Table 1 Emergency department disposition-related review studies
Source | ED disposition | Database | Analytic strategy | Articles included | Specific disease/condition | Major findings |
|---|---|---|---|---|---|---|
Shung et al. [11] | Admission and mortality | Embase, Medline, Cochrane, Central, Web of Science, WHO COVID-19 Global Literature on Coronavirus Disease, and Google Scholar | Systematic review | 14 | Gastrointestinal bleeding | The median AUC for mortality was 0.84, with AI yielding higher AUCs. Machine learning demonstrated superior performance compared to clinical risk scores for mortality in cases of upper gastrointestinal bleeding. |
Guo et al. [12] | Mortality | PubmMed | Systematic review | 335 | Heart failure | Machine learning helps identify heart failure patients and assess their risk for readmission and mortality accurately. |
Kareemi et al. [13] | Admission and mortality | Medline, Embase, Central, and CINAHL | Systematic review | 23 | No | Machine learning might surpass standard care in predicting outcomes for emergency department patients in diverse clinical scenarios. |
Naemi et al. [14] | Mortality | PubMed, Scopus, and Embase | Systematic review | 15 | No | Eight recommendations for future research to enhance the practical implementation of machine learning in various domains. |
Buttia et al. [15] | ICU admission, intubation, high-flow nasal therapy, extracorporeal membrane oxygenation, and mortality | Embase, Medline, Cochrane Central, Web of Science, WHO COVID-19 Global Literature on Coronavirus Disease, and Google Scholar | Systematic review | 314 | COVID-19 | Several clinical prognostic models for COVID-19, described in the literature, suffer from limited generalizability and applicability due to unresolved statistical and methodological concerns. |
Chen et al. [16] | ICU admission | PubMed, Embase, Cochrane Library, and Web of Science | Systematic review | 10 | No | Machine learning excels in identifying and predicting critically ill patients in emergency department triage. |
Issaiy et al. [17] | ICU admission | PubMed, Embase, Scopus, and Web of Science | Systematic review | 29 | Acute appendicitis | The artificial neural network exhibited high performance across the majority of cases. |
Larburu et al. [18] | Admission | Scopus, PubMed, and Google Scholar | Systematic review | 14 | No | Artificial intelligence models improve emergency department care and ease healthcare system burdens. |
Olender et al. [19] | Mortality | PubMed, Embase, Web of Science, Scopus, and Proquest | Systematic review and meta-analysis | 37 | Older adults (+ 65) | Machine learning models demonstrate strong discriminatory power in predicting mortality. |
Zhang et al. [20] | Mortality | PubMed, Embase, Cochrane Library, and Web of Science | Systematic review and meta-analysis | 50 | Sepsis | Machine learning methods exhibit notably high accuracy in predicting mortality risk among sepsis patients. |