitoring systems using wearables and smart devices will support detection and prevention. Finally, interdisciplinary collaboration among experts in medicine, data science, and engineering will be essential in addressing complex problems and developing more effective asthma management tools.

Conclusions

The systematic review indicates great potential for ML in pediatric asthma exacerbation management, including risk identification, diagnosis, and personalized care. However, challenges such as data quality, algorithm optimization, and interdisciplinary collaboration need to be addressed in clinical practice. Future work should prioritize model robustness, data security, and clinical testing to advance the field.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

ED:

Emergency Department

HRQL:

Health-Related Quality of Life

ML:

Machine Learning

EHRs:

Electronic Health Records

RCT:

Randomized Controlled Trials

EPHPP:

Effective Public Health Practice Project

ERS:

European Respiratory Society

ATS:

American Thoracic Society

PEFR:

Peak Expiratory Flow Rate

BN:

Bayesian Network

RF:

Random Forest

DT:

Decision Tree

NN:

Neural Network

SVM:

Support Vector Machine

AUC:

Area Under the Curve

PPV:

Positive Predictive Value

NPV:

Negative Predictive Value

MMHC:

Max-Min Hill-Climbing

GP:

Gaussian process

ANN:

Artificial Neural Network

XAI:

Explainable Artificial Intelligence

SHAP:

SHapley Additive explanation

LSTM:

Long Short-Term Memory

VOCs:

Volatile Organic Compounds

CP:

Computable Phenotype

auto ML:

Automated Machine Learning

PBDT:

Pattern Based Decision Tree

PBCAR:

Pattern Based Class-Association Rule

AI-CDS:

Artificial Intelligence-assisted Clinical Decision Support

A-GPS:

Asthma Guidance and Prediction System

COPD:

Chronic Obstructive Pulmonary Disease

LIME:

Local Interpretable Model-Agnostic Explanations

ICD:

International Classification of Diseases

AUPRC:

Area Under Precision-Recall Curve

References

  1. Alharbi ET, Nadeem F, Cherif A.Predictive models for personalized asthma attacks based on patient’s biosignals and environmental factors: a systematic review. BMC Med Inform Decis Mak. 2021;21(1):345.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Horne EMF, McLean S, Alsallakh MA, Davies GA, Price BD, Sheikh, et al. Defining clinical subtypes of adult asthma using electronic health records: analysis of a large UK primary care database with external validation. Int J Med Inform. 2023;170:104942.

  3. Global Initiative for Asthma. Global Strategy for Asthma Management and Prevention, 2024. May 2024. Available from: www.ginasthma.org.

  4. Rodríguez-Martínez CE, Sossa-Briceño MP, Forno E.Composite predictive models for asthma exacerbations or asthma deterioration in pediatric asthmatic patients: a systematic review of the literature. Pediatr Pulmonol. 2023;58(10):2703–18.

    Article  PubMed  Google Scholar 

  5. Dexheimer JW, Abramo TJ, Arnold DH, Johnson K, Shyr Y, Ye F, et al. Implementation and evaluation of an integrated computerized asthma management system in a pediatric emergency department: a randomized clinical trial. Int J Med Inform 2014;83:805–13.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Centers for Disease Control and Prevention, National Center for Environmental Health, Asthma Attacks among People with Current Asthma, 2014–2017. 2019. Available from: https://www.cdc.gov/asthma/asthma_stats/attacks-current-asthma.htm#:%7E:text=In%20an%20asthma%20attack%2C%20the%20airways%20become%20inflamed%2C,The%20percentage%20of%20asthma%20attacks%20varied%20by%20state. 22 Jun 2024.

  7. QuickStats: percentage of All Emergency Department (ED) Visits Made by Patients with Asthma, by Sex and Age Group - National Hospital Ambulatory Medical Care Survey, United States 2014-2015. MMWR Morb Mortal Wkly Rep. 2018, 67 (5), 167.

  8. Dexheimer JW, Abramo TJ, Arnold DH, Johnson KB, Shyr Y, Ye F, et al. An asthma management system in a pediatric emergency department. Int J Med Inform 2013;82:230–38.

    Article  PubMed  Google Scholar 

  9. Sanders DL, Aronsky D. Detecting asthma exacerbations in a pediatric emergency department using a Bayesian network. AMIA Annu Symp Proc. 2006;684–88.

  10. Dexheimer JW, Brown LE, Leegon J, Aronsky D.Comparing decision support methodologies for identifying asthma exacerbations. Stud Health Technol Inform. 2007;129(2):880–84.

    PubMed  Google Scholar 

  11. Xu M, Tantisira KG, Wu A, Litonjua AA, Chu JH, Himes BE, et al. Genome Wide Association Study to predict severe asthma exacerbations in children using random forests classifiers. BMC Med Genet. 2011;12:90.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Farion KJ, Wilk S, Michalowski W, O’Sullivan D, Sayyad-Shirabad J.Comparing predictions made by a prediction model, clinical score, and physicians. Appl Clin Inform. 2013;4(3):376–91.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Robroeks CM, Van Berkel JJ, Jöbsis Q, Van Schooten FJ, Dallinga JW, Wouters EF, et al. Exhaled volatile organic compounds predict exacerbations of childhood asthma in a 1-year prospective study. Eur Respir J 2013;42:98–106.

    Article  CAS  PubMed  Google Scholar 

  14. Luo G, Stone BL, Fassl B, Maloney CG, Gesteland PH, Yerram SR, et al. Predicting asthma control deterioration in children. BMC Med Inform Decis Mak. 2015;15:84.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Van Vliet D, Alonso A, Rijkers G, Heynens J, Rosias P, Muris J, et al. Prediction of asthma exacerbations in children by innovative exhaled inflammatory markers: results of a longitudinal study. PLoS One. 2015;10(3).

  16. Toti G, Vilalta R, Lindner P, Lefer B, Macias C, Price D. Analysis of correlation between pediatric asthma exacerbation and exposure to pollutant mixtures with association rule mining. Artif Intell Med. 2016;74:44–52.

    Article  PubMed  Google Scholar 

  17. Gardeux V, Berghout J, Achour I, Schissler AG, Li Q, Kenost C, et al. A genome-by-environment interaction classifier for precision medicine: personal transcriptome response to rhinovirus identifies children prone to asthma exacerbations. J Am Med Inform Assoc 2017;24:1116–26.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Van Vliet D, Smolinska A, Jöbsis Q, Rosias P, Muris J, Dallinga J, et al. Can exhaled volatile organic compounds predict asthma exacerbations in children? J Breath Res 2017;11:01601.

    Article  Google Scholar 

  19. Patel SJ, Chamberlain DB, Chamberlain JM.A machine learning approach to predicting need for hospitalization for pediatric asthma exacerbation at the time of emergency department triage. Acad Emerg Med. 2018;25(12):1463–70.

    Article  PubMed  Google Scholar 

  20. Spyroglou II SG, Rigas AG, Paraskakis EN.Evaluation of Bayesian classifiers in asthma exacerbation prediction after medication discontinuation. BMC Res Notes. 2018;11(1):522.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Kim D, Cho S, Tamil L, Song DJ, Seo S. Predicting asthma attacks: effects of indoor PM concentrations on peak expiratory flow rates of asthmatic children. IEEE Access. 2019;8:8791–97.

    Article  Google Scholar 

  22. Okubo Y, Horimukai K, Michihata N, Morita K, Matsui H, Fushimi K, et al. Recent practice patterns and variations in children hospitalized for asthma exacerbation in Japan. Int Arch Allergy Immunol 2020;181:926–33.

    Article  CAS  PubMed  Google Scholar 

  23. Seol HY, Shrestha P, Muth JF, Wi CI, Sohn S, Ryu E, et al. Artificial intelligence-assisted clinical decision support for childhood asthma management: a randomized clinical trial. PLoS One. 2021;16(8).

  24. Sills MR, Ozkaynak M, Jang H. Predicting hospitalization of pediatric asthma patients in emergency departments using machine learning. Int J Med Inform. 2021;151:104468.

    Article  PubMed  Google Scholar 

  25. Juhn YJ, Ryu E, Wi CI, Katherine King S, Malik M, Romero-Brufau S, et al. Assessing socioeconomic bias in machine learning algorithms in health care: a case study of the HOUSES index. J Am Med Inform Assoc 2022;29:1142–51.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Hurst JH, Zhao C, Hostetler HP, Ghiasi Gorveh M, Lang JE, Goldstein BA.Environmental and clinical data utility in pediatric asthma exacerbation risk prediction models. BMC Med Inform Decis Mak. 2022;22(1):108.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Rezaeiahari M, Brown CC, Eyimina A, Perry TT, Goudie A, Boyd M, et al. Predicting pediatric severe asthma exacerbations: an administrative claims-based predictive model. J Asthma 2024;61:203–11.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Lee CH, Chen JC, Tseng VS.A novel data mining mechanism considering bio-signal and environmental data with applications on asthma monitoring. Comput Methods Programs Biomed. 2011;101(1):44–61.

    Article  PubMed  Google Scholar 

  29. Overgaard SM, Peterson KJ, Wi CI, Kshatriya BSA, Ohde JW, Brereton T, et al. A Technical Performance Study and Proposed Systematic and Comprehensive Evaluation of an ML-based CDS Solution for Pediatric Asthma. AMIA Jt Summits Transl Sci Proc. 2022;2022:25–35. Published 2022 May 23

    PubMed  PubMed Central  Google Scholar 

  30. Harmon I, Brailsford J, Sanchez-Cano I, Fishe J.Development of a Computable Phenotype for Prehospital Pediatric Asthma Encounters. Prehosp Emerg Care. 2025;29(1):10–21.

    Article  PubMed  Google Scholar 

  31. Emeryk A, Derom E, Janeczek K, Kuźnar-Kamińska B, Zelent A, Łukaszyk M, et al. Home Monitoring of Asthma Exacerbations in Children and Adults With Use of an AI-Aided Stethoscope. Ann Fam Med 2023;21:517–25.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Patel D, Hall GL, Broadhurst D, Smith A, Schultz A, Foong RE. Does machine learning have a role in the prediction of asthma in children? Paediatr Respir Rev. 2022;41:51–60.

    PubMed  Google Scholar 

  33. Tsang KCH, Pinnock H, Wilson AM, Shah SA. Application of Machine Learning Algorithms for Asthma Management with mHealth: a Clinical Review. J Asthma Allergy. 2022;15:855–73.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Choi SG, Oh M, Park DH, Lee B, Lee YH, Jee SH, et al. Comparisons of the prediction models for undiagnosed diabetes between machine learning versus traditional statistical methods. Sci Rep 2023;13:13101.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Ley C, Martin RK, Pareek A, Groll A, Seil R, Tischer T.Machine learning and conventional statistics: making sense of the differences. Knee Surg Sports Traumatol Arthrosc. 2022;30(3):753–57.

    Article  PubMed  Google Scholar 

  36. Lowden R, Turner S.Past asthma exacerbation in children predicting future exacerbation: a systematic review. ERJ Open Res. 2022;8(4):1–17.

    Article  Google Scholar 

  37. Effective Public Health Practice Project (EPHPP). Quality Assessment Tool for Quantitative Studies. 2009. Available from: https://www.ephpp.ca/quality-assessment-tool-for-quantitative-studies/. 1 Jul 2024.

  38. Kothalawala DM, Kadalayil L, Weiss VB, Kyyaly MA, Arshad SH, Holloway JW, et al. Prediction models for childhood asthma: a systematic review. Pediatr Allergy Immunol 2020;31:616–27.

    Article  PubMed  Google Scholar 

  39. Ma L, Tibble H. Primary Care Asthma Attack Prediction Models for Adults: a Systematic Review of Reported Methodologies and Outcomes. J Asthma Allergy. 2024;17:181–94.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Xiong S, Chen W, Jia X, Jia Y, Liu C.Machine learning for prediction of asthma exacerbations among asthmatic patients: a systematic review and meta-analysis. BMC Pulm Med. 2023;23(1):278.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Xiang Y, Ji H, Zhou Y, Li F, Du J, Rasmy L, et al. Asthma exacerbation prediction and risk factor analysis based on a time-sensitive, attentive neural network: retrospective cohort study. J Med Internet Res. 2020;22(7).

  42. Finkelstein J, Jeong IC.Machine learning approaches to personalize early prediction of asthma exacerbations. Ann N Y Acad Sci. 2017;1387(1):153–65.

    Article  PubMed  Google Scholar 

  43. Yin H, Wang K, Yang R, Tan YF, Li Q, Zhu W, et al. A machine learning model for predicting acute exacerbation of in-home chronic obstructive pulmonary disease patients. Comput Methods Programs Biomed. 2024;246:108005.

    Article  PubMed  Google Scholar 

  44. Zeng S, Arjomandi M, Tong Y, Liao ZC, Luo G. Developing a machine learning model to predict severe chronic obstructive pulmonary disease exacerbations: retrospective cohort study. J Med Internet Res. 2022;24(1).

  45. Adibi A, Sin DD, Safari A, Johnson KM, Aaron SD, FitzGerald JM, et al. The acute COPD exacerbation prediction tool (ACCEPT): a modelling study. Lancet Respir Med 2020;8:1013–21.

    Article  PubMed  Google Scholar 

  46. Gu M, Liu Y, Sun H, Sun H, Fang Y, Chen L, et al. Using machine learning to predict the risk of short-term and long-term death in acute kidney injury patients after commencing CRRT. BMC Nephrol 2024;25:245.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Yue S, Li S, Huang X, Liu J, Hou X, Zhao Y, et al. Machine learning for the prediction of acute kidney injury in patients with sepsis. J Transl Med 2022;20:215.

    Article  PubMed  PubMed Central  Google Scholar 

  48. Ahsan MM, Siddique Z. Machine learning-based heart disease diagnosis: a systematic literature review. Artif Intell Med. 2022;128:102289.

    Article  PubMed  Google Scholar 

  49. Battineni G, Sagaro GG, Chinatalapudi N, Amenta F.Applications of machine learning predictive models in the chronic disease diagnosis. J Pers Med. 2020;10(2):21.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Kumar N, Narayan Das N, Gupta D, Gupta K, Bindra J.Efficient automated disease diagnosis using machine learning models. J Healthc Eng. 2021;2021(1):9983652.

    PubMed  PubMed Central  Google Scholar 

  51. Sebastiani M, Vacchi C, Manfredi A, Cassone G.Personalized medicine and machine learning: a roadmap for the future. J Clin Med. 2022;11(14):4110.

    Article  PubMed  PubMed Central  Google Scholar 

  52. Wu Y, Li L, Xin B, Hu Q, Dong X, Li Z.Application of machine learning in personalized medicine. Intell Pharm. 2023;1(3):152–56.

    Google Scholar 

  53. Mizan T, Taghipour S. Medical resource allocation planning by integrating machine learning and optimization models. Artif Intell Med. 2022;134:102430.

    Article  PubMed  Google Scholar 

  54. Kaliappan S, Maranan R, Malladi A, Yamsani N Optimizing Resource Allocation in Healthcare Systems for Efficient Pandemic Management using Machine Learning and Artificial Neural Networks. In: 2024 International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC); 2024:1–7.

  55. Fuhlbrigge A, Peden D, Apter AJ, Boushey HA, CA Jr C, Gern J, et al. Asthma outcomes: exacerbations. J Allergy Clin Immunol. 2012;129(3 Suppl).

  56. Herrera-Luis E, Forno E, Celedón JC, Pino-Yanes M.Asthma Exacerbations: the Genes Behind the Scenes. J Investig Allergol Clin Immunol. 2023;33(2):76–94.

    Article  CAS  PubMed  Google Scholar 

  57. Castillo JR, Peters SP, Busse WW.Asthma exacerbations: pathogenesis, prevention, and treatment. J Allergy Clin Immunol Pract. 2017;5(4):918–27.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Jahan J, Li J, Newcomb P.Social and built environmental factors associated with adult asthma hospital admission in the North Central Texas. City Built Environ. 2024;2(1):8.

    Article  Google Scholar 

  59. Akar-Ghibril N, Phipatanakul W.The indoor environment and childhood asthma. Curr Allergy Asthma Rep. 2020;20(9):43.

    Article  PubMed  Google Scholar 

  60. Frasca M, Torre DL, Pravettoni G, Cutica I. Explainable and interpretable artificial intelligence in medicine: a systematic bibliometric review. Discov Artif Intell. 2024;4:15.

    Article  Google Scholar 

  61. Ali S, Abuhmed T, El-Sappagh S, Muhammad K, Alonso-Moral JM, Confalonieri R, et al. Explainable Artificial Intelligence (XAI): what we know and what is left to attain Trustworthy Artificial Intelligence. Information Fusion. 2023;99:101805.

    Article  Google Scholar 

  62. Habiba U, Habib MK, Bogner J, Fritzsch J, Wagner S.How do ML practitioners perceive explainability? an interview study of practices and challenges. Empir Software Eng. 2025;30(1):18.

    Article  Google Scholar 

  63. Ekpo RH, Osamor VC, Azeta AA, Ikeakanam E, Amos BO.Machine learning classification approach for asthma prediction models in children. Health Technol. 2023;13(1):1–10.

    Article  Google Scholar 

  64. Das S, Nayak SP, Sahoo B, Nayak SC. Machine Learning in Healthcare Analytics: a State-of-the-Art Review. Arch Computat Methods Eng. 2024;31:3923–62.

    Google Scholar 

  65. Çorbacıoğlu ŞK, Aksel G.Receiver operating characteristic curve analysis in diagnostic accuracy studies: a guide to interpreting the area under the curve value. Turk J Emerg Med. 2023;23(4):195–98.

    Article  PubMed  PubMed Central  Google Scholar 

  66. Deng F, Huang J, Yuan X, Cheng C, Zhang L.Performance and efficiency of machine learning algorithms for analyzing rectangular biomedical data. Lab Invest. 2021;101(4):430–41.

    Article  CAS  PubMed  Google Scholar 

  67. Dinov ID. Model Performance Assessment, Validation, and Improvement. In: Data Science and Predictive Analytics. The Springer Series in Applied Machine Learning. Cham: Springer; 2023. p. 477–531.

    Chapter  Google Scholar 

  68. Salmi M, Atif D, Oliva D, Abraham A, Ventura S. Handling imbalanced medical datasets: review of a decade of research. Artif Intell Rev. 2024;57:273.

    Article  Google Scholar 

  69. Shreffler J, Huecker MR. Diagnostic Testing Accuracy: sensitivity, Specificity, Predictive Values and Likelihood Ratios. In: StatPearls. Treasure Island (FL): StatPearls Publishing. March 6, 2023.

  70. Skolnik N, Yawn BP, Correia de Sousa J, Vázquez MMM, Barnard A, Wright WL, et al. Best practice advice for asthma exacerbation prevention and management in primary care: an international expert consensus. npj Prim. Care Respir Med. 2024;34:39.

    Google Scholar 

Download references

Acknowledgements

The authors thank all included studies’ authors in this study.

Funding

This work was supported by the Start-Up Research Fund of Southeast University [grant numbers RF1028623262]; the Fundamental Research Funds for the Central Universities [grant numbers 2242024RCB0051]; the “Zhishan” Yong Scholar Programs of Southeast University; the Project of Guangdong Basic and Applied Basic Research Foundation [grant numbers 2022A1515220148]; and the Project of Guangdong Basic and Applied Basic Research Foundation [grant numbers 2022A1515110662]. The funding source had no role in study design; the collection, analysis, and interpretation of data; the writing of the report; or the decision to submit the article for publication.

Author information

Authors and Affiliations

  1. School of Public Health, Southeast University, 87, Dingjiaqiao Road, Gulou District, Nanjing, 210009, China

    Chunni Zhou, Liu Shuai, Lijun Fan, Wei Du & Meng Li

  2. State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Macao, China

    Hao Hu & Carolina Oi Lam Ung

  3. School of Public Health and Management, Guangzhou University of Chinese Medicine, Guangzhou, China

    Yunfeng Lai

  4. Department of Child and Adolescent Health Promotion, Jiangsu Provincial Center for Disease Control and Prevention, 172, Jiangsu Road, Gulou District, Nanjing, 210009, China

    Yan Wang

Authors
  1. Chunni Zhou
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Liu Shuai
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Hao Hu
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Carolina Oi Lam Ung
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Yunfeng Lai
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Lijun Fan
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Wei Du
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Yan Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Meng Li
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

M. L. conceived this article. C. Z. and L. S. searched and selected literature, extracted data and assessed the quality of the included studies. C. Z. wrote the original manuscript. M. L. reviewed and amended the original manuscript. All authors contributed to and have approved the final manuscript.

Corresponding authors

Correspondence to Yan Wang or Meng Li.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Clinical trial number

Not applicable.

Competing interest

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.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

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/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhou, C., Shuai, L., Hu, H. et al. Applications of machine learning approaches for pediatric asthma exacerbation management: a systematic review. BMC Med Inform Decis Mak 25, 170 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12911-025-02990-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12911-025-02990-0

Keywords

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us